CN114225662A - Flue gas desulfurization and denitrification optimization control method based on hysteresis model - Google Patents
Flue gas desulfurization and denitrification optimization control method based on hysteresis model Download PDFInfo
- Publication number
- CN114225662A CN114225662A CN202111483083.0A CN202111483083A CN114225662A CN 114225662 A CN114225662 A CN 114225662A CN 202111483083 A CN202111483083 A CN 202111483083A CN 114225662 A CN114225662 A CN 114225662A
- Authority
- CN
- China
- Prior art keywords
- data
- desulfurization
- model
- power plant
- flue gas
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000006477 desulfuration reaction Methods 0.000 title claims abstract description 206
- 230000023556 desulfurization Effects 0.000 title claims abstract description 205
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 title claims abstract description 118
- 239000003546 flue gas Substances 0.000 title claims abstract description 118
- 238000000034 method Methods 0.000 title claims abstract description 112
- 238000005457 optimization Methods 0.000 title claims abstract description 20
- MWUXSHHQAYIFBG-UHFFFAOYSA-N nitrogen oxide Inorganic materials O=[N] MWUXSHHQAYIFBG-UHFFFAOYSA-N 0.000 claims abstract description 291
- RAHZWNYVWXNFOC-UHFFFAOYSA-N Sulphur dioxide Chemical compound O=S=O RAHZWNYVWXNFOC-UHFFFAOYSA-N 0.000 claims abstract description 136
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 claims abstract description 87
- 239000002002 slurry Substances 0.000 claims abstract description 77
- 238000012706 support-vector machine Methods 0.000 claims abstract description 62
- 238000005507 spraying Methods 0.000 claims abstract description 44
- 238000004088 simulation Methods 0.000 claims abstract description 43
- 229910021529 ammonia Inorganic materials 0.000 claims abstract description 39
- 238000010521 absorption reaction Methods 0.000 claims abstract description 38
- 238000003745 diagnosis Methods 0.000 claims abstract description 35
- 238000011217 control strategy Methods 0.000 claims abstract description 13
- 238000012806 monitoring device Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 46
- 230000008859 change Effects 0.000 claims description 38
- 230000008569 process Effects 0.000 claims description 35
- 230000000875 corresponding effect Effects 0.000 claims description 33
- 238000010219 correlation analysis Methods 0.000 claims description 30
- 238000010801 machine learning Methods 0.000 claims description 30
- 238000012549 training Methods 0.000 claims description 30
- 230000004927 fusion Effects 0.000 claims description 29
- 238000004422 calculation algorithm Methods 0.000 claims description 28
- 239000010440 gypsum Substances 0.000 claims description 26
- 229910052602 gypsum Inorganic materials 0.000 claims description 26
- 238000005259 measurement Methods 0.000 claims description 24
- 239000013598 vector Substances 0.000 claims description 22
- 235000019738 Limestone Nutrition 0.000 claims description 21
- 235000005770 birds nest Nutrition 0.000 claims description 21
- 239000006028 limestone Substances 0.000 claims description 21
- 235000005765 wild carrot Nutrition 0.000 claims description 21
- 230000004044 response Effects 0.000 claims description 19
- 238000001514 detection method Methods 0.000 claims description 18
- 238000004458 analytical method Methods 0.000 claims description 16
- 238000007781 pre-processing Methods 0.000 claims description 16
- 238000002347 injection Methods 0.000 claims description 14
- 239000007924 injection Substances 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 14
- 238000006243 chemical reaction Methods 0.000 claims description 12
- 230000001276 controlling effect Effects 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000002159 abnormal effect Effects 0.000 claims description 10
- 241000544061 Cuculus canorus Species 0.000 claims description 9
- 230000002596 correlated effect Effects 0.000 claims description 9
- 238000003066 decision tree Methods 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 8
- 238000003860 storage Methods 0.000 claims description 8
- 238000012544 monitoring process Methods 0.000 claims description 7
- 238000007500 overflow downdraw method Methods 0.000 claims description 7
- VTYYLEPIZMXCLO-UHFFFAOYSA-L Calcium carbonate Chemical compound [Ca+2].[O-]C([O-])=O VTYYLEPIZMXCLO-UHFFFAOYSA-L 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 239000003245 coal Substances 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 6
- 230000018044 dehydration Effects 0.000 claims description 6
- 238000006297 dehydration reaction Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 4
- 238000011161 development Methods 0.000 claims description 4
- 238000011049 filling Methods 0.000 claims description 4
- 238000002360 preparation method Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 3
- 238000007664 blowing Methods 0.000 claims description 3
- 229910000019 calcium carbonate Inorganic materials 0.000 claims description 3
- 238000010531 catalytic reduction reaction Methods 0.000 claims description 3
- 238000002485 combustion reaction Methods 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 238000004134 energy conservation Methods 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 239000001301 oxygen Substances 0.000 claims description 3
- 229910052760 oxygen Inorganic materials 0.000 claims description 3
- 238000005295 random walk Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 239000004071 soot Substances 0.000 claims description 3
- 238000004065 wastewater treatment Methods 0.000 claims description 3
- 238000000738 capillary electrophoresis-mass spectrometry Methods 0.000 abstract 1
- 230000003647 oxidation Effects 0.000 description 6
- 238000007254 oxidation reaction Methods 0.000 description 6
- 208000005156 Dehydration Diseases 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 229910052815 sulfur oxide Inorganic materials 0.000 description 5
- VHUUQVKOLVNVRT-UHFFFAOYSA-N Ammonium hydroxide Chemical group [NH4+].[OH-] VHUUQVKOLVNVRT-UHFFFAOYSA-N 0.000 description 4
- 235000011114 ammonium hydroxide Nutrition 0.000 description 4
- 239000007789 gas Substances 0.000 description 4
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 2
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 2
- 238000003915 air pollution Methods 0.000 description 2
- 239000006227 byproduct Substances 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 239000002250 absorbent Substances 0.000 description 1
- 230000002745 absorbent Effects 0.000 description 1
- 238000003916 acid precipitation Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000010790 dilution Methods 0.000 description 1
- 239000012895 dilution Substances 0.000 description 1
- 238000004090 dissolution Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000012717 electrostatic precipitator Substances 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000004868 gas analysis Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000003595 mist Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000006722 reduction reaction Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/346—Controlling the process
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/46—Removing components of defined structure
- B01D53/48—Sulfur compounds
- B01D53/50—Sulfur oxides
- B01D53/501—Sulfur oxides by treating the gases with a solution or a suspension of an alkali or earth-alkali or ammonium compound
- B01D53/502—Sulfur oxides by treating the gases with a solution or a suspension of an alkali or earth-alkali or ammonium compound characterised by a specific solution or suspension
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/74—General processes for purification of waste gases; Apparatus or devices specially adapted therefor
- B01D53/77—Liquid phase processes
- B01D53/78—Liquid phase processes with gas-liquid contact
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/74—General processes for purification of waste gases; Apparatus or devices specially adapted therefor
- B01D53/86—Catalytic processes
- B01D53/8621—Removing nitrogen compounds
- B01D53/8625—Nitrogen oxides
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/74—General processes for purification of waste gases; Apparatus or devices specially adapted therefor
- B01D53/86—Catalytic processes
- B01D53/90—Injecting reactants
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Environmental & Geological Engineering (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Analytical Chemistry (AREA)
- General Chemical & Material Sciences (AREA)
- Oil, Petroleum & Natural Gas (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Treating Waste Gases (AREA)
Abstract
The invention relates to a hysteresis model-based flue gas desulfurization and denitrification optimization control method, which comprises the steps of establishing a power plant desulfurization and denitrification system simulation model according to a modular modeling and control strategy through dynamic simulation software; establishing a first lag time prediction model for the time delay of measuring the existence of the PH value of an absorption tower in a desulfurization system, and establishing a second lag time prediction model for the time delay of measuring the existence of the concentration of nitrogen oxides in flue gas at the inlet of a denitrification system through a flue gas on-line monitoring device (CEMS); predicting the concentration of sulfur dioxide at a desulfurization outlet and the concentration of nitrogen oxide at a denitration inlet in a support vector machine model; controlling the spraying amount of the slurry according to the predicted value of the concentration of sulfur dioxide and the combination of a first lag time prediction model, and controlling the ammonia spraying amount according to the predicted value of the concentration of nitrogen oxide and the combination of a second lag time prediction model; and issuing the control parameters to a simulation model of the desulfurization and denitrification system for intelligent diagnosis.
Description
Technical Field
The invention belongs to the technical field of desulfurization and denitrification, and particularly relates to a hysteresis model-based flue gas desulfurization and denitrification optimization control method.
Background
With the rapid development of various industries, the accompanying air pollution is also more serious, the emission of sulfur dioxide and nitrogen oxides is always high, and the problems of air pollution and acid rain are also serious, and a thermal power plant is one of the main sources of the emission of the sulfur dioxide and the nitrogen oxides, so that the control of the emission of the sulfur dioxide and the nitrogen oxides in the power plant is urgent. The state department also correspondingly puts forward the work plan of enhancing pollutant emission reduction and continuously promoting desulfurization and denitrification in the electric power industry, the newly-built coal-fired unit is required to comprehensively implement desulfurization and denitrification to realize standard emission, the existing coal-fired unit without the desulfurization and denitrification facility needs to be matched with the construction of a flue gas desulfurization and denitrification facility, and the unit which cannot stably achieve the standard emission needs to be implemented and transformed.
The common process of the desulfurization system is limestone-gypsum wet flue gas desulfurization, the whole process mainly comprises an absorption tower system, a limestone slurry preparation system and a gypsum dehydration treatment system, the absorption tower is provided with two outlets, one outlet is a gypsum slurry outlet, and the real-time PH value is detected, namely the PH value of the gypsum slurry outlet is detected; the other outlet is a clean flue gas outlet, and the concentration of the second oxidation building is detected, namely the concentration of sulfur dioxide at the clean flue gas outlet is detected; however, the flue gas reaction of the absorption tower is a large-lag slow dynamic process, and meanwhile, the desulfurization system is a complex control system, the conventional PID control strategy sets the pH value or adjusts the slurry spraying amount according to experience, the limestone slurry spraying amount is difficult to accurately control, and the slurry pH value is difficult to effectively control and ensure that the value is in an effective range.
The common process of the denitration system is SCR denitration, the main influencing factor is ammonia water amount, the reaction is incomplete due to too little ammonia water amount, the concentration of nitrogen oxide at an outlet exceeds the standard, the ammonia water amount is too much, and unreacted ammonia water can be discharged out of the system along with flue gas, so that the atmospheric pollution and the blockage of downstream equipment are caused; the high-efficient steady operation of deNOx systems realizes the key that the flue gas nitrogen oxide emission concentration of power plant is up to standard, because deNOx systems has characteristics such as nonlinearity, hysteresis quality, traditional PID control technique is difficult to maintain the deNOx systems export nitrogen oxide concentration stable, produces too big export nitrogen oxide concentration fluctuation, leads to frequently appearing export nitrogen oxide concentration phenomenon that exceeds standard on the one hand, and on the other hand is for guaranteeing export nitrogen oxide concentration standard rate, needs the increase to spout ammonia volume, and the deNOx systems spouts ammonia volume cost and improves thereupon.
How to solve the control difficult problem that the SOx/NOx control system hysteresis characteristic of power plant brought, guarantee SOx/NOx control outlet nitrogen oxide concentration effective control, reduce SOx/NOx control outlet nitrogen oxide concentration fluctuation, the accurate control thick liquid volume of spraying, it is the important direction of SOx/NOx control system operation regulation and control of power plant to reduce the deNOx control system and spout the ammonia cost.
Based on the technical problems, a new flue gas desulfurization and denitration optimization control method based on a hysteresis model needs to be designed.
Disclosure of Invention
The invention aims to solve the technical problem of providing a flue gas desulfurization and denitrification optimization control method based on a hysteresis model, which can truly simulate the actual working scene of desulfurization and denitrification of a power plant, and can visually display and acquire the actual operation condition of desulfurization and denitrification of the power plant and adjust the system parameters.
The technical scheme adopted by the invention is as follows:
a flue gas desulfurization and denitrification optimization control method based on a hysteresis model comprises the following steps: step 1: carrying out model construction on all parts for desulfurization and denitrification of the power plant, building a control system according to a field control strategy, and building a complete power plant desulfurization and denitrification system simulation model; step 2: establishing a first lag time prediction model for the time delay existing in the determination of the PH value of the absorption tower in the power plant desulfurization system, and establishing a second lag time prediction model for the time delay existing in the determination of the concentration of the nitrogen oxides in the flue gas at the inlet of the power plant denitrification system through a flue gas on-line monitoring device (CEMS); and step 3: predicting the concentrations of sulfur dioxide in the flue gas at the outlet of the power plant desulfurization system and nitrogen oxide in the flue gas at the inlet of the denitration system; and 4, step 4: controlling the slurry spraying amount of the desulfurization system according to the predicted value of the sulfur dioxide concentration and the combination of a first lag time prediction model, and controlling the ammonia spraying amount of the denitrification system according to the predicted value of the nitrogen oxide concentration and the combination of a second lag time prediction model; and 5: and issuing the control parameters of the slurry spraying amount and the ammonia spraying amount to a power plant desulfurization and denitrification system simulation model for intelligent diagnosis.
Further, in step 1, after the dynamic simulation software is used for building models of all the desulfurization and denitrification components of the power plant according to the modular modeling method, a corresponding control system is built according to a field control strategy, and a complete power plant desulfurization and denitrification system simulation model is built, and the method specifically comprises the following steps:
the power plant desulfurization system selects a limestone-gypsum wet desulfurization system, which at least comprises a flue gas system, an absorption tower system, a limestone slurry preparation system, a gypsum slurry dehydration system, a wastewater treatment system and an electrical system; the power plant denitration system is an SCR (selective catalytic reduction) flue gas denitration system, and at least comprises a flue gas system, an SCR reactor system, an acoustic wave soot blowing system and a liquid ammonia storage and supply system;
according to the mass conservation, momentum conservation and energy conservation equations, in the modeling process, the dynamic simulation software selects corresponding component modules from the model library according to the process flows of the limestone-gypsum wet desulphurization system and the SCR flue gas denitration system, connects the component modules, inputs initial data and completes the model construction of the power plant desulphurization and denitration system;
and (3) building an analog quantity control system, a sequence control system and a logic control system according to a field control strategy, and configuring by adopting a basic algorithm module to realize the same function as the actual control system and build a complete power plant desulfurization and denitrification system simulation model.
Further, the power plant SOx/NOx control system simulation model still includes:
in the process of model development and debugging, physical data acquired by an actual power plant desulfurization and denitrification system is compared with virtual data acquired based on a power plant desulfurization and denitrification simulation model, whether an error exceeds a threshold value is judged, if the error exceeds the threshold value, the virtual data with larger error is classified through cluster learning, corresponding historical data is used as input, error learning is carried out through a neural network, a correction coefficient is output to correct the error data of the virtual data, and virtual-real fusion is carried out on the corrected virtual data and the physical data to generate a verified power plant desulfurization and denitrification simulation model.
Further, in step 2, establishing a first lag time prediction model for the time delay of measuring the existence of the PH value of the absorption tower in the power plant desulfurization system by using a variable point detection model, a time window sliding model, a correlation analysis model and a machine learning model comprises: establishing a slurry pH value response lag time identification algorithm flow by adopting a variable point detection, time window sliding and correlation analysis method and establishing a first lag time prediction model by adopting a machine learning model;
the flow of the slurry pH value response lag time identification algorithm comprises the following steps:
selecting a working condition that the concentration value of sulfur dioxide at the outlet of the absorption tower changes after the pH value of the slurry of the absorption tower is adjusted as an identification object;
the time window deltat is equally divided into two equally spaced timesWindow Δ ti1And Δ ti2Gradually sliding forwards on a time axis, calculating the average difference value of the sulfur dioxide concentration in two time windows, and if the average difference value exceeds a set threshold value, taking the moment as a working condition change point tiIf the time window is smaller than the set threshold, continuing to slide the time window forwards until a working condition change point is detected or the time window slides to a cut-off time point;
based on the operating condition change point tiAnd a time window delta t, respectively acquiring a slurry PH value time sequence and a sulfur dioxide concentration time sequence from the start of working condition change to the end of the time window;
gradually advancing the sulfur dioxide concentration time sequence, setting a maximum moving step number k, obtaining a new sulfur dioxide concentration sequence through the advancement and constructing a sulfur dioxide concentration time lag matrix V;
calculating the PH value time sequence of the slurry and the Pearson correlation coefficient r of each column in the matrix V, wherein the delay time corresponding to the maximum correlation coefficient is the PH value response lag time t under the working condition1;
Establishing a first lag time prediction model using a machine learning model comprises:
acquiring original data characteristics in a power plant desulfurization system, preprocessing the original data characteristics, substituting the preprocessed data characteristics into the flow of the slurry PH value response lag time identification algorithm to carry out PH value delay identification, and acquiring the relationship between delay time and different operation data characteristics; the raw data features include at least: the load capacity of the boiler, the air supply quantity of the boiler, the flow rate of limestone slurry, the input quantity of the limestone slurry, the sulfur dioxide content, the calcium carbonate content in limestone and the distance data characteristics from the absorption tower to a PH measuring point;
the operation data characteristics which are obtained by identification and can cause the PH value change are converted into characteristics with more working condition characteristics in a characteristic conversion mode, the correlation among the data characteristics is reduced, and the influence caused by dimension is eliminated by carrying out normalization processing on the converted data characteristics;
performing correlation analysis on the original operation data characteristics by adopting a correlation analysis method to obtain correlation coefficients of each operation data characteristic and PH value response lag time, wherein the higher the correlation coefficient is, the most correlation between the data characteristic and the lag time is shown;
fusing the operating data characteristics by adopting a characteristic fusion method according to the height of the correlation coefficient to form new fusion characteristics, taking the original operating data characteristics and the new fusion characteristics as sample data, and inputting a training set in the sample data into a machine learning model according to a preset proportion to establish a first lag time prediction model under different operating data change conditions; and calculating the PH value response lag time according to different operation data characteristics through the first lag time prediction model.
Further, in step 2, establishing a second lag time prediction model for the time delay of the flue gas nitrogen oxide concentration determination of the inlet of the power plant denitration system through the flue gas online monitoring device CEMS by adopting a variable point detection model, a time window sliding model, a correlation analysis model and a machine learning model comprises the following steps: establishing a CEMS determination lag time identification algorithm flow by adopting a variable point detection method, a time window sliding method and a correlation analysis method, and establishing a second lag time prediction model by adopting a machine learning model;
the CEMS determination lag time identification algorithm flow comprises the following steps:
selecting a working condition that the CEMS measured value changes after the concentration of the inlet nitrogen oxide changes as an identification object; the measurement of the concentration of the nitrogen oxides passes through the heat tracing guide pipe and the analysis cabinet, and the flow of the flue gas in the heat tracing guide pipe and the concentration measurement in the analysis cabinet have certain time lag;
the time window Δ t' is equally divided into two equally spaced time windows Δ ti1' and Δ ti2Gradually sliding forward on a time axis, calculating the average difference value of the CEMS measured values in two time windows, and if the average difference value exceeds a set threshold value, determining the moment as a working condition change point tiIf the time window is smaller than the set threshold, continuing to slide the time window forwards until a working condition change point is detected or the time window slides to a cut-off time point;
based on the operating condition change point ti'and a time window delta t' are respectively obtained, wherein a nitrogen oxide concentration value time sequence and a CEMS measurement value time sequence from the start of working condition change to the end of the time window are obtained;
gradually advancing the CEMS measured value time sequence, setting the maximum moving step number k, obtaining a new CEMS measured value sequence through the advancing and constructing a CEMS measured value time lag matrix V';
calculating the time sequence of the concentration value of the nitrogen oxide and the Pearson correlation coefficient r 'of each column in the matrix V', wherein the delay time corresponding to the maximum correlation coefficient is the lag time t of the measurement of the concentration of the nitrogen oxide under the working condition2;
Establishing a second lag time prediction model using a machine learning model comprises:
collecting original data characteristics in a denitration system of a power plant, preprocessing the original data characteristics, substituting the preprocessed data characteristics into a CEMS (continuous emission monitoring system) determination lag time identification algorithm flow for delay identification, and acquiring the relation between delay time and different operation data characteristics; the raw data features include at least: boiler load, coal type, coal feeding amount, combustion temperature, air volume and flue gas volume;
the method comprises the steps of converting the identified operating data characteristics capable of causing the concentration change of the nitrogen oxides into characteristics with more working condition characteristics in a characteristic conversion mode, reducing the correlation among the data characteristics, and eliminating the influence caused by dimension by carrying out normalization processing on the converted data characteristics;
performing correlation analysis on the original operation data characteristics by adopting a correlation analysis method to obtain correlation coefficients of each operation data characteristic and the determination lag time of the concentration value of the nitric oxide, wherein the higher the correlation coefficient is, the most correlation between the data characteristics and the lag time is shown;
fusing the operating data characteristics by adopting a characteristic fusion method according to the height of the correlation coefficient to form new fusion characteristics, taking the original operating data characteristics and the new fusion characteristics as sample data, and inputting a training set in the sample data into a machine learning model according to a preset proportion to establish second lag time prediction models under different operating data change conditions; and calculating the determined lag time of the CEMS according to different operating data characteristics by using a second lag time prediction model.
Furthermore, the XGboost model is selected as the machine learning model, and is an integrated learning algorithm adopting a boosting method, namely a base learnerSelecting a CART decision tree, applying k CART functions { f }1,f2,…,fkAdding to form an integrated tree model; the target function of the model consists of a loss function and a regular term, and the loss function is approximated by second-order Taylor expansion; optimizing key parameters to improve the accuracy of model prediction, wherein the key parameters comprise the maximum depth of a tree, subsamples, the column number ratio of random sampling of each tree, the minimum leaf node sample weight and the learning rate;
the method comprises the steps of constructing a model, starting from a root node, sequencing training set data according to each data feature, calculating the profit of each feature by adopting a greedy method, selecting the feature with the maximum profit as a splitting feature, mapping the training set data to corresponding leaf nodes, performing recursion on the generated leaf nodes until a limiting condition is reached, finishing the generation process of a decision tree, calculating the weight of the leaf nodes of the decision tree by using first-order and second-order derivatives of a loss function, taking the weight as a fitting target of the next tree, performing recursion repeatedly until the condition is met, and finishing the establishment of the model.
Further, in step 3, after collecting the historical operating parameters of the desulfurization system and the denitration system of the power plant, selecting the operating parameters strongly related to the desulfurization of the power plant and inputting the operating parameters into the constructed first support vector machine model to predict the concentration of sulfur dioxide in the flue gas at the outlet of the desulfurization system of the power plant, and specifically comprising the following steps:
taking the collected historical operating parameters of the power plant desulfurization system as sample data, performing correlation analysis on the sample data, removing the sample data of which the correlation with the sulfur dioxide concentration in the flue gas at the outlet of the limestone-gypsum wet desulfurization system is smaller than a preset value, and taking the residual sample data as operating data strongly correlated with the desulfurization system; the historical operating parameters of the desulfurization system at least comprise sulfur dioxide concentration at an inlet, nitrogen oxide concentration, unit load, limestone slurry circulating pump current, slurry supply quantity, flue gas sulfur dioxide concentration at an outlet of an absorption tower and slurry pH value;
performing data preprocessing on the operation data strongly related to the desulfurization system, and constructing a first support vector machine model by using the preprocessed data;
collecting real-time operation data related to power plant desulfurization and inputting the real-time operation data into a constructed first support vector machine model to obtain a predicted value of the concentration of sulfur dioxide in flue gas at an outlet of a power plant desulfurization system;
wherein the data preprocessing comprises: filling missing values and abnormal values of the operation data strongly related to the desulfurization system and normalizing the operation data to obtain a pretreated desulfurization data sequence, and recording the pretreated desulfurization data sequence as F ═ F1,f2,f3,…,fn],fiThe desulfurization data of the ith time point in the desulfurization data sequence after treatment is obtained;
carrying out wavelet threshold denoising processing on the desulfurization data sequence F, carrying out wavelet decomposition on noisy data with noise to obtain real data information, and marking as P ═ P1,p2,p3,…,pm],piThe desulfurization data is the desulfurization data of the ith time point in the real desulfurization data sequence.
Further, in step 3, selecting an operation parameter strongly related to denitration of the power plant, inputting the operation parameter into a second support vector machine model constructed to predict the concentration of nitrogen oxide in flue gas at an inlet of a denitration system of the power plant, and specifically comprising the following steps:
the collected historical operating parameters of the power plant denitration system are used as sample data, correlation between the sample data and the concentration of nitrogen oxides in flue gas at an inlet of the power plant denitration system is calculated by adopting a Pearson correlation coefficient, and a data combination with high correlation is selected according to the correlation and is used as operating data strongly related to the denitration system; historical operating data of the denitration system at least comprises ammonia injection mass flow, boiler load, SCR inlet flue gas temperature, SCR inlet flue gas oxygen content, SCR inlet nitrogen oxide concentration and SCR denitration efficiency;
performing data preprocessing on operating data strongly related to the denitration system, and constructing a second support vector machine model by using the preprocessed data;
collecting real-time operation data related to power plant denitration and inputting the real-time operation data into a constructed second support vector machine model to obtain a predicted value of the concentration of nitrogen oxides in flue gas at an inlet of a denitration system of a power plant;
wherein, select strong correlation and extremely strong relevant data as the data that the correlation with deNOx systems is high, the computational formula is:
x is the input sample data characteristic, Y is the nitrogen oxide concentration at the inlet, cov (X, Y) represents the covariance of X, Y; sigmaXAnd σYIs the standard deviation of X and Y respectively, and rho represents the correlation coefficient between two variables, and the value range is [ -1,1](ii) a When rho is more than or equal to 0.8<1, referred to as very strong correlation; when rho is more than or equal to 0.6<0.8, called strong correlation; when rho is more than or equal to 0.4<At 0.6, it is said to be moderately correlated; when rho is more than or equal to 0.2<0.4, referred to as weak correlation; when rho is more than or equal to 0.0<At 0.2, it is referred to as very weakly correlated or uncorrelated.
Further, constructing the first support vector machine model and the second support vector machine model includes: determining the optimal support vector machine parameters by adopting a cuckoo optimization method: initializing parameters of a cuckoo optimization algorithm, and searching the position of a bird nest by step length self-adaptive and dynamically adjusted Laevice flight according to the parameters of the cuckoo optimization method:1,2, …, n; wherein x isi (t+1)The position of the ith bird nest in the tth generation; a is step control quantity, is used for controlling the search range of the step, and obeys positive space distribution; l (lambda) is a Levy random walk path; the step self-adaptive dynamic adjustment strategy is as follows:
stepi=stepmin+(stepmax-stepmin)di
wherein stepiStep for the current search stepmaxStep being the maximum value of the step sizeminIs the minimum value of the step size, niIs the position of the ith bird nest, nbestThe current minimum fitness corresponds to the nest position of the nest, dmaxThe maximum value of the distance between the bird nest corresponding to the current minimum fitness and other bird nests is obtained;
training a support vector machine model by adopting a training set in the preprocessed data, calculating the fitness of each bird nest position, and keeping the bird nest corresponding to the minimum fitness to the next iteration;
judging whether the minimum fitness meets a preset termination condition, if so, determining the nest position of the nest corresponding to the minimum fitness to be the determined optimal support vector machine parameter, and if not, removing a plurality of nests with the highest fitness and readjusting the nest positions;
training a support vector machine model according to the determined optimal support vector machine parameters: establishing a kernel function-based support vector machine training program, forming a mapping relation between an input variable and an output variable through a support vector machine model, performing learning training on training sample data by using the support vector machine training program, and obtaining N support vectors X through learning and trainingi *I is 0,1, …, N, forming a support vector machine model:
wherein, Xi *Support vector, Y, representing a desulfurization or denitrification system of a power plantiSulfur dioxide concentration representing a support vector of a desulfurization system of a power plant or nitrogen oxide concentration, alpha, of a support vector of a denitrification systemiAnd the coefficient represents the ith support vector, X is input pre-processed desulfurization data or denitration data, Y (X) represents the predicted value of the concentration of sulfur dioxide of the support vector of the desulfurization system of the power plant or the predicted value of the concentration of nitrogen oxide of the support vector of the denitration system, K (·) represents the kernel function of the support vector machine, and the kernel function selects one of a Gaussian function, a polynomial function, a linear function and a radial basis function.
Further, step 5 includes in detail: after the slurry spraying amount control parameter, the ammonia spraying amount control parameter and the relevant configuration parameter of the operation of the power plant desulfurization and denitrification system are input into the power plant desulfurization and denitrification system simulation model, the obtained real-time operation parameter of the power plant desulfurization and denitrification system is compared with the simulation result data of the simulation model through the set expert diagnosis module to obtain a deviation, and pre-alarming is realized through whether the deviation exceeds a preset threshold value or not;
the expert diagnosis module is internally provided with an intelligent diagnosis strategy, judges related running states, data deviation and pre-alarm information conditions through preset logics, comprehensively outputs diagnosis preliminary result information, calls expert base knowledge information for comparison, analyzes whether conclusion information obtained by the intelligent diagnosis strategy is related or consistent with the expert base knowledge information, and outputs a diagnosis analysis result, running guidance or a task list; the knowledge information of the expert database comprises stored preset knowledge and information that an abnormal fault occurs; the pre-alarm comprises a parameter exceeding a preset threshold, time for the parameter exceeding the preset threshold and abnormal fault information.
The invention has the positive effects that:
(1) according to the method, dynamic simulation software is used for building models of all parts for desulfurization and denitrification of the power plant according to a modular modeling method, and building corresponding control systems according to a field control strategy, so that a complete simulation model of the desulfurization and denitrification system of the power plant is built, the actual working scene of desulfurization and denitrification of the power plant can be simulated really, the actual operation condition of desulfurization and denitrification of the power plant can be displayed and known visually, and system parameters can be adjusted;
(2) in the process of developing and debugging the model, physical data acquired by an actual power plant desulfurization and denitrification system is compared with virtual data acquired based on the power plant desulfurization and denitrification simulation model, whether an error exceeds a threshold value is judged, if the error exceeds the threshold value, the virtual data with larger error is classified through cluster learning, corresponding historical data is used as input, error learning is carried out through a neural network, a correction coefficient is output to correct error data of the virtual data, the corrected virtual data and the physical data are subjected to virtual-real fusion to generate a verified power plant desulfurization and denitrification simulation model, the error can be corrected through the neural network after the virtual-real data are compared and analyzed, the precision and the accuracy of the power plant desulfurization and denitrification simulation model are improved, and a foundation is laid for prediction control of a subsequent desulfurization and denitrification system;
(3) according to the method, a first lag time prediction model is established for the time delay existing in the process of determining the PH value of the absorption tower in the power plant desulfurization system by adopting a variable point detection model, a time window sliding model, a correlation analysis model and a machine learning model, a second lag time prediction model is established for the time delay existing in the process of determining the concentration of the nitrogen oxide in the flue gas at the inlet of the power plant denitrification system by the flue gas on-line monitoring device CEMS, the lag time existing in the power plant desulfurization and denitrification system can be analyzed, calculated and the prediction model is established, and the corresponding lag influence data characteristics and the lag time can be rapidly and effectively obtained;
(4) according to the method, after historical operating parameters of a power plant desulfurization system and a denitration system are collected, operating parameters strongly related to power plant desulfurization are selected and input into a constructed first support vector machine model to predict the concentration of sulfur dioxide in flue gas at an outlet of the power plant desulfurization system, operating parameters strongly related to power plant denitration are selected and input into a constructed second support vector machine model to predict the concentration of nitrogen oxide in flue gas at an inlet of the power plant denitration system, the concentration of sulfur dioxide in flue gas at the outlet of the desulfurization system can be predicted through the support vector machine model, the concentration of nitrogen oxide at the inlet of the denitration system can be predicted, and the accuracy of predicted values is improved;
(5) according to the method, the slurry spraying amount of a desulfurization system is controlled by combining a predicted value of the sulfur dioxide concentration with a first lag time prediction model, the ammonia spraying amount of a denitrification system is controlled by combining a predicted value of the nitrogen oxide concentration with a second lag time prediction model, the concentration of the nitrogen oxide at a desulfurization outlet can be effectively controlled by combining the first lag time, the slurry spraying amount is accurately controlled, the pH value of the slurry is controlled within an effective range, the fluctuation of the concentration of the nitrogen oxide at the denitrification outlet is reduced by combining the second lag time, the ammonia spraying amount is accurately controlled, and the ammonia spraying cost of the denitrification system is reduced;
(6) according to the invention, the control parameters of the slurry spraying amount and the ammonia spraying amount are issued to the power plant desulfurization and denitrification system simulation model for intelligent diagnosis, expert base knowledge information and an intelligent diagnosis strategy are set in an expert diagnosis module to compare the real-time operation parameters and simulation data of the system to realize alarming and diagnosis, and a diagnosis analysis result, an operation guide or a task list are output, so that the effective processing and diagnosis analysis of the power plant desulfurization and denitrification system data are realized.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a process flow diagram of a limestone-gypsum wet desulfurization system of the present invention;
FIG. 3 is a schematic view of an inlet CEMS device in a flue gas denitration system by an SCR method.
Detailed Description
As shown in fig. 1, this embodiment 1 provides a flue gas desulfurization and denitration optimization control method based on a hysteresis model, where the flue gas desulfurization and denitration optimization control method includes:
after the dynamic simulation software is used for modeling all the components for desulfurization and denitrification of the power plant according to a modular modeling method, a corresponding control system is built according to a field control strategy, and a complete simulation model for the desulfurization and denitrification system of the power plant is built; the power plant desulfurization system adopts a limestone-gypsum wet desulfurization system, and the power plant denitration system adopts an SCR flue gas denitration system;
respectively establishing a first lag time prediction model for the time delay of determining the existence of the PH value of an absorption tower in a power plant desulfurization system and establishing a second lag time prediction model for the time delay of determining the existence of the concentration of the nitrogen oxides in the flue gas at the inlet of a power plant denitrification system through a flue gas on-line monitoring system (CEMS) by adopting a variable point detection model, a time window sliding model, a correlation analysis model and a machine learning model;
after historical operating parameters of a power plant desulfurization system and a denitration system are collected, selecting operating parameters strongly related to power plant desulfurization and inputting the operating parameters into a constructed first support vector machine model to predict the concentration of sulfur dioxide in flue gas at an outlet of the power plant desulfurization system, and selecting operating parameters strongly related to power plant denitration and inputting the operating parameters into a constructed second support vector machine model to predict the concentration of nitrogen oxide in flue gas at an inlet of the power plant denitration system;
controlling the slurry spraying amount of a desulfurization system according to the predicted value of the sulfur dioxide concentration and the first lag time prediction model, and controlling the ammonia spraying amount of a denitrification system according to the predicted value of the nitrogen oxide concentration and the second lag time prediction model;
and issuing the control parameters of the slurry spraying amount and the ammonia spraying amount to a power plant desulfurization and denitrification system simulation model for model verification and intelligent diagnosis.
The slurry spraying amount of the desulfurization system is controlled by combining the predicted value of the sulfur dioxide concentration with a first lag time prediction model: the set value is the output target value, and in the actual system, the concentration of sulfur dioxide in the flue gas at the outlet is the output target value; the operable variable is the spraying amount of the slurry, the size of the operable variable can be adjusted by the controller to act on a target object, the output reaches a desired value, the influence of the pH value measurement delay is considered in the control process, and the pH value detection needs to be carried out in advance for a certain time so as to open the slurry control valve in advance to control the spraying amount of the slurry. And controlling the ammonia injection amount of the denitration system by combining a second lag time prediction model according to the nitrogen oxide concentration predicted value: the method comprises the steps of calculating a first ammonia injection amount of a current operation condition according to an inlet nitrogen oxide concentration predicted value, sending the first ammonia injection amount to a denitration system, measuring an actual measurement value of an outlet nitrogen oxide concentration, inputting the actual measurement value and an outlet nitrogen oxide concentration set value into a controller after deviation is carried out, obtaining a second ammonia injection amount, sending the second ammonia injection amount to the denitration system, controlling the ammonia injection amount of a denitration reactor by the denitration system according to the first ammonia injection amount and the second ammonia injection amount, considering the delay influence of nitrogen oxide concentration value measurement in the control process, and needing to measure the nitrogen oxide concentration value in advance for a certain time so as to control the ammonia injection amount in advance.
In this embodiment, after the dynamic simulation software is used to model each part of the power plant for desulfurization and denitrification according to the modular modeling method, and then a corresponding control system is built according to the field control strategy, a complete power plant desulfurization and denitrification system simulation model is built, which includes:
the power plant desulfurization system selects a limestone-gypsum wet desulfurization system, which at least comprises a flue gas system, an absorption tower system, a limestone slurry preparation system, a gypsum slurry dehydration system, a wastewater treatment system and an electrical system; the power plant denitration system is an SCR (selective catalytic reduction) flue gas denitration system, and at least comprises a flue gas system, an SCR reactor system, an acoustic wave soot blowing system and a liquid ammonia storage and supply system;
according to the mass conservation, momentum conservation and energy conservation equations, in the modeling process, the dynamic simulation software selects corresponding component modules from the model library according to the process flows of the limestone-gypsum wet desulphurization system and the SCR flue gas denitration system, connects the component modules, inputs initial data and completes the model construction of the power plant desulphurization and denitration system;
and (3) building an analog quantity control system, a sequence control system and a logic control system according to a field control strategy, and configuring by adopting a basic algorithm module to realize the same function as the actual control system and build a complete power plant desulfurization and denitrification system simulation model.
According to the method, dynamic simulation software is used for building models of all parts for desulfurization and denitrification of the power plant according to a modular modeling method, and building corresponding control systems according to a field control strategy, so that a complete simulation model of the desulfurization and denitrification system of the power plant is built, the actual working scene of desulfurization and denitrification of the power plant can be simulated really, the actual operation condition of desulfurization and denitrification of the power plant can be displayed and known visually, and system parameters can be adjusted.
As shown in fig. 2, it should be noted that the main process sequence of desulfurization and denitrification is denitrification-dedusting-desulfurization. The process flow of the limestone-gypsum wet desulphurization system adopted by the patent comprises the following steps: the raw flue gas that comes out from the boiler afterbody passes through deNOx systems and dry-type electrostatic precipitator back, enters into the absorption tower under the draught fan effect, and whole absorption tower collects absorption and oxidation as an organic whole, and upper portion is the absorption region, and the lower part is the oxidation region. The oxidation fan at the bottom of the tower continuously blows oxidation air, the slurry circulating pump continuously pumps limestone slurry to the upper spraying layer from the bottom of the absorption tower, flue gas entering the absorption tower is in reverse contact with the limestone slurry sprayed from the upper part, and after mist carried in the gas is removed by a demister at the upper part of the absorption tower, the fully reacted clean flue gas is directly discharged into the inner space of the Haylor air cooling system from the top of the absorption tower and finally discharged into the atmosphere along with water vapor in the cooling tower. The density of the gypsum slurry in the slurry of the absorption tower is continuously increased along with the reaction, when the density reaches a certain value, a gypsum discharge pump is started to send the gypsum slurry to a gypsum dehydration system, byproduct gypsum is formed after dehydration, and the residual slurry returns to the system for recycling, so that the utilization rate of the desulfurization absorbent is improved.
In the technological process of a limestone-gypsum wet flue gas desulfurization system, the main processes are as follows: the sulfur dioxide and limestone slurry are subjected to chemical reactions such as dissolution, oxidation and the like to obtain a byproduct gypsum, and the processes are the main processes of the whole process flow.
The SCR method flue gas denitration process flow comprises the following steps: liquid ammonia tank car is sent into the liquid ammonia storage tank through the compressor of unloading with liquid ammonia, and liquid ammonia in the storage tank enters into liquid ammonia evaporator through self pressure to evaporate into the ammonia through the water-bath heating, and then is sent into SCR reaction zone after getting into in the ammonia buffer tank steady voltage. Before entering the SCR reactor, the air and ammonia gas sent by the dilution fan are uniformly mixed and then introduced into the SCR reactor to participate in chemical reaction.
In this embodiment, the power plant desulfurization and denitrification system simulation model further includes:
in the process of model development and debugging, physical data acquired by an actual power plant desulfurization and denitrification system is compared with virtual data acquired based on a power plant desulfurization and denitrification simulation model, whether an error exceeds a threshold value is judged, if the error exceeds the threshold value, the virtual data with larger error is classified through cluster learning, corresponding historical data is used as input, error learning is carried out through a neural network, a correction coefficient is output to correct the error data of the virtual data, and virtual-real fusion is carried out on the corrected virtual data and the physical data to generate a verified power plant desulfurization and denitrification simulation model.
In the process of developing and debugging the model, after comparing and analyzing the virtual data and the real data, the error is corrected by adopting the neural network, so that the precision and the accuracy of the power plant desulfurization and denitrification simulation model are improved, and a foundation is laid for the prediction control of a subsequent desulfurization and denitrification system.
In this embodiment, the establishing a first lag time prediction model for a time delay of determining the existence of the PH value of the absorption tower in the power plant desulfurization system by using the variable point detection, the time window sliding, the correlation analysis and the machine learning model includes: establishing a slurry pH value response lag time identification algorithm flow by adopting a variable point detection, time window sliding and correlation analysis method and establishing a first lag time prediction model by adopting a machine learning model;
the flow of the slurry pH value response lag time identification algorithm comprises the following steps:
selecting a working condition that the concentration value of sulfur dioxide at the outlet of the absorption tower changes after the pH value of the slurry of the absorption tower is adjusted as an identification object;
the time window Δ t is equally divided into two equally spaced time windows Δ ti1And Δ ti2Gradually sliding forwards on a time axis, calculating the average difference value of the sulfur dioxide concentration in two time windows, and if the average difference value exceeds a set threshold value, taking the moment as a working condition change point tiIf the time window is smaller than the set threshold, continuing to slide the time window forwards until a working condition change point is detected or the time window slides to a cut-off time point;
based on the operating condition change point tiAnd a time window delta t, respectively acquiring a slurry PH value time sequence and a sulfur dioxide concentration time sequence from the start of working condition change to the end of the time window;
gradually advancing the sulfur dioxide concentration time sequence, setting a maximum moving step number k, obtaining a new sulfur dioxide concentration sequence through the advancement and constructing a sulfur dioxide concentration time lag matrix V;
calculating the PH value time sequence of the slurry and the Pearson correlation coefficient r of each column in the matrix V, wherein the delay time corresponding to the maximum correlation coefficient is the PH value response lag time t under the working condition1;
Establishing a first lag time prediction model using a machine learning model comprises:
acquiring original data characteristics in a power plant desulfurization system, preprocessing the original data characteristics, substituting the preprocessed data characteristics into the flow of the slurry PH value response lag time identification algorithm to carry out PH value delay identification, and acquiring the relationship between delay time and different operation data characteristics; the raw data features include at least: the load capacity of the boiler, the air supply quantity of the boiler, the flow rate of limestone slurry, the input quantity of the limestone slurry, the sulfur dioxide content, the calcium carbonate content in limestone and the distance data characteristics from the absorption tower to a PH measuring point;
the operation data characteristics which are obtained by identification and can cause the PH value change are converted into characteristics with more working condition characteristics in a characteristic conversion mode, the correlation among the data characteristics is reduced, and the influence caused by dimension is eliminated by carrying out normalization processing on the converted data characteristics;
performing correlation analysis on the original operation data characteristics by adopting a correlation analysis method to obtain correlation coefficients of each operation data characteristic and PH value response lag time, wherein the higher the correlation coefficient is, the most correlation between the data characteristic and the lag time is shown;
fusing the operating data characteristics by adopting a characteristic fusion method according to the height of the correlation coefficient to form new fusion characteristics, taking the original operating data characteristics and the new fusion characteristics as sample data, and inputting a training set in the sample data into a machine learning model according to a preset proportion to establish a first lag time prediction model under different operating data change conditions; and calculating the PH value response lag time according to different operation data characteristics through the first lag time prediction model.
It should be noted that, in the desulfurization system, due to the installation position of the PH measuring element, the pure delay of the PH detection is caused by the time required for the detection by the dedicated PH detector and the reaction time of the limestone slurry and the sulfur dioxide, and the pure delay causes the time delay of the PH value measured by the electrode of the PH detector because the measurement signal cannot reflect the change of the PH value of the absorption liquid in the absorption tower in time.
In this embodiment, the establishing a second lag time prediction model for the time delay of the flue gas nitrogen oxide concentration at the inlet of the denitration system of the power plant through the flue gas online monitoring system CEMS by using the variable point detection, the time window sliding, the correlation analysis and the machine learning model includes: establishing a CEMS determination lag time identification algorithm flow by adopting a variable point detection method, a time window sliding method and a correlation analysis method, and establishing a second lag time prediction model by adopting a machine learning model;
the CEMS determination lag time identification algorithm flow comprises the following steps:
selecting a working condition that the CEMS measured value changes after the concentration of the inlet nitrogen oxide changes as an identification object; the measurement of the concentration of the nitrogen oxides passes through the heat tracing guide pipe and the analysis cabinet, and the flowing of the flue gas in the heat tracing guide pipe and the concentration measurement in the analysis cabinet have certain time lag;
the time window Δ t' is equally divided into two equally spaced time windows Δ ti1' and Δ ti2Gradually sliding forward on a time axis, calculating the average difference value of the CEMS measured values in two time windows, and if the average difference value exceeds a set threshold value, determining the moment as a working condition change point tiIf the time window is smaller than the set threshold, continuing to slide the time window forwards until a working condition change point is detected or the time window slides to a cut-off time point;
based on the operating condition change point ti'and a time window delta t' are respectively obtained, wherein a nitrogen oxide concentration value time sequence and a CEMS measurement value time sequence from the start of working condition change to the end of the time window are obtained;
gradually advancing the CEMS measured value time sequence, setting the maximum moving step number k, obtaining a new CEMS measured value sequence through the advancing and constructing a CEMS measured value time lag matrix V';
calculating the time sequence of the concentration value of the nitrogen oxide and the Pearson correlation coefficient r 'of each column in the matrix V', wherein the delay time corresponding to the maximum correlation coefficient is the lag time t of the measurement of the concentration of the nitrogen oxide under the working condition2;
Establishing a second lag time prediction model using a machine learning model comprises:
acquiring original data characteristics in a denitration system of a power plant, preprocessing the original data characteristics, substituting the preprocessed data characteristics into the CEMS determination lag time identification algorithm flow to perform delay identification, and acquiring the relation between delay time and different operation data characteristics; the raw data features include at least: boiler load, coal type, coal feeding amount, combustion temperature, air volume and flue gas volume;
the method comprises the steps of converting the identified operating data characteristics capable of causing the concentration change of the nitrogen oxides into characteristics with more working condition characteristics in a characteristic conversion mode, reducing the correlation among the data characteristics, and eliminating the influence caused by dimension by carrying out normalization processing on the converted data characteristics;
performing correlation analysis on the original operation data characteristics by adopting a correlation analysis method to obtain correlation coefficients of each operation data characteristic and the determination lag time of the concentration value of the nitric oxide, wherein the higher the correlation coefficient is, the most correlation between the data characteristics and the lag time is shown;
fusing the operating data characteristics by adopting a characteristic fusion method according to the height of the correlation coefficient to form new fusion characteristics, taking the original operating data characteristics and the new fusion characteristics as sample data, and inputting a training set in the sample data into a machine learning model according to a preset proportion to establish second lag time prediction models under different operating data change conditions; and calculating the determined lag time of the CEMS according to different operating data characteristics by the second lag time prediction model.
According to the method, a first lag time prediction model is established for the time delay existing in the process of determining the PH value of the absorption tower in the power plant desulfurization system by adopting a variable point detection model, a time window sliding model, a correlation analysis model and a machine learning model, a second lag time prediction model is established for the time delay existing in the process of determining the concentration of the nitrogen oxide in the flue gas at the inlet of the power plant denitrification system by the flue gas on-line monitoring device CEMS, the lag time existing in the power plant desulfurization and denitrification system can be analyzed, calculated and established, and the corresponding lag influence data characteristics and the lag time can be rapidly and effectively obtained.
As shown in fig. 3, it should be noted that an online flue gas monitoring system (CEMS) is used for measuring the concentration of nitrogen oxides in the denitration system, and a measurement delay of a certain time length may exist in the CEMS measurement process. The CEMS flue gas on-line monitoring system extracts gas from a flue in a heat pipe extraction and sampling mode, and the gas is subjected to dust removal, heating and,Links such as heat preservation and the like are guided to a pretreatment system to remove particulate matters and H2O, corrosive gas's processing, carry flue gas analysis appearance at last, in this processing procedure, can see that the measurement of flue gas nitrogen oxide concentration will pass through heat tracing pipe and analysis cabinet, the flow of flue gas heat tracing pipe again and the measurement of concentration all need certain time in the analysis cabinet, thereby lead to CEMS to measure and to have certain time delay, this measurement time delay can exert an influence to the control of deNOx systems ammonia injection volume afterwards, thereby make unable timely response among the ammonia injection volume control process, the degree of difficulty of ammonia injection volume control has been increased, the fluctuation of deNOx systems export nitrogen oxide concentration also can be bigger, when the change of entry nitrogen oxide concentration is more rapid, measurement lag error is bigger.
It should be noted that, the data with strong and weak correlation in the original data can be found through the data correlation strength analysis and calculation, however, the lower correlation coefficient cannot represent that there is no relation between the data feature and the output quantity, and only that the linear correlation degree of the data feature and the output quantity is not high, but there may be some non-linear correlations, and the fusion feature related to the output quantity can be studied through a data fusion mode.
The fusion method based on linear regression and multi-layer perceptron algorithm is essentially to construct a basic prediction model y (f) (x) and learn the prediction model y (f) (x)As a new fusion feature, if the mapping function f (-) is a linear (e.g., linear regression) function, the obtained fusion feature will also be a linear combination of the original features; if the mapping function f (-) is a non-linear (e.g., multi-layered perceptron) function, the resulting fused feature will be a non-linear combination of the original features.
Besides the linear fusion of the features, the nonlinear relation in the original features can be extracted through a multilayer perceptron, the feature data is input into a model to fit the concentration of nitrogen oxide, and then the nonlinear feature fusion feature F is obtainedMLP. The nodes of the hidden layer are set to be 100, the iteration times are set to be 200 generations, and the relu function is used as the activation function. For passing indexThe fusion approach combines features, the nature of which is different from the two above-described methods based on map fitting.
The method based on exponential fusion is to combine original features together in the form of exponential power to form new fusion features, such asWherein f is1,f2,…fkFor the original features to be screened, a, b, …, m are parameters to be determined for each feature. The fusion feature space is obtained by searching different parameter combinations in one parameter space. And calculating the correlation coefficient between each exponential fusion feature and the delay time in the fusion feature space, wherein the feature with the maximum correlation coefficient is used as the final fusion feature.
In this embodiment, the XGBoost model is selected as the machine learning model, and is an integrated learning algorithm using boosting method, and the base learner selects the CART decision tree and applies k CART functions { f ″1,f2,…,fkAdding to form an integrated tree model; the target function of the model consists of a loss function and a regular term, and the loss function is approximated by second-order Taylor expansion; optimizing key parameters to improve the accuracy of model prediction, wherein the key parameters comprise the maximum depth of a tree, subsamples, the column number ratio of random sampling of each tree, the minimum leaf node sample weight and the learning rate;
the method comprises the steps of constructing a model, starting from a root node, sequencing training set data according to each data feature, calculating the profit of each feature by adopting a greedy method, selecting the feature with the maximum profit as a splitting feature, mapping the training set data to corresponding leaf nodes, performing recursion on the generated leaf nodes until a limiting condition is reached, finishing the generation process of a decision tree, calculating the weight of the leaf nodes of the decision tree by using first-order and second-order derivatives of a loss function, taking the weight as a fitting target of the next tree, performing recursion repeatedly until the condition is met, and finishing the establishment of the model.
In this embodiment, after acquiring the historical operating parameters of the desulfurization system and the denitration system of the power plant, selecting the operating parameters strongly related to the desulfurization of the power plant and inputting the selected operating parameters into the constructed first support vector machine model to predict the sulfur dioxide concentration in the flue gas at the outlet of the desulfurization system of the power plant, specifically including:
taking the collected historical operating parameters of the power plant desulfurization system as sample data, performing correlation analysis on the sample data, removing the sample data of which the correlation with the sulfur dioxide concentration in the flue gas at the outlet of the limestone-gypsum wet desulfurization system is less than a preset value, and taking the residual sample data as operating data strongly correlated with the desulfurization system; the historical operating parameters of the desulfurization system at least comprise sulfur dioxide concentration at an inlet, nitrogen oxide concentration, unit load, limestone slurry circulating pump current, slurry supply quantity, flue gas sulfur dioxide concentration at an outlet of an absorption tower and slurry pH value;
performing data preprocessing on the operation data strongly related to the desulfurization system, and constructing a first support vector machine model by using the preprocessed data;
collecting real-time operation data related to power plant desulfurization and inputting the real-time operation data into the constructed first support vector machine model to obtain a predicted value of the concentration of sulfur dioxide in flue gas at an outlet of a desulfurization system of a power plant.
Wherein the data preprocessing comprises: filling missing values and abnormal values of the operation data strongly related to the desulfurization system and normalizing the operation data to obtain a pretreated desulfurization data sequence, and recording the pretreated desulfurization data sequence as F ═ F1,f2,f3,…,fn],fiThe desulfurization data of the ith time point in the desulfurization data sequence after treatment is obtained;
carrying out wavelet threshold denoising processing on the desulfurization data sequence F, carrying out wavelet decomposition on noisy data with noise to obtain real data information, and marking as P ═ P1,p2,p3,…,pm],piThe desulfurization data is the desulfurization data of the ith time point in the real desulfurization data sequence.
In this embodiment, the selecting of the operating parameter strongly related to the denitration of the power plant and inputting the operating parameter into the second support vector machine model to predict the concentration of nitrogen oxide in flue gas at the inlet of the denitration system of the power plant specifically includes:
the collected historical operating parameters of the power plant denitration system are used as sample data, correlation between the sample data and the concentration of nitrogen oxides in flue gas at an inlet of the power plant denitration system is calculated by adopting a Pearson correlation coefficient, and a data combination with high correlation is selected according to the correlation to be used as operating data strongly related to the denitration system; the historical operating data of the denitration system at least comprises ammonia spraying mass flow, boiler load, SCR inlet flue gas temperature, SCR inlet flue gas oxygen content, SCR inlet nitrogen oxide concentration and SCR denitration efficiency;
performing data preprocessing on the operating data strongly related to the denitration system, and constructing a second support vector machine model by using the preprocessed data;
collecting real-time operation data related to power plant denitration and inputting the real-time operation data into a constructed second support vector machine model to obtain a predicted value of the concentration of nitrogen oxides in flue gas at an inlet of a denitration system of a power plant;
wherein, select strong correlation and extremely strong relevant data as the data that the correlation with deNOx systems is high, the computational formula is:
x is the input sample data characteristic, Y is the nitrogen oxide concentration at the inlet, cov (X, Y) represents the covariance of X, Y; sigmaXAnd σYIs the standard deviation of X and Y respectively, and rho represents the correlation coefficient between two variables, and the value range is [ -1,1](ii) a When rho is more than or equal to 0.8<1, referred to as very strong correlation; when rho is more than or equal to 0.6<0.8, called strong correlation; when rho is more than or equal to 0.4<At 0.6, it is said to be moderately correlated; when rho is more than or equal to 0.2<0.4, referred to as weak correlation; when rho is more than or equal to 0.0<At 0.2, it is referred to as very weakly correlated or uncorrelated.
It should be noted that the data preprocessing of the operation data strongly related to the denitration system includes: removing repeated data, time completion and resampling, missing value filling, abnormal value replacement processing and the like.
In this embodiment, the constructing the first support vector machine model and the second support vector machine model includes:
determining the optimal support vector machine parameters by adopting a cuckoo optimization method: initializing parameters of a cuckoo optimization algorithm, and searching the position of a bird nest by step length self-adaptive and dynamically adjusted Laevice flight according to the parameters of the cuckoo optimization method:1,2, …, n; wherein x isi (t+1)The position of the ith bird nest in the tth generation; a is step control quantity, is used for controlling the search range of the step, and obeys positive space distribution; l (lambda) is a Levy random walk path; the step self-adaptive dynamic adjustment strategy is as follows:
stepi=stepmin+(stepmax-stepmin)di
wherein stepiStep for the current search stepmaxStep being the maximum value of the step sizeminIs the minimum value of the step size, niIs the position of the ith bird nest, nbestThe current minimum fitness corresponds to the nest position of the nest, dmaxThe maximum value of the distance between the bird nest corresponding to the current minimum fitness and other bird nests is obtained;
training a support vector machine model by adopting a training set in the preprocessed data, calculating the fitness of each bird nest position, and keeping the bird nest corresponding to the minimum fitness to the next iteration;
judging whether the minimum fitness meets a preset termination condition, if so, determining the nest position of the nest corresponding to the minimum fitness to be the determined optimal support vector machine parameter, and if not, removing a plurality of nests with the highest fitness and readjusting the nest positions;
training a support vector machine model according to the determined optimal support vector machine parameters: establishing a kernel function-based support vector machine training program, and forming a mapping relation by an input variable and an output variable through a support vector machine modelThe system utilizes a training program of a support vector machine to carry out learning training on training sample data, and N support vectors X are obtained through learning and trainingi *I is 0,1, …, N, forming a support vector machine model:
wherein, Xi *Support vector, Y, representing a desulfurization or denitrification system of a power plantiSulfur dioxide concentration representing a support vector of a desulfurization system of a power plant or nitrogen oxide concentration, alpha, of a support vector of a denitrification systemiAnd the coefficient represents the ith support vector, X is input pre-processed desulfurization data or denitration data, Y (X) represents the predicted value of the concentration of sulfur dioxide of the support vector of the desulfurization system of the power plant or the predicted value of the concentration of nitrogen oxide of the support vector of the denitration system, K (·) represents the kernel function of the support vector machine, and the kernel function selects one of a Gaussian function, a polynomial function, a linear function and a radial basis function.
According to the method, after historical operating parameters of a power plant desulfurization system and a denitration system are collected, operating parameters strongly related to power plant desulfurization are selected and input into a constructed first support vector machine model to predict the concentration of sulfur dioxide in flue gas at an outlet of the power plant desulfurization system, operating parameters strongly related to power plant denitration are selected and input into a constructed second support vector machine model to predict the concentration of nitrogen oxide in flue gas at an inlet of the power plant denitration system, the concentration of sulfur dioxide in flue gas at an outlet of the desulfurization system can be predicted through the support vector machine model, the concentration of nitrogen oxide at an inlet of the denitration system can be predicted, and accuracy of predicted values is improved.
In practical application, the correctness of the predicted sulfur dioxide concentration in the flue gas at the outlet and the predicted nitrogen oxide concentration in the flue gas at the inlet is evaluated through the root mean square error RMSE and the average absolute percentage error MAPE, and the calculation formula is as follows:
wherein, yiIs the actual value of the concentration of sulfur dioxide at the outlet or the concentration of nitrogen oxide in the flue gas at the inlet,in order to predict the predicted value of the concentration of the sulfur dioxide at the outlet or the concentration of the nitrogen oxide in the flue gas at the inlet, n is the number of prediction samples, and the smaller the RMSE and MAPE values are, the closer the predicted value of the concentration of the sulfur dioxide at the outlet or the predicted value of the concentration of the nitrogen oxide in the flue gas at the inlet is to the real value, the higher the precision is.
In this embodiment, the control parameters of the slurry spraying amount and the ammonia spraying amount are issued to the power plant desulfurization and denitrification system simulation model for model verification and intelligent diagnosis, including:
after the slurry spraying amount control parameter, the ammonia spraying amount control parameter and the relevant configuration parameter of the operation of the power plant desulfurization and denitrification system are input into the power plant desulfurization and denitrification system simulation model, the obtained real-time operation parameter of the power plant desulfurization and denitrification system is compared with the simulation result data of the simulation model through the set expert diagnosis module to obtain a deviation, and pre-alarming is realized through whether the deviation exceeds a preset threshold value or not;
the expert diagnosis module is internally provided with an intelligent diagnosis strategy, judges related running states, data deviation and pre-alarm information conditions through preset logics, comprehensively outputs diagnosis preliminary result information, calls expert base knowledge information for comparison, analyzes whether conclusion information obtained by the intelligent diagnosis strategy is related or consistent with the expert base knowledge information, and outputs a diagnosis analysis result, running instructions or a task list; the knowledge information of the expert database comprises stored preset knowledge and information that an abnormal fault occurs; the pre-alarm comprises a parameter exceeding a preset threshold, time for the parameter exceeding the preset threshold and abnormal fault information.
According to the invention, the slurry spraying amount of the desulfurization system is controlled by combining the predicted value of the sulfur dioxide concentration with the first lag time prediction model, the ammonia spraying amount of the denitrification system is controlled by combining the predicted value of the nitrogen oxide concentration with the second lag time prediction model, the concentration of the nitrogen oxide at the desulfurization outlet can be effectively controlled by combining the first lag time, the slurry spraying amount is accurately controlled, the pH value of the slurry is controlled within an effective range, the fluctuation of the concentration of the nitrogen oxide at the denitrification outlet is reduced by combining the second lag time, the ammonia spraying amount is accurately controlled, and the ammonia spraying cost of the denitrification system is reduced.
According to the invention, the control parameters of the slurry spraying amount and the ammonia spraying amount are issued to the power plant desulfurization and denitrification system simulation model for intelligent diagnosis, expert base knowledge information and an intelligent diagnosis strategy are set in an expert diagnosis module to compare the real-time operation parameters and simulation data of the system to realize alarming and diagnosis, and a diagnosis analysis result, an operation guide or a task list are output, so that the effective processing and diagnosis analysis of the power plant desulfurization and denitrification system data are realized.
In the several embodiments provided in the present application, it should be understood that the disclosed system and method may be implemented in other ways. The system embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, the functional modules in the embodiments of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In light of the foregoing description of the preferred embodiment of the present invention, many modifications and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The technical scope of the present invention is not limited to the content of the specification, and must be determined according to the scope of the claims.
Claims (10)
1. A flue gas desulfurization and denitrification optimization control method based on a hysteresis model is characterized by comprising the following steps: step 1: carrying out model construction on all parts for desulfurization and denitrification of the power plant, building a control system according to a field control strategy, and building a complete power plant desulfurization and denitrification system simulation model; step 2: establishing a first lag time prediction model for the time delay existing in the determination of the PH value of the absorption tower in the power plant desulfurization system, and establishing a second lag time prediction model for the time delay existing in the determination of the concentration of the nitrogen oxides in the flue gas at the inlet of the power plant denitrification system through a flue gas on-line monitoring device (CEMS); and step 3: predicting the concentrations of sulfur dioxide in the flue gas at the outlet of the power plant desulfurization system and nitrogen oxide in the flue gas at the inlet of the denitration system; and 4, step 4: controlling the slurry spraying amount of the desulfurization system according to the predicted value of the sulfur dioxide concentration and the combination of a first lag time prediction model, and controlling the ammonia spraying amount of the denitrification system according to the predicted value of the nitrogen oxide concentration and the combination of a second lag time prediction model; and 5: and issuing the control parameters of the slurry spraying amount and the ammonia spraying amount to a power plant desulfurization and denitrification system simulation model for intelligent diagnosis.
2. The optimal control method for flue gas desulfurization and denitrification based on the hysteresis model as claimed in claim 1, wherein in step 1, dynamic simulation software is used for building a model of each component for desulfurization and denitrification of the power plant according to a modular modeling method, and a corresponding control system is built according to a field control strategy to build a complete simulation model of the desulfurization and denitrification system of the power plant, which specifically comprises the following steps:
the power plant desulfurization system selects a limestone-gypsum wet desulfurization system, which at least comprises a flue gas system, an absorption tower system, a limestone slurry preparation system, a gypsum slurry dehydration system, a wastewater treatment system and an electrical system; the power plant denitration system is an SCR (selective catalytic reduction) flue gas denitration system, and at least comprises a flue gas system, an SCR reactor system, an acoustic wave soot blowing system and a liquid ammonia storage and supply system;
according to the mass conservation, momentum conservation and energy conservation equations, in the modeling process, the dynamic simulation software selects corresponding component modules from the model library according to the process flows of the limestone-gypsum wet desulphurization system and the SCR flue gas denitration system, connects the component modules, inputs initial data and completes the model construction of the power plant desulphurization and denitration system;
and (3) building an analog quantity control system, a sequence control system and a logic control system according to a field control strategy, and configuring by adopting a basic algorithm module to realize the same function as the actual control system and build a complete power plant desulfurization and denitrification system simulation model.
3. The optimal control method for desulfurization and denitrification of flue gas based on the hysteresis model as claimed in claim 2, wherein the simulation model of the power plant desulfurization and denitrification system further comprises:
in the process of model development and debugging, physical data acquired by an actual power plant desulfurization and denitrification system is compared with virtual data acquired based on a power plant desulfurization and denitrification simulation model, whether an error exceeds a threshold value is judged, if the error exceeds the threshold value, the virtual data with larger error is classified through cluster learning, corresponding historical data is used as input, error learning is carried out through a neural network, a correction coefficient is output to correct the error data of the virtual data, and virtual-real fusion is carried out on the corrected virtual data and the physical data to generate a verified power plant desulfurization and denitrification simulation model.
4. The method for optimally controlling the desulfurization and the denitrification of the flue gas based on the lag model as recited in claim 1, wherein in the step 2, the step of establishing the first lag time prediction model for the time delay of the measurement of the PH value of the absorption tower in the power plant desulfurization system by adopting the variable point detection, the time window sliding, the correlation analysis and the machine learning model comprises the following steps: establishing a slurry pH value response lag time identification algorithm flow by adopting a variable point detection, time window sliding and correlation analysis method and establishing a first lag time prediction model by adopting a machine learning model;
the flow of the slurry pH value response lag time identification algorithm comprises the following steps:
selecting a working condition that the concentration value of sulfur dioxide at the outlet of the absorption tower changes after the pH value of the slurry of the absorption tower is adjusted as an identification object;
the time window Δ t is equally divided into two equally spaced time windows Δ ti1And Δ ti2Gradually sliding forwards on a time axis, calculating the average difference value of the sulfur dioxide concentration in two time windows, and if the average difference value exceeds a set threshold value, taking the moment as a working condition change point tiIf the value is less than the set threshold value, the operation is continued to be carried forwardSliding the time window until a working condition change point is detected or the time window slides to a cut-off time point;
based on the operating condition change point tiAnd a time window delta t, respectively acquiring a slurry PH value time sequence and a sulfur dioxide concentration time sequence from the start of working condition change to the end of the time window;
gradually advancing the sulfur dioxide concentration time sequence, setting a maximum moving step number k, obtaining a new sulfur dioxide concentration sequence through the advancement and constructing a sulfur dioxide concentration time lag matrix V;
calculating the PH value time sequence of the slurry and the Pearson correlation coefficient r of each column in the matrix V, wherein the delay time corresponding to the maximum correlation coefficient is the PH value response lag time t under the working condition1;
Establishing a first lag time prediction model using a machine learning model comprises:
acquiring original data characteristics in a power plant desulfurization system, preprocessing the original data characteristics, substituting the preprocessed data characteristics into the flow of the slurry PH value response lag time identification algorithm to carry out PH value delay identification, and acquiring the relationship between delay time and different operation data characteristics; the raw data features include at least: the load capacity of the boiler, the air supply quantity of the boiler, the flow rate of limestone slurry, the input quantity of the limestone slurry, the sulfur dioxide content, the calcium carbonate content in limestone and the distance data characteristics from the absorption tower to a PH measuring point;
the operation data characteristics which are obtained by identification and can cause the PH value change are converted into characteristics with more working condition characteristics in a characteristic conversion mode, the correlation among the data characteristics is reduced, and the influence caused by dimension is eliminated by carrying out normalization processing on the converted data characteristics;
performing correlation analysis on the original operation data characteristics by adopting a correlation analysis method to obtain correlation coefficients of each operation data characteristic and PH value response lag time, wherein the higher the correlation coefficient is, the most correlation between the data characteristic and the lag time is shown;
fusing the operating data characteristics by adopting a characteristic fusion method according to the height of the correlation coefficient to form new fusion characteristics, taking the original operating data characteristics and the new fusion characteristics as sample data, and inputting a training set in the sample data into a machine learning model according to a preset proportion to establish a first lag time prediction model under different operating data change conditions; and calculating the PH value response lag time according to different operation data characteristics through the first lag time prediction model.
5. The method of claim 1, wherein in step 2, establishing a second lag time prediction model for the time delay of the inlet of the plant denitration system for determining the existence of the concentration of nitrogen oxides in flue gas by a flue gas on-line monitoring system (CEMS) by using a variable point detection model, a time window sliding model, a correlation analysis model and a machine learning model comprises: establishing a CEMS determination lag time identification algorithm flow by adopting a variable point detection method, a time window sliding method and a correlation analysis method, and establishing a second lag time prediction model by adopting a machine learning model;
the CEMS determination lag time identification algorithm flow comprises the following steps:
selecting a working condition that the CEMS measured value changes after the concentration of the inlet nitrogen oxide changes as an identification object; the measurement of the concentration of the nitrogen oxides passes through the heat tracing guide pipe and the analysis cabinet, and the flow of the flue gas in the heat tracing guide pipe and the concentration measurement in the analysis cabinet have certain time lag;
the time window Δ t' is equally divided into two equally spaced time windows Δ ti1' and Δ ti2Gradually sliding forward on a time axis, calculating the average difference value of the CEMS measured values in two time windows, and if the average difference value exceeds a set threshold value, determining the moment as a working condition change point tiIf the time window is smaller than the set threshold, continuing to slide the time window forwards until a working condition change point is detected or the time window slides to a cut-off time point;
based on the operating condition change point ti'and a time window delta t' are respectively obtained, wherein a nitrogen oxide concentration value time sequence and a CEMS measurement value time sequence from the start of working condition change to the end of the time window are obtained;
gradually advancing the CEMS measured value time sequence, setting the maximum moving step number k, obtaining a new CEMS measured value sequence through the advancing and constructing a CEMS measured value time lag matrix V';
calculating the time sequence of the concentration value of the nitrogen oxide and the Pearson correlation coefficient r 'of each column in the matrix V', wherein the delay time corresponding to the maximum correlation coefficient is the lag time t of the measurement of the concentration of the nitrogen oxide under the working condition2;
Establishing a second lag time prediction model using a machine learning model comprises:
collecting original data characteristics in a denitration system of a power plant, preprocessing the original data characteristics, substituting the preprocessed data characteristics into a CEMS (continuous emission monitoring system) determination lag time identification algorithm flow for delay identification, and acquiring the relation between delay time and different operation data characteristics; the raw data features include at least: boiler load, coal type, coal feeding amount, combustion temperature, air volume and flue gas volume;
the method comprises the steps of converting the identified operating data characteristics capable of causing the concentration change of the nitrogen oxides into characteristics with more working condition characteristics in a characteristic conversion mode, reducing the correlation among the data characteristics, and eliminating the influence caused by dimension by carrying out normalization processing on the converted data characteristics;
performing correlation analysis on the original operation data characteristics by adopting a correlation analysis method to obtain correlation coefficients of each operation data characteristic and the determination lag time of the concentration value of the nitric oxide, wherein the higher the correlation coefficient is, the most correlation between the data characteristics and the lag time is shown;
fusing the operating data characteristics by adopting a characteristic fusion method according to the height of the correlation coefficient to form new fusion characteristics, taking the original operating data characteristics and the new fusion characteristics as sample data, and inputting a training set in the sample data into a machine learning model according to a preset proportion to establish second lag time prediction models under different operating data change conditions; and calculating the determined lag time of the CEMS according to different operating data characteristics by using a second lag time prediction model.
6. The optimal control method for flue gas desulfurization and denitration based on the hysteresis model as claimed in claim 4 or 5, wherein the XGboost model is selected as the machine learning model, and is an integrated learning algorithm based on a boosting methodThe learner selects a CART decision tree, applying k CART functions { f }1,f2,…,fkAdding to form an integrated tree model; the target function of the model consists of a loss function and a regular term, and the loss function is approximated by second-order Taylor expansion; optimizing key parameters to improve the accuracy of model prediction, wherein the key parameters comprise the maximum depth of a tree, subsamples, the column number ratio of random sampling of each tree, the minimum leaf node sample weight and the learning rate;
the method comprises the steps of constructing a model, starting from a root node, sequencing training set data according to each data feature, calculating the profit of each feature by adopting a greedy method, selecting the feature with the maximum profit as a splitting feature, mapping the training set data to corresponding leaf nodes, performing recursion on the generated leaf nodes until a limiting condition is reached, finishing the generation process of a decision tree, calculating the weight of the leaf nodes of the decision tree by using first-order and second-order derivatives of a loss function, taking the weight as a fitting target of the next tree, performing recursion repeatedly until the condition is met, and finishing the establishment of the model.
7. The optimal control method for flue gas desulfurization and denitration based on the hysteresis model as claimed in claim 1, wherein in step 3, after collecting the historical operating parameters of the power plant desulfurization system and the denitration system, selecting the operating parameters strongly related to the power plant desulfurization and inputting the selected operating parameters into the constructed first support vector machine model to predict the sulfur dioxide concentration in the flue gas at the outlet of the power plant desulfurization system, specifically comprising:
taking the collected historical operating parameters of the power plant desulfurization system as sample data, performing correlation analysis on the sample data, removing the sample data of which the correlation with the sulfur dioxide concentration in the flue gas at the outlet of the limestone-gypsum wet desulfurization system is smaller than a preset value, and taking the residual sample data as operating data strongly correlated with the desulfurization system; the historical operating parameters of the desulfurization system at least comprise sulfur dioxide concentration at an inlet, nitrogen oxide concentration, unit load, limestone slurry circulating pump current, slurry supply quantity, flue gas sulfur dioxide concentration at an outlet of an absorption tower and slurry pH value;
performing data preprocessing on the operation data strongly related to the desulfurization system, and constructing a first support vector machine model by using the preprocessed data;
collecting real-time operation data related to power plant desulfurization and inputting the real-time operation data into a constructed first support vector machine model to obtain a predicted value of the concentration of sulfur dioxide in flue gas at an outlet of a power plant desulfurization system;
wherein the data preprocessing comprises: filling missing values and abnormal values of the operation data strongly related to the desulfurization system and normalizing the operation data to obtain a pretreated desulfurization data sequence, and recording the pretreated desulfurization data sequence as F ═ F1,f2,f3,…,fn],fiThe desulfurization data of the ith time point in the desulfurization data sequence after treatment is obtained;
carrying out wavelet threshold denoising processing on the desulfurization data sequence F, carrying out wavelet decomposition on noisy data with noise to obtain real data information, and marking as P ═ P1,p2,p3,…,pm],piThe desulfurization data is the desulfurization data of the ith time point in the real desulfurization data sequence.
8. The optimal control method for flue gas desulfurization and denitration based on the hysteresis model as claimed in claim 1, wherein in step 3, the operation parameters strongly related to denitration of the power plant are selected and input into the constructed second support vector machine model to predict the concentration of nitrogen oxides in flue gas at the inlet of the denitration system of the power plant, and the method specifically comprises the following steps:
the collected historical operating parameters of the power plant denitration system are used as sample data, correlation between the sample data and the concentration of nitrogen oxides in flue gas at an inlet of the power plant denitration system is calculated by adopting a Pearson correlation coefficient, and a data combination with high correlation is selected according to the correlation and is used as operating data strongly related to the denitration system; historical operating data of the denitration system at least comprises ammonia injection mass flow, boiler load, SCR inlet flue gas temperature, SCR inlet flue gas oxygen content, SCR inlet nitrogen oxide concentration and SCR denitration efficiency;
performing data preprocessing on operating data strongly related to the denitration system, and constructing a second support vector machine model by using the preprocessed data;
collecting real-time operation data related to power plant denitration and inputting the real-time operation data into a constructed second support vector machine model to obtain a predicted value of the concentration of nitrogen oxides in flue gas at an inlet of a denitration system of a power plant;
wherein, select strong correlation and extremely strong relevant data as the data that the correlation with deNOx systems is high, the computational formula is:
x is the input sample data characteristic, Y is the nitrogen oxide concentration at the inlet, cov (X, Y) represents the covariance of X, Y; sigmaXAnd σYIs the standard deviation of X and Y respectively, and rho represents the correlation coefficient between two variables, and the value range is [ -1,1](ii) a When rho is more than or equal to 0.8<1, referred to as very strong correlation; when rho is more than or equal to 0.6<0.8, called strong correlation; when rho is more than or equal to 0.4<At 0.6, it is said to be moderately correlated; when rho is more than or equal to 0.2<0.4, referred to as weak correlation; when rho is more than or equal to 0.0<At 0.2, it is referred to as very weakly correlated or uncorrelated.
9. The optimization control method for flue gas desulfurization and denitration based on the hysteresis model as claimed in claim 7 or 8, wherein the constructing of the first support vector machine model and the second support vector machine model comprises: determining the optimal support vector machine parameters by adopting a cuckoo optimization method: initializing parameters of a cuckoo optimization algorithm, and searching the position of a bird nest by step length self-adaptive and dynamically adjusted Laevice flight according to the parameters of the cuckoo optimization method:wherein x isi (t +1)The position of the ith bird nest in the tth generation; a is step control quantity, is used for controlling the search range of the step, and obeys positive space distribution; l (lambda) is a Levy random walk path; the step self-adaptive dynamic adjustment strategy is as follows:
stepi=stepmin+(stepmax-stepmin)di
wherein stepiStep for the current search stepmaxStep being the maximum value of the step sizeminIs the minimum value of the step size, niIs the position of the ith bird nest, nbestThe current minimum fitness corresponds to the nest position of the nest, dmaxThe maximum value of the distance between the bird nest corresponding to the current minimum fitness and other bird nests is obtained;
training a support vector machine model by adopting a training set in the preprocessed data, calculating the fitness of each bird nest position, and keeping the bird nest corresponding to the minimum fitness to the next iteration;
judging whether the minimum fitness meets a preset termination condition, if so, determining the nest position of the nest corresponding to the minimum fitness to be the determined optimal support vector machine parameter, and if not, removing a plurality of nests with the highest fitness and readjusting the nest positions;
training a support vector machine model according to the determined optimal support vector machine parameters: establishing a kernel function-based support vector machine training program, forming a mapping relation between an input variable and an output variable through a support vector machine model, performing learning training on training sample data by using the support vector machine training program, and obtaining N support vectors X through learning and trainingi *I is 0,1, …, N, forming a support vector machine model:
wherein, Xi *Support vector, Y, representing a desulfurization or denitrification system of a power plantiSulfur dioxide concentration representing a support vector of a desulfurization system of a power plant or nitrogen oxide concentration, alpha, of a support vector of a denitrification systemiCoefficients representing the ith support vector, X being the input preconditionsAnd Y (X) represents a predicted value of the concentration of sulfur dioxide of a supporting vector of the desulfurization system of the power plant or a predicted value of the concentration of nitrogen oxide of a supporting vector of the denitrification system, K (-) represents a kernel function of the supporting vector machine, and the kernel function is one of a Gaussian function, a polynomial function, a linear function and a radial basis function.
10. The optimal control method for desulfurization and denitrification of flue gas based on the hysteresis model as claimed in claim 1, wherein the step 5 comprises in detail: after the slurry spraying amount control parameter, the ammonia spraying amount control parameter and the relevant configuration parameter of the operation of the power plant desulfurization and denitrification system are input into the power plant desulfurization and denitrification system simulation model, the obtained real-time operation parameter of the power plant desulfurization and denitrification system is compared with the simulation result data of the simulation model through the set expert diagnosis module to obtain a deviation, and pre-alarming is realized through whether the deviation exceeds a preset threshold value or not;
the expert diagnosis module is internally provided with an intelligent diagnosis strategy, judges related running states, data deviation and pre-alarm information conditions through preset logics, comprehensively outputs diagnosis preliminary result information, calls expert base knowledge information for comparison, analyzes whether conclusion information obtained by the intelligent diagnosis strategy is related or consistent with the expert base knowledge information, and outputs a diagnosis analysis result, running guidance or a task list; the knowledge information of the expert database comprises stored preset knowledge and information that an abnormal fault occurs; the pre-alarm comprises a parameter exceeding a preset threshold, time for the parameter exceeding the preset threshold and abnormal fault information.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111483083.0A CN114225662B (en) | 2021-12-07 | 2021-12-07 | Hysteresis model-based flue gas desulfurization and denitrification optimal control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111483083.0A CN114225662B (en) | 2021-12-07 | 2021-12-07 | Hysteresis model-based flue gas desulfurization and denitrification optimal control method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114225662A true CN114225662A (en) | 2022-03-25 |
CN114225662B CN114225662B (en) | 2024-02-13 |
Family
ID=80753603
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111483083.0A Active CN114225662B (en) | 2021-12-07 | 2021-12-07 | Hysteresis model-based flue gas desulfurization and denitrification optimal control method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114225662B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114708924A (en) * | 2022-03-28 | 2022-07-05 | 大唐环境产业集团股份有限公司 | Model construction method and device for predicting soot blowing interval time of soot blower in SCR system |
CN115146833A (en) * | 2022-06-14 | 2022-10-04 | 北京全应科技有限公司 | Method for predicting generation concentration of boiler nitrogen oxide |
CN115828757A (en) * | 2022-12-12 | 2023-03-21 | 福建中锐汉鼎数字科技有限公司 | Flood discharge hysteresis characteristic construction and selection method for basin water level prediction |
CN117190173A (en) * | 2023-09-25 | 2023-12-08 | 天津大学 | Optimal control method and control system for flue gas recirculation and boiler coupling system |
CN117273497A (en) * | 2023-11-20 | 2023-12-22 | 一夫科技股份有限公司 | Production optimization method and system for high-strength gypsum |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102565274A (en) * | 2012-01-17 | 2012-07-11 | 广东电网公司电力科学研究院 | Modifying method for catalyst activity in power plant SCR (selective catalytic reduction) denitration system |
CN106000041A (en) * | 2016-05-30 | 2016-10-12 | 安徽工业大学 | Ammonia process desulfurization spraying tower system and method for determining SO2 absorption mass transfer coefficient |
CN108636094A (en) * | 2018-07-12 | 2018-10-12 | 浙江大学 | A kind of accurate PREDICTIVE CONTROL in wet desulfurizing process and energy conserving system and method |
CN108803309A (en) * | 2018-07-02 | 2018-11-13 | 大唐环境产业集团股份有限公司 | It is a kind of that ammonia optimization method and system are intelligently sprayed based on the SCR denitration of hard measurement and model adaptation |
CN111589301A (en) * | 2020-05-29 | 2020-08-28 | 广东电科院能源技术有限责任公司 | Method, device, equipment and storage medium for predicting SCR denitration performance of coal-fired power plant |
CN112206640A (en) * | 2020-09-16 | 2021-01-12 | 西安热工研究院有限公司 | Limestone slurry pH value concentration fly-up detection system, method, control system and desulfurization system |
CN112580250A (en) * | 2020-11-12 | 2021-03-30 | 山东纳鑫电力科技有限公司 | Thermal power generating unit denitration system based on deep learning and optimization control method |
CN112892209A (en) * | 2021-01-25 | 2021-06-04 | 山东戈尔环境科技有限公司 | Ammonia spraying control system |
CN113082954A (en) * | 2021-04-07 | 2021-07-09 | 浙江大学 | Whole-process intelligent operation regulation and control system of wet desulphurization device |
WO2021208396A1 (en) * | 2020-04-13 | 2021-10-21 | 西安热工研究院有限公司 | Efficient double-inlet channel rotational flow atomization desulfurization nozzle |
-
2021
- 2021-12-07 CN CN202111483083.0A patent/CN114225662B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102565274A (en) * | 2012-01-17 | 2012-07-11 | 广东电网公司电力科学研究院 | Modifying method for catalyst activity in power plant SCR (selective catalytic reduction) denitration system |
CN106000041A (en) * | 2016-05-30 | 2016-10-12 | 安徽工业大学 | Ammonia process desulfurization spraying tower system and method for determining SO2 absorption mass transfer coefficient |
CN108803309A (en) * | 2018-07-02 | 2018-11-13 | 大唐环境产业集团股份有限公司 | It is a kind of that ammonia optimization method and system are intelligently sprayed based on the SCR denitration of hard measurement and model adaptation |
CN108636094A (en) * | 2018-07-12 | 2018-10-12 | 浙江大学 | A kind of accurate PREDICTIVE CONTROL in wet desulfurizing process and energy conserving system and method |
WO2021208396A1 (en) * | 2020-04-13 | 2021-10-21 | 西安热工研究院有限公司 | Efficient double-inlet channel rotational flow atomization desulfurization nozzle |
CN111589301A (en) * | 2020-05-29 | 2020-08-28 | 广东电科院能源技术有限责任公司 | Method, device, equipment and storage medium for predicting SCR denitration performance of coal-fired power plant |
CN112206640A (en) * | 2020-09-16 | 2021-01-12 | 西安热工研究院有限公司 | Limestone slurry pH value concentration fly-up detection system, method, control system and desulfurization system |
CN112580250A (en) * | 2020-11-12 | 2021-03-30 | 山东纳鑫电力科技有限公司 | Thermal power generating unit denitration system based on deep learning and optimization control method |
CN112892209A (en) * | 2021-01-25 | 2021-06-04 | 山东戈尔环境科技有限公司 | Ammonia spraying control system |
CN113082954A (en) * | 2021-04-07 | 2021-07-09 | 浙江大学 | Whole-process intelligent operation regulation and control system of wet desulphurization device |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114708924A (en) * | 2022-03-28 | 2022-07-05 | 大唐环境产业集团股份有限公司 | Model construction method and device for predicting soot blowing interval time of soot blower in SCR system |
CN115146833A (en) * | 2022-06-14 | 2022-10-04 | 北京全应科技有限公司 | Method for predicting generation concentration of boiler nitrogen oxide |
CN115828757A (en) * | 2022-12-12 | 2023-03-21 | 福建中锐汉鼎数字科技有限公司 | Flood discharge hysteresis characteristic construction and selection method for basin water level prediction |
CN115828757B (en) * | 2022-12-12 | 2024-02-23 | 福建中锐汉鼎数字科技有限公司 | Flood discharge hysteresis characteristic structure and selection method for drainage basin water level prediction |
CN117190173A (en) * | 2023-09-25 | 2023-12-08 | 天津大学 | Optimal control method and control system for flue gas recirculation and boiler coupling system |
CN117190173B (en) * | 2023-09-25 | 2024-03-29 | 天津大学 | Optimal control method and control system for flue gas recirculation and boiler coupling system |
CN117273497A (en) * | 2023-11-20 | 2023-12-22 | 一夫科技股份有限公司 | Production optimization method and system for high-strength gypsum |
CN117273497B (en) * | 2023-11-20 | 2024-02-09 | 一夫科技股份有限公司 | Production optimization method and system for high-strength gypsum |
Also Published As
Publication number | Publication date |
---|---|
CN114225662B (en) | 2024-02-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114225662B (en) | Hysteresis model-based flue gas desulfurization and denitrification optimal control method | |
CN114191953A (en) | Flue gas desulfurization and denitrification control method based on convolutional neural network and XGboost | |
CN111804146B (en) | Intelligent ammonia injection control method and intelligent ammonia injection control device | |
CN110716512A (en) | Environmental protection equipment performance prediction method based on coal-fired power plant operation data | |
CN113433911B (en) | Accurate control system and method for ammonia spraying of denitration device based on accurate concentration prediction | |
CN113175678B (en) | Garbage incineration monitoring method and device | |
CN116688754A (en) | Ship flue gas desulfurization automatic control system and method thereof | |
CN112613237B (en) | CFB unit NOx emission concentration prediction method based on LSTM | |
CN112742187A (en) | Method and device for controlling pH value in desulfurization system | |
CN112465239B (en) | Desulfurization system operation optimization method based on improved PSO-FCM algorithm | |
CN112364562B (en) | Flue gas environment-friendly island cooperative control method and system | |
CN111860701B (en) | Denitration system working condition discrimination preprocessing method based on clustering method | |
CN113440990A (en) | EMD-LSTM based outlet SO2Concentration prediction method | |
CN113469449A (en) | Optimizing control method and system for desulfurization system | |
CN115034129A (en) | NOx emission concentration soft measurement method for thermal power plant denitration device | |
CN113450880A (en) | Desulfurization system inlet SO2Intelligent concentration prediction method | |
CN114186708A (en) | Circulating fluidized bed unit SO based on PSO-ELM2Concentration prediction method | |
Hasnen et al. | Semi-supervised adaptive PLS soft-sensor with PCA-based drift correction method for online valuation of NOx emission in industrial water-tube boiler | |
WO2024130992A1 (en) | Online soft measurement method for dioxin (dxn) emission concentration during mswi process | |
Yin et al. | Enhancing deep learning for the comprehensive forecast model in flue gas desulfurization systems | |
CN111008735A (en) | Predictive emission management system and method | |
CN116312869A (en) | Method, device and system for predicting nitrogen oxides in catalytic cracking regenerated flue gas | |
CN113869359B (en) | Method for predicting nitrogen oxides in urban solid waste incineration process based on modularized neural network | |
CN115201408A (en) | Method for predicting concentration of sulfur dioxide at desulfurization outlet under all working conditions | |
JP4211173B2 (en) | Method for calculating the concentration of SO3 gas in the flue |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |