CN116738862A - Boiler scaling fault prediction method, device, equipment and medium - Google Patents
Boiler scaling fault prediction method, device, equipment and medium Download PDFInfo
- Publication number
- CN116738862A CN116738862A CN202310854521.2A CN202310854521A CN116738862A CN 116738862 A CN116738862 A CN 116738862A CN 202310854521 A CN202310854521 A CN 202310854521A CN 116738862 A CN116738862 A CN 116738862A
- Authority
- CN
- China
- Prior art keywords
- thermal resistance
- boiler
- model
- dirt
- target
- 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
- 238000000034 method Methods 0.000 title claims abstract description 88
- 239000002245 particle Substances 0.000 claims abstract description 137
- 238000004088 simulation Methods 0.000 claims abstract description 79
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 32
- 238000011002 quantification Methods 0.000 claims abstract description 6
- 230000007704 transition Effects 0.000 claims description 43
- 238000009826 distribution Methods 0.000 claims description 42
- 238000005070 sampling Methods 0.000 claims description 26
- 230000006870 function Effects 0.000 claims description 24
- 238000004140 cleaning Methods 0.000 claims description 19
- 238000009499 grossing Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 7
- 238000000342 Monte Carlo simulation Methods 0.000 claims description 6
- 238000012952 Resampling Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 abstract description 26
- 238000012546 transfer Methods 0.000 abstract description 10
- 238000010586 diagram Methods 0.000 description 8
- 230000008859 change Effects 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 239000000498 cooling water Substances 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 230000000750 progressive effect Effects 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 239000012530 fluid Substances 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000009991 scouring Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a boiler scaling fault prediction method, a device, equipment and a medium, which comprise the following steps: after a target dirt thermal resistance model is built on the basis of a least square method by adopting historical dirt thermal resistance of the boiler, updating the target dirt thermal resistance model in real time on the basis of a particle swarm algorithm, inputting a plurality of simulated dirt thermal resistances generated by the target dirt thermal resistance model into a simulation temperature obtained by a boiler simulation loop, carrying out uncertainty factor quantification on all the simulation temperatures to build a state transfer model, carrying out multi-step prediction according to the state transfer model and a preset temperature failure limit value to determine the failure moment of the boiler, and carrying out difference value operation on the failure moment and the prediction moment to output the residual service life of the boiler. In the whole boiler scaling fault prediction process, a target scaling thermal resistance model is updated in real time, and multi-step prediction is performed after the uncertainty of the simulation temperature is effectively quantified based on differential modeling, so that the accuracy of boiler structure fault prediction is improved.
Description
Technical Field
The invention relates to the technical field of boilers, in particular to a method, a device, equipment and a medium for predicting boiler scaling faults.
Background
In the running process of the boiler, when a great amount of dirt blocks the water wall pipe, the safety accident of the boiler is easily caused. In order to prevent and reduce the adverse effect of scale on the boiler, it is necessary to take measures for scale prevention and removal such as periodic cleaning, treatment of scale-causing fluids, and the like. Although the damage of the dirt is relieved, the service life of the equipment is easily influenced by frequent start-stop, so that the normal operation period of the equipment is shortened, and therefore, the prediction of the influence of the dirt on the heat exchange performance of the boiler by adopting a fault prediction technology is a key for scientifically arranging the dirt cleaning.
The traditional fault prediction technology builds a fault prediction model by adopting a large amount of historical scaling fault data, on one hand, the historical scaling fault data is limited by the high acquisition cost of key equipment data in partial actual operation, and on the other hand, the acquired historical scaling fault data has strong uncertainty, so that modeling errors of the fault prediction model are caused, and the accuracy of scaling fault prediction is reduced.
Disclosure of Invention
The invention provides a boiler scaling fault prediction method, device, equipment and medium, which solve the technical problem that the accuracy of scaling fault prediction is lower due to the fact that a large amount of historical scaling fault data with incompleteness and uncertainty are relied on when a fault prediction model is constructed to predict the faults of a boiler in the prior art.
The invention provides a boiler scaling fault prediction method, which comprises the following steps:
constructing a target dirt thermal resistance model by adopting a historical dirt thermal resistance model and an initial dirt thermal resistance model of the boiler according to a least square method, and updating the target dirt thermal resistance model in real time based on a particle swarm algorithm;
inputting a plurality of simulated dirt thermal resistances related to the prediction time generated by the target dirt thermal resistance model into a boiler simulation loop, and determining a plurality of simulated temperatures;
carrying out uncertainty factor quantification on all the simulation temperatures, and constructing a state transition model;
performing multi-step prediction according to the state transition model and a preset temperature failure limit value, and determining the failure moment of the boiler;
and carrying out difference value operation on the failure moment and the predicted moment, and outputting the residual service life of the boiler.
Optionally, the step of constructing a target fouling thermal resistance model by adopting a historical fouling thermal resistance model and an initial fouling thermal resistance model of the boiler according to a least square method and updating the target fouling thermal resistance model in real time based on a particle swarm algorithm comprises the following steps:
acquiring a historical dirt thermal resistance and an initial dirt thermal resistance model of the boiler;
constructing a loss function related to dirt thermal resistance based on a least square method, and adopting the historical dirt thermal resistance to carry out equation solving on the loss function to determine a first thermal resistance parameter;
Updating the initial dirt thermal resistance model by adopting the first thermal resistance parameter to generate a target dirt thermal resistance model;
acquiring actual dirt thermal resistance of a boiler and the operation time of the actual dirt thermal resistance after cleaning in real time;
inputting the target dirt thermal resistance model by adopting the operation time after cleaning, and solving the corresponding theoretical dirt thermal resistance;
when the thermal resistance difference value between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference value threshold, updating the target dirt thermal resistance model based on a particle swarm algorithm, and generating a new target dirt thermal resistance model.
Optionally, when the thermal resistance difference between any one of the theoretical thermal resistances and the corresponding actual thermal resistance exceeds a preset thermal resistance difference threshold, updating the target thermal resistance model based on a particle swarm algorithm, and generating a new target thermal resistance model, including:
when the thermal resistance difference value between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference value threshold, constructing a neighborhood range by taking the first thermal resistance parameter as the center;
generating a particle swarm in the neighborhood range, and initializing the speed and the position of each initial particle in the particles;
Evaluating the fitness of each initial particle through a preset fitness function, taking the position of the initial particle with the minimum fitness in the particle swarm as a global optimal position, carrying out iterative updating on the speed and the position of each initial particle through a particle swarm algorithm, and taking the global optimal position when the iterative updating meets the iteration stopping condition as a second thermal resistance parameter;
updating the target dirt thermal resistance model by adopting the second thermal resistance parameter to generate a new target dirt thermal resistance model.
Optionally, the step of inputting a plurality of simulated scale thermal resistances associated with the predicted time generated by the target scale thermal resistance model into a boiler simulation loop, and determining a plurality of simulated temperatures includes:
acquiring the predicted time of the boiler, sequentially adopting an associated target dirt thermal resistance model to operate according to a preset time step from the predicted time, and sequentially generating a plurality of simulated dirt thermal resistances;
and sequentially inputting the simulated dirt thermal resistance into a boiler simulation loop of the boiler to perform heat exchange simulation, and outputting a plurality of corresponding simulation temperatures.
Optionally, the step of quantifying uncertainty factors of all the simulation temperatures and constructing a state transition model includes:
Carrying out data smoothing on all the simulation temperatures by adopting a filter fitting method, and correspondingly generating a plurality of smoothing temperatures;
respectively carrying out difference operation on each simulation temperature and the corresponding smooth temperature to determine a plurality of first temperature differences;
performing variance operation on all the first temperature differences to determine a first variance, and constructing an observed Gaussian uncertainty meeting the first variance;
performing differential operation on all the smooth temperatures to obtain a plurality of differential temperatures, and performing curve fitting by adopting all the differential temperatures to generate a differential model;
respectively carrying out difference operation on each differential temperature and the differential fitting temperature correspondingly constructed through the differential model, and determining a plurality of second temperature differences;
performing variance operation on the basis of all the second temperature differences to determine a second variance, and constructing a state Gaussian uncertainty meeting the second variance;
and carrying out differential modeling according to the observed Gaussian uncertainty, the differential model and the state Gaussian uncertainty to generate a state transition model.
Optionally, the step of performing multi-step prediction according to the state transition model and a preset temperature failure limit value to determine the failure time of the boiler includes:
Acquiring prior probability distribution corresponding to the state transition model and a preset temperature failure limit value of the boiler, and generating a sampling particle set by the prior probability distribution;
constructing posterior probability distribution by adopting the sampling particle set according to a Monte Carlo method;
carrying out weight normalization and particle resampling on the sampled particle sets to determine target sampled particle sets;
performing multi-step prediction by adopting the target sampling particle set and the posterior probability distribution, and outputting predicted temperature probability distribution;
calculating a plurality of failure probabilities of the target sample particle set according to the temperature failure limit value and the predicted temperature probability distribution;
and calculating expected values based on all the failure probabilities, and determining the failure moment of the boiler.
The second aspect of the invention provides a boiler scale fault prediction device, comprising:
the target dirt thermal resistance model construction module is used for constructing a target dirt thermal resistance model according to a least square method by adopting a historical dirt thermal resistance model and an initial dirt thermal resistance model of the boiler, and updating the target dirt thermal resistance model in real time based on a particle swarm algorithm;
the simulated temperature generation module is used for inputting a plurality of simulated dirt thermal resistances related to the predicted time generated by the target dirt thermal resistance model into a boiler simulation loop to determine a plurality of simulated temperatures;
The state transition model construction module is used for quantifying uncertainty factors of all the simulation temperatures and constructing a state transition model;
the failure moment determining module is used for carrying out multi-step prediction according to the state transition model and a preset temperature failure limit value, and determining the failure moment of the boiler;
and the remaining usable life output module is used for carrying out difference value operation on the failure moment and the predicted moment and outputting the remaining usable life of the boiler.
Optionally, the simulation temperature generation module is specifically configured to:
acquiring the predicted time of the boiler, sequentially adopting an associated target dirt thermal resistance model to operate according to a preset time step from the predicted time, and sequentially generating a plurality of simulated dirt thermal resistances;
and sequentially inputting the simulated dirt thermal resistance into a boiler simulation loop of the boiler to perform heat exchange simulation, and outputting a plurality of corresponding simulation temperatures.
An electronic device according to a third aspect of the present invention includes a memory and a processor, where the memory stores a computer program, and the computer program when executed by the processor causes the processor to execute the steps of the boiler fouling fault prediction method according to any one of the first aspect of the present invention.
A fourth aspect of the present invention provides a computer readable storage medium having stored thereon a computer program which, when executed, implements the boiler fouling fault prediction method according to any one of the first aspects of the present invention.
From the above technical scheme, the invention has the following advantages:
according to the invention, after a target dirt thermal resistance model is constructed based on a least square method by adopting historical dirt thermal resistance of a boiler, the target dirt thermal resistance model is updated in real time based on a particle swarm algorithm, a plurality of simulated dirt thermal resistances are generated from a prediction moment through the target dirt thermal resistance model, all simulated dirt thermal resistances are input into a boiler simulation loop to obtain corresponding simulation temperatures, uncertainty quantification is carried out on all simulation temperatures to construct a state transition model, multi-step prediction is carried out according to the state transition model and a preset temperature failure limit value to determine the failure moment of the boiler, difference operation is carried out between the failure moment and the prediction moment, and the residual service life of the boiler is output. In the whole boiler scaling fault prediction process, a target scaling thermal resistance model is updated in real time according to actual scaling thermal resistance of a boiler, and multi-step prediction is performed after uncertainty of simulation temperature is effectively quantified based on differential modeling, so that accuracy of boiler structure fault prediction is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained from these drawings without inventive faculty for a person skilled in the art.
FIG. 1 is a flow chart showing the steps of a method for predicting boiler fouling failure according to a first embodiment of the present invention;
FIG. 2 is a graph showing the trend of the growth of the progressive scale thermal resistance according to the first embodiment of the present invention;
FIG. 3 is a schematic diagram of a prediction result of remaining usable life of a boiler in a full life cycle according to an embodiment of the present invention;
FIG. 4 is a flow chart illustrating a method for predicting boiler fouling failure according to a second embodiment of the present invention;
FIG. 5 is a schematic diagram of a change curve of theoretical thermal resistance and actual thermal resistance of fouling according to a second embodiment of the present invention;
FIG. 6 is a schematic diagram of a first temperature difference according to a second embodiment of the present invention;
FIG. 7 is a schematic diagram of probability distribution for observing Gaussian uncertainty according to a second embodiment of the invention;
FIG. 8 is a schematic diagram of a differential model according to a second embodiment of the present invention;
FIG. 9 is a schematic diagram of probability distribution of state Gaussian uncertainty according to a second embodiment of the invention;
FIG. 10 is a block diagram of a boiler fouling failure prediction apparatus according to an embodiment of the present invention.
Detailed Description
The embodiment of the invention provides a boiler scaling fault prediction method, device, equipment and medium, which are used for solving the technical problem that the accuracy of scaling fault prediction is lower due to the fact that a large amount of historical scaling fault data with incompleteness and uncertainty are relied on when a fault prediction model is constructed to perform fault prediction on a boiler in the prior art.
In order to make the objects, features and advantages of the present invention more comprehensible, the technical solutions in the embodiments of the present invention are described in detail below with reference to the accompanying drawings, and it is apparent that the embodiments described below are only some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, fig. 1 is a flowchart illustrating a method for predicting a boiler fouling failure according to an embodiment of the present invention.
The invention provides a boiler scaling fault prediction method, which comprises the following steps:
step 101, constructing a target dirt thermal resistance model by adopting a historical dirt thermal resistance model and an initial dirt thermal resistance model of the boiler according to a least square method, and updating the target dirt thermal resistance model in real time based on a particle swarm algorithm.
Historical dirt thermal resistance refers to actual dirt thermal resistance corresponding to the time of historic recording of different running time after cleaning. The operation time after cleaning refers to the operation time of the boiler after cleaning dirt on the heat exchange surface.
An initial fouling thermal resistance model refers to a progressive empirical model reflecting the change in fouling thermal resistance over time.
The target fouling thermal resistance model refers to an initial fouling thermal resistance model reflecting the actual fouling thermal resistance change of the boiler.
In the embodiment of the invention, in the process of forming dirt of the boiler water wall pipe, the method can be divided into two action forms, wherein one action form is that dirt substances are deposited on the heat exchange surface to cause the increase of thermal resistance, the other action form is that the dirt part is peeled off the heat exchange surface to cause the decrease of thermal resistance under the scouring action of fluid, and the growth trend of the thermal resistance of the dirt is shown in figure 2. The method comprises the steps of adopting a progressive empirical model of the fouling resistance as an initial fouling resistance model of the boiler, updating the initial fouling resistance model into a target fouling resistance model by using a least square method according to the historical fouling resistance of the boiler, and then performing rolling optimization on parameters of the target fouling resistance model by using a particle swarm algorithm so as to obtain the target fouling resistance model which accords with the actual fouling resistance change of the boiler.
Preferably, the fouling thermal resistance model can be established based on a digital twin technology under the condition that a large amount of historical fouling thermal resistance is not relied on, and the operation data acquired in real time by the digital twin body is continuously and automatically evolved and updated based on a particle swarm algorithm.
Step 102, inputting a plurality of simulated scale thermal resistances related to the predicted time generated by the target scale thermal resistance model into a boiler simulation loop, and determining a plurality of simulated temperatures.
The predicted time refers to the post-cleaning run time for which the prediction of boiler fouling was made.
The simulated fouling resistance refers to the fouling resistance generated by a target fouling resistance model when boiler fouling prediction is performed.
The boiler simulation loop is used for reflecting the dynamic process of the simulated unit boiler in the operation process, and in the simulation process, the total heat transfer coefficient of the simulated unit boiler is changed by changing the dirt heat resistance, so that the simulation temperature with corresponding change can be obtained. The simulation temperature comprises state indexes such as smoke exhaust temperature, pipe wall temperature and the like which can represent the scaling fault degree, and only any temperature value is selected as the state index to carry out subsequent scaling prediction.
It can be understood that as the thermal resistance of the fouling changes with time, the heat exchange of the boiler water wall tube is affected, the heat transferred to the cooling water through the heating surface is reduced, the heat transfer efficiency is reduced, the exhaust gas temperature is increased, the heat absorbed from the exhaust gas cannot be quickly transferred to the cooling water through the tube wall, and the temperature of the tube wall is also increased and is in an overtemperature state. Therefore, the pipe wall temperature and the exhaust gas temperature can be used as state indexes for representing the scale fault degree.
In the embodiment of the invention, starting from the prediction time, generating a plurality of prediction times according to a preset time step, inputting all the prediction times into a corresponding target dirt thermal resistance model for operation, outputting a plurality of corresponding simulated dirt thermal resistances, and inputting all the simulated dirt thermal resistances into a boiler simulation loop for heat exchange process simulation to obtain a plurality of corresponding simulation temperatures.
It will be appreciated that for a heat transfer process of a boiler waterwall tube, the total heat transfer coefficient is the sum of the heat transfer coefficients of the individual series links that make up the entire heat transfer process, i.e., the total heat transfer coefficient is specifically:
wherein K is g R is the total heat transfer coefficient g R is the total heat resistance 0 Is smoke side heat resistance, R w Is the thermal resistance of the wall surface of the water-cooled wall, R 1 Is used for cooling the side thermal resistance of the working medium.
And 103, quantifying uncertainty factors of all simulation temperatures, and constructing a state transition model.
In the embodiment of the invention, modeling is carried out on a plurality of obtained simulation temperatures which change along with time in a differential model decomposition mode, and uncertainty factors existing in modeling are effectively quantified to construct a state transition model.
And 104, performing multi-step prediction according to the state transition model and a preset temperature failure limit value, and determining the failure moment of the boiler.
Temperature failure limit refers to a simulated temperature that characterizes the extent of boiler fouling failure to failure and requires cleaning.
The failure time refers to the predicted time when the simulated temperature of the boiler reaches the temperature failure limit value.
In the embodiment of the invention, the corresponding temperature failure limit value is obtained according to the data type of the selected simulation temperature, the multi-step prediction is carried out in a recursive mode by adopting a particle filtering algorithm according to a state transition model, the prediction process is continuously executed until the predicted value exceeds the temperature failure limit value, and the failure moment of the boiler is determined.
And 105, performing difference operation on the failure time and the predicted time, and outputting the residual service life of the boiler.
The remaining usable life RUL refers to the length of time from the predicted time to the failure time.
In the embodiment of the present invention, CL of RUL refers to the remaining available lifetime confidence interval, RUL prediction refers to the remaining available lifetime prediction value, accery limit refers to the upper and lower error limit, and group turn refers to the actual remaining available lifetime. As shown in fig. 3, the prediction operation is performed under the full life cycle, the cycle recursion is performed to obtain the failure time, the difference value operation is performed between the failure time and the prediction time, the remaining usable life of the boiler corresponding to the prediction time can be obtained, and the structural failure condition of the boiler can be determined according to the remaining usable life so as to arrange the maintenance time of the boiler.
In the embodiment of the invention, after a target fouling thermal resistance model is built on the basis of a least square method by adopting historical fouling thermal resistance of a boiler, updating the target fouling thermal resistance model in real time on the basis of a particle swarm algorithm, starting to generate a plurality of simulated fouling thermal resistances from a prediction time through the target fouling thermal resistance model, inputting all the simulated fouling thermal resistances into a boiler simulation loop to obtain corresponding simulation temperatures, carrying out uncertainty quantification on all the simulation temperatures to build a state transition model, carrying out multi-step prediction according to the state transition model and a preset temperature failure limit value to determine the failure time of the boiler, carrying out difference value operation on the failure time and the prediction time, and outputting the residual service life of the boiler. In the whole boiler scaling fault prediction process, a target scaling thermal resistance model is updated in real time according to actual scaling thermal resistance of a boiler, and multi-step prediction is performed after uncertainty of simulation temperature is effectively quantified based on differential modeling, so that accuracy of boiler structure fault prediction is improved.
Referring to fig. 4, fig. 4 is a flowchart illustrating a method for predicting boiler fouling failure according to a second embodiment of the present invention.
Step 401, constructing a target dirt thermal resistance model by adopting a historical dirt thermal resistance model and an initial dirt thermal resistance model of the boiler according to a least square method, and updating the target dirt thermal resistance model in real time based on a particle swarm algorithm.
Optionally, step 401 comprises the sub-steps of:
s1, acquiring a historical dirt thermal resistance and an initial dirt thermal resistance model of a boiler;
s2, constructing a loss function related to dirt thermal resistance based on a least square method, and adopting historical dirt thermal resistance to carry out equation solving on the loss function to determine a first thermal resistance parameter;
s3, updating an initial dirt thermal resistance model by adopting a first thermal resistance parameter to generate a target dirt thermal resistance model;
s3, acquiring actual dirt thermal resistance of the boiler and the operation time after cleaning of the actual dirt thermal resistance in real time;
s4, inputting a target dirt thermal resistance model by using the operation time after cleaning, and solving a corresponding theoretical dirt thermal resistance;
s5, updating the target dirt thermal resistance model based on a particle swarm algorithm when the thermal resistance difference between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference threshold value, and generating a new target dirt thermal resistance model.
In the embodiment of the invention, the initial fouling thermal resistance model is specifically: r is R f =f (τ, a, B) =a (1-exp (-bτ)); wherein R is f The scale thermal resistance is τ is the running time after cleaning, A is a first initial sub-thermal resistance parameter, and B is a second initial sub-thermal resistance parameter. It will be appreciated that a and B are parameters related to factors such as the structural design of the boiler, the quality of the cooling water, the speed of the cooling water flowing through the heat exchange surface, etc., which are inconvenient to measure directly in actual operation and need to be determined in order to be able to predict the fouling characteristics of the boiler.
The actual scale thermal resistance refers to the scale thermal resistance in the actual operation process of the boiler. Theoretical fouling resistance refers to the fouling resistance calculated in the target thermal resistance model corresponding to the actual fouling resistance.
The purpose of the least square method is to minimize the sum of square errors between theoretical and actual fouling thermal resistance values solved by the target fouling thermal resistance model, namely:wherein n is the total heat resistance of historical dirt, R i For the ith historical fouling resistance, θ is the thermal resistance parameter set { A, B }, ++>θ, which is the minimum sum of square errors. The construction of the loss function E (θ) with respect to the thermal resistance of fouling based on the least square method is specifically: />
It will be appreciated that the use of historical fouling thermal resistance to solve the loss function involves: and deriving the loss function with respect to the parameter vector, enabling the derivative to be zero to obtain an equation set, substituting the historical dirt thermal resistance into the equation set, and solving to obtain a first thermal resistance parameter.
The first thermal resistance parameter comprises a first sub-thermal resistance parameter A 0 And a second sub-thermal resistance parameter B 0 . After the historical dirt thermal resistance and the initial dirt thermal resistance model are adopted to conduct curve fitting by a least square method to determine a first thermal resistance parameter, the first initial sub-thermal resistance parameter is updated by adopting the first sub-thermal resistance parameter, and the second initial sub-thermal resistance parameter is updated by adopting the second sub-thermal resistance parameter, so that the initial dirt thermal resistance model is updated. The target dirt thermal resistance model determined by adopting the least square method and the historical dirt thermal resistance is a smooth curve, and specifically comprises the following steps: r is R f ＝A 0 (1-exp(-B 0 τ))。
Preferably, substep S5 comprises:
when the thermal resistance difference between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference threshold, constructing a neighborhood range by taking the first thermal resistance parameter as the center;
generating a particle swarm in a neighborhood range, and initializing the speed and the position of each initial particle in the particle swarm;
evaluating the fitness of each initial particle through a preset fitness function, taking the position of the initial particle with the minimum fitness in the particle swarm as a global optimal position, carrying out iterative updating on the speed and the position of each initial particle through a particle swarm algorithm, and taking the global optimal position when the iterative updating reaches the iteration stopping condition as a second thermal resistance parameter;
updating the target dirt thermal resistance model by adopting the second thermal resistance parameter to generate a new target dirt thermal resistance model.
The thermal resistance difference threshold refers to a difference threshold when the difference between the theoretical dirt thermal resistance and the actual dirt thermal resistance is in a normal range.
In the embodiment of the invention, after a target dirt thermal resistance model is determined by adopting a least square method, the actual dirt thermal resistance of the boiler and the post-cleaning running time related to the actual dirt thermal resistance are obtained in real time, and the post-cleaning running time related to the actual dirt thermal resistance is input into the target dirt thermal resistance model to calculate the corresponding theoretical dirt thermal resistance. As shown in fig. 5, the ordinate is the thermal resistance of the dirt, fit refers to the theoretical thermal resistance of the dirt, data refers to the actual thermal resistance of the dirt, the difference between the theoretical thermal resistance of the dirt and the actual thermal resistance of the dirt is calculated, when any difference exceeds a preset thermal resistance difference threshold, a particle swarm algorithm is adopted to further optimize the first thermal resistance parameter in order to further adapt to the change of the actual thermal resistance of the dirt of the boiler.
Respectively with a first thermal resistance parameter A 0 And B 0 Building a neighborhood range for the center [ A ] 0 -ε,A 0 +ε]And [ B ] 0 -ε,B 0 +ε]Where ε is the neighborhood radius, and is an arbitrary value greater than 0, set to create the initial range of particles. Randomly generating a certain number of initial particles in a neighborhood range to form a particle swarm, and initializing the speed V of each initial particle in the particle swarm j ＝(v Aj ,v Bj ) And position X j ＝(A j ,B j ) Wherein j= (1, 2, l, n), a j ∈[A 0 -ε,A 0 +ε]，B j ∈[B 0 -ε,B 0 +ε]J is the j-th initial particle, N is the total number of initial particles, A j The position of the jth initial particle, B, which is the first sub-thermal resistance parameter j The position, X, of the jth initial particle, which is a second sub-thermal resistance parameter j For the position of the jth initial particle, v Aj The speed, v, of the jth initial particle, which is a first sub-thermal resistance parameter Bj The speed of the jth initial particle, V, which is a second sub-thermal resistance parameter j Is the velocity of the jth initial particle.
And evaluating the fitness of each initial particle through a preset fitness function, wherein the fitness function is set as the sum of square errors of the thermal resistance of each cleaning theory dirt and the actual thermal resistance of the dirt, which are calculated by the initial particles. And taking the position of the initial particle with the minimum fitness in the particle swarm in the iteration process as a global optimal position, and taking the position corresponding to the minimum fitness obtained by calculating each initial particle in the iteration process as an individual historical optimal position. And determining a first individual optimal position of each initial particle and a first global optimal position of a particle group by calculating a first fitness of each initial particle, updating the speed and the position of the initial particle by adopting the first individual optimal position and the first global optimal position, and calculating a second fitness of each initial particle and the statistics iteration times. And traversing all the second fitness to determine a second individual optimal position of each initial particle and a second global optimal position of the particle swarm. Judging whether stopping conditions are met, if all the second fitness does not reach a preset fitness threshold value and the iteration times do not reach a preset time threshold value, taking the second fitness, the second individual optimal position and the second global optimal position as a new first fitness, a new first individual optimal position and a new first global optimal position, and jumping to execute the step of updating the speed and the position of the initial particle by adopting the first individual optimal position and the first global optimal position; and if all the second fitness reaches a preset fitness threshold or the iteration number reaches a preset number threshold, taking the second global optimal position as a second thermal resistance parameter.
The process of iteratively updating the speed and the position of each initial particle through a particle swarm algorithm comprises the following steps: v (V) j ＝wV j +c 1 r 1 (pbest j -X j )+c 2 r 2 (gbest-X j )；X j ＝X j +V j The method comprises the steps of carrying out a first treatment on the surface of the Wherein w is inertial weight, c 1 C is the first learning factor 2 R is the second learning factor 1 And r 2 Is distributed in [0,1 ]]Random number, pbest between j And as the individual historical optimal position of the jth initial particle, gbest is the global optimal position.
The second thermal resistance parameter comprises a third sub-thermal resistance parameter A best And a fourth sub-thermal resistance parameter B best Updating the first sub-thermal resistance parameter by using the third sub-thermal resistance parameter and updating the second sub-thermal resistance parameter by using the fourth sub-thermal resistance parameter so as to update the target dirt thermal resistance model determined by using the least square method, thereby generatingThe new target fouling thermal resistance model is specifically: r is R f ＝A best (1-exp(-B best τ)). If the difference between the theoretical dirt thermal resistance and the actual dirt thermal resistance exceeds the thermal resistance difference threshold again, the third sub-thermal resistance parameter A obtained by the previous round of optimization best And a fourth sub-thermal resistance parameter B best And constructing a neighborhood range for the center, and skipping to execute the step of updating the target dirt thermal resistance model based on the particle swarm algorithm to generate a new target dirt thermal resistance model.
It will be appreciated that the preset number of iterations threshold refers to the maximum number of iterations and the fitness threshold refers to the target value of fitness.
Step 402, obtaining a predicted time of the boiler, and sequentially adopting an associated target fouling thermal resistance model to operate according to a preset time step from the predicted time to sequentially generate a plurality of simulated fouling thermal resistances.
The time step refers to the unit duration of two adjacent predicted moments.
In the embodiment of the invention, after the predicted time of the boiler is obtained, a plurality of predicted times are built according to the preset time step from the predicted time, all the predicted times are sequentially input into the corresponding target dirt thermal resistance model for operation, and a plurality of corresponding simulated dirt thermal resistances are sequentially generated.
It can be understood that, because the target fouling thermal resistance model is in a real-time updated state based on the actual fouling thermal resistance and the theoretical fouling thermal resistance, different target fouling thermal resistance models before and after updating may be correspondingly updated according to different prediction moments when calculating a plurality of simulated fouling thermal resistances.
Step 403, sequentially inputting the simulated dirt thermal resistance into a boiler simulation loop of the boiler to perform heat exchange simulation, and outputting a plurality of corresponding simulation temperatures.
In the embodiment of the invention, all the simulated thermal resistances are sequentially input into a boiler simulation loop to perform dynamic simulation of the heat exchange process, and a plurality of corresponding simulation temperatures are output.
And step 404, quantifying uncertainty factors of all simulation temperatures, and constructing a state transition model.
Optionally, step 404 includes the sub-steps of:
carrying out data smoothing on all simulation temperatures by adopting a filter fitting method, and correspondingly generating a plurality of smoothing temperatures;
respectively carrying out difference operation on each simulation temperature and the corresponding smooth temperature to determine a plurality of first temperature differences;
performing variance operation on all the first temperature differences to determine a first variance, and constructing an observed Gaussian uncertainty meeting the first variance;
performing differential operation on all the smooth temperatures to obtain a plurality of differential temperatures, and performing curve fitting by adopting all the differential temperatures to generate a differential model;
respectively carrying out difference operation on each differential temperature and the differential fitting temperature correspondingly constructed through the differential model, and determining a plurality of second temperature differences;
performing variance operation on the basis of all second temperature differences to determine a second variance, and constructing a state Gaussian uncertainty meeting the second variance;
and carrying out differential modeling according to the observed Gaussian uncertainty, the differential model and the state Gaussian uncertainty to generate a state transition model.
Smoothing temperature refers to the simulated temperature after the data smoothing process.
The filter fitting method refers to the Savitzky-Golay algorithm, basically, in which data is fitted with a polynomial function in a sliding window, and then the data point at the center of the window is represented by the value of the polynomial function, in which the data can be smoothed by using the smoothness of the polynomial function, and the degree of smoothing can be controlled according to the order of the polynomial function and the window size.
The observed gaussian uncertainty refers to the observed uncertainty that satisfies a gaussian distribution with a mean of 0 and a first variance.
The state gaussian uncertainty refers to the state uncertainty that satisfies a gaussian distribution with a mean of 0 and a variance of a second variance.
In the embodiment of the invention, the filter fitting method is adopted to carry out data smoothing on all simulation temperatures, so as to eliminate uncertainty factors existing in each simulation temperature and obtain a plurality of smooth temperatures which still retain main characteristic information in the simulation temperatures. As shown in fig. 6, subtracting the corresponding smooth temperatures from each simulated temperature to obtain a plurality of first temperature differences, performing variance operation on all the first temperature differences to determine a first variance, and constructing a probability distribution for observing the gaussian uncertainty according to the first variance, as shown in fig. 7. As shown in fig. 8, the smooth temperatures at any two adjacent moments with a fixed interval are subjected to differential operation, so as to obtain a plurality of differential temperatures raw data with time-varying rules, and a corresponding differential model is constructed in a curve fitting manner. And calculating a difference fitting temperature fit data corresponding to each difference temperature by adopting a difference model, subtracting the corresponding difference fitting temperature from each difference temperature to determine a plurality of second temperature differences, performing variance operation on the basis of all the first temperature differences to determine a second variance, and constructing a state Gaussian uncertainty according to the second variance, as shown in fig. 9. And carrying out differential modeling on a differential rule adopted by approximate description based on the state variation quantity between any two adjacent moments according to the observed Gaussian uncertainty, the differential model and the state Gaussian uncertainty, and generating a state transition model.
The difference rule is specifically:
the state transition model comprises a state model and an observation model, wherein the state model specifically comprises the following components:
x t ＝f ′ (x t ,t-1)+x t-1 +ω t ；
the observation model specifically comprises the following steps:
y t ＝x t +v t ；
wherein t is the t-th predicted time, x t For the state vector of the simulated temperature at the predicted time t, x t-1 To simulate the state vector of the temperature at the predicted time t-1, f ′ () Is a differential model, omega t To observe Gaussian uncertainty, y t To predict the observed value of the simulated temperature in the t state vector at time, v t Is a state gaussian uncertainty.
And 405, performing multi-step prediction according to the state transition model and a preset temperature failure limit value, and determining the failure moment of the boiler.
Optionally, step 405 comprises the sub-steps of:
acquiring prior probability distribution corresponding to a state transition model and a preset temperature failure limit value of a boiler, and generating a sampling particle set by the prior probability distribution;
constructing posterior probability distribution by adopting a sampling particle set according to a Monte Carlo method;
carrying out weight normalization and particle resampling on the sampled particle sets, and determining target sampled particle sets;
performing multi-step prediction by adopting a target sampling particle set and posterior probability distribution, and outputting predicted temperature probability distribution;
calculating a plurality of failure probabilities of the target sampling particle set according to the temperature failure limit value and the predicted temperature probability distribution;
And calculating expected values based on all failure probabilities, and determining the failure moment of the boiler.
The predicted temperature probability distribution refers to a probability distribution in a multi-step prediction of the predicted simulated temperature.
Failure probability refers to the probability that the predicted simulated temperature exceeds the temperature failure limit. For example, the probability of failure at the predicted time t can be understood as the ratio of the number of particles at time t to the total number of particles reaching the temperature failure limit.
In the embodiment of the invention, the particle filtering algorithm adopts a Monte Carlo method to simulate the actual probability density, and a group of particle sets are used for representing the posterior probability distribution of states, wherein the historical optimal state vector sequence of the simulation temperature can be usedThe historical observation sequence of the simulation temperature can be expressed by y 1:t ＝{y 1 ,y 2 ,…,y t And } represents. Determining a corresponding prior probability distribution by means of a state transition model>Wherein (1)>Is the optimal state vector, y 1:t-1 The observation value of the simulation temperature from the 1 st predicted time to the t-1 st predicted time. Generating sample particle set from a priori probability distribution>Wherein k is the kth sampling particle, M is the total number of sampling particles, < + >>Sample particles being a sample particle group, +.>Is the weight of the sampled particle. Constructing posterior probability distribution by adopting sampling particle set according to Monte Carlo method Wherein->Wherein y is 1:t For the observation of the simulated temperature at the 1 st predicted time, δ () is a dirac function.
The dirac functional relation is specifically:
taking the state transition probability density distribution defined by the state model as an importance distribution, and obtaining the weight of the sampling particles as follows:wherein->Is a likelihood function of the observation process determined by the observation model.
Weight normalization is carried out on the sampling particle group:
the resampling mode is adopted to alleviate the particle degradation phenomenon existing in the particle filtering: in interval [0,1 ]]Taking a random number u, sequentially summing the weights of the sampling particles, wherein the sum of the weights of the current D-1 sampling particles is smaller than uAnd the sum of the weights of the first D sampling particles is equal to or greater than u +.>When D is E [0, N]The D-th particle is taken as a new sampling particle after resampling +.>Repeating the process for N times to obtain a resampled target sampling particle setAnd update the weight to +.>/>
The target sampling particle group and posterior probability distribution are adopted for multi-step prediction, the predicted temperature probability distribution is output, and the calculation process of the predicted temperature probability distribution specifically comprises the following steps:
wherein, the liquid crystal display device comprises a liquid crystal display device,a state vector of the simulated temperature after p steps, < >>For the predicted temperature probability distribution after p steps, < > >Representing the state transition process of a multi-step predictive process.
According to the temperature failure limit value and the predicted temperature probability distribution, calculating a plurality of failure probabilities corresponding to the target sampling particle set, wherein the calculation process of the failure probabilities specifically comprises the following steps:
wherein p is f F is the failure probability f Is a temperature failure limit.
And carrying out expected value calculation based on all failure probabilities and corresponding prediction moments in multi-step prediction, and determining an expected value calculation result as the failure moment of the boiler.
And 406, performing difference operation on the failure time and the predicted time, and outputting the residual service life of the boiler.
In the embodiment of the present invention, the implementation process of step 406 is similar to that of step 105, and will not be repeated here.
In the embodiment of the invention, after a target fouling thermal resistance model is built on the basis of a least square method by adopting historical fouling thermal resistance of a boiler, updating the target fouling thermal resistance model in real time on the basis of a particle swarm algorithm, starting to generate a plurality of simulated fouling thermal resistances from a prediction moment through the target fouling thermal resistance model, inputting all the simulated fouling thermal resistances into a boiler simulation loop to obtain corresponding simulation temperatures, quantifying uncertainty factors on the basis of all the simulation temperatures, executing differential modeling to build a state transition model, carrying out multi-step prediction to determine the failure moment of the boiler by adopting a particle filter algorithm according to the state transition model and a preset temperature failure limit value, carrying out difference operation on the failure moment and the prediction moment, and outputting the residual usable life of the boiler. In the whole boiler scaling fault prediction process, a target scaling thermal resistance model is updated in real time according to actual scaling thermal resistance of a boiler, and multi-step prediction is performed after uncertainty of simulation temperature is effectively quantified based on differential modeling, so that accuracy of boiler structure fault prediction is improved.
Referring to fig. 10, fig. 10 is a block diagram illustrating a boiler scale fault prediction apparatus according to an embodiment of the present invention.
A boiler fouling fault prediction apparatus comprising:
the target dirty thermal resistance model construction module 1001 is configured to construct a target dirty thermal resistance model according to a least square method by adopting a historical dirty thermal resistance and an initial dirty thermal resistance model of the boiler, and update the target dirty thermal resistance model in real time based on a particle swarm algorithm;
a simulation temperature generation module 1002, configured to input a plurality of simulated fouling thermal resistances associated with a prediction time generated by the target fouling thermal resistance model into a boiler simulation loop, and determine a plurality of simulation temperatures;
the state transition model construction module 1003 is configured to perform uncertainty factor quantification on all simulation temperatures, and construct a state transition model;
the failure moment determining module 1004 is configured to perform multi-step prediction according to the state transition model and a preset temperature failure limit value, and determine a failure moment of the boiler;
and the remaining usable life output module 1005 is used for performing difference operation on the failure time and the predicted time and outputting the remaining usable life of the boiler.
Optionally, the target fouling thermal resistance model building module 1001 includes:
the first data acquisition unit is used for acquiring a historical dirt thermal resistance and an initial dirt thermal resistance model of the boiler;
The first thermal resistance parameter determining unit is used for constructing a loss function related to dirt thermal resistance based on a least square method, carrying out equation solving on the loss function by adopting historical dirt thermal resistance, and determining a first thermal resistance parameter;
the target dirt thermal resistance model generating unit is used for updating the initial dirt thermal resistance model by adopting the first thermal resistance parameter to generate a target dirt thermal resistance model;
the second data acquisition unit is used for acquiring the actual dirt thermal resistance of the boiler and the operation time after cleaning of the actual dirt thermal resistance in real time;
the theoretical dirt thermal resistance calculation unit is used for inputting a target dirt thermal resistance model by adopting the running time after cleaning to solve the corresponding theoretical dirt thermal resistance;
the target dirt thermal resistance model updating unit is used for updating the target dirt thermal resistance model based on a particle swarm algorithm when the thermal resistance difference between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference threshold value, and generating a new target dirt thermal resistance model.
Optionally, the target fouling thermal resistance model updating unit is specifically configured to:
when the thermal resistance difference between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference threshold, constructing a neighborhood range by taking the first thermal resistance parameter as the center;
Generating a particle swarm in a neighborhood range, and initializing the speed and the position of each initial particle in the particles;
evaluating the fitness of each initial particle through a preset fitness function, taking the position of the initial particle with the minimum fitness in the particle swarm as a global optimal position, carrying out iterative updating on the speed and the position of each initial particle through a particle swarm algorithm, and taking the global optimal position when the iterative updating reaches the iteration stopping condition as a second thermal resistance parameter;
updating the target dirt thermal resistance model by adopting the second thermal resistance parameter to generate a new target dirt thermal resistance model.
Optionally, the simulation temperature generation module 1002 is specifically configured to:
acquiring the predicted time of the boiler, sequentially adopting an associated target dirt thermal resistance model to operate according to a preset time step from the predicted time, and sequentially generating a plurality of simulated dirt thermal resistances;
the simulated dirt thermal resistance is sequentially input into a boiler simulation loop of the boiler to perform heat exchange simulation, and a plurality of corresponding simulation temperatures are output.
Optionally, the state transition model building module 1003 is specifically configured to:
carrying out data smoothing on all simulation temperatures by adopting a filter fitting method, and correspondingly generating a plurality of smoothing temperatures;
respectively carrying out difference operation on each simulation temperature and the corresponding smooth temperature to determine a plurality of first temperature differences;
Performing variance operation on all the first temperature differences to determine a first variance, and constructing an observed Gaussian uncertainty meeting the first variance;
performing differential operation on all the smooth temperatures to obtain a plurality of differential temperatures, and performing curve fitting by adopting all the differential temperatures to generate a differential model;
respectively carrying out difference operation on each differential temperature and the differential fitting temperature correspondingly constructed through the differential model, and determining a plurality of second temperature differences;
performing variance operation on the basis of all second temperature differences to determine a second variance, and constructing a state Gaussian uncertainty meeting the second variance;
and carrying out differential modeling according to the observed Gaussian uncertainty, the differential model and the state Gaussian uncertainty to generate a state transition model.
Optionally, the failure moment determining module 1004 is specifically configured to:
acquiring prior probability distribution corresponding to a state transition model and a preset temperature failure limit value of a boiler, and generating a sampling particle set by the prior probability distribution;
constructing posterior probability distribution by adopting a sampling particle set according to a Monte Carlo method;
carrying out weight normalization and particle resampling on the sampled particle sets, and determining target sampled particle sets;
performing multi-step prediction by adopting a target sampling particle set and posterior probability distribution, and outputting predicted temperature probability distribution;
Calculating a plurality of failure probabilities of the target sampling particle set according to the temperature failure limit value and the predicted temperature probability distribution;
and calculating expected values based on all failure probabilities, and determining the failure moment of the boiler.
The embodiment of the application also provides electronic equipment, which comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the boiler scale fault prediction method according to any embodiment of the application.
Embodiments of the present application also provide a computer readable storage medium having a computer program stored thereon, which when executed implements a boiler fouling fault prediction method according to any of the embodiments of the present application.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.
In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform 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, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for predicting boiler fouling failure, comprising:
constructing a target dirt thermal resistance model by adopting a historical dirt thermal resistance model and an initial dirt thermal resistance model of the boiler according to a least square method, and updating the target dirt thermal resistance model in real time based on a particle swarm algorithm;
inputting a plurality of simulated dirt thermal resistances related to the prediction time generated by the target dirt thermal resistance model into a boiler simulation loop, and determining a plurality of simulated temperatures;
carrying out uncertainty factor quantification on all the simulation temperatures, and constructing a state transition model;
performing multi-step prediction according to the state transition model and a preset temperature failure limit value, and determining the failure moment of the boiler;
And carrying out difference value operation on the failure moment and the predicted moment, and outputting the residual service life of the boiler.
2. The boiler scale fault prediction method according to claim 1, wherein the step of constructing a target scale thermal resistance model by using a historical scale thermal resistance and an initial scale thermal resistance model of a boiler according to a least square method and updating the target scale thermal resistance model in real time based on a particle swarm algorithm comprises the steps of:
acquiring a historical dirt thermal resistance and an initial dirt thermal resistance model of the boiler;
constructing a loss function related to dirt thermal resistance based on a least square method, and adopting the historical dirt thermal resistance to carry out equation solving on the loss function to determine a first thermal resistance parameter;
updating the initial dirt thermal resistance model by adopting the first thermal resistance parameter to generate a target dirt thermal resistance model;
acquiring actual dirt thermal resistance of a boiler and the operation time of the actual dirt thermal resistance after cleaning in real time;
inputting the target dirt thermal resistance model by adopting the operation time after cleaning, and solving the corresponding theoretical dirt thermal resistance;
when the thermal resistance difference value between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference value threshold, updating the target dirt thermal resistance model based on a particle swarm algorithm, and generating a new target dirt thermal resistance model.
3. The method of claim 2, wherein the step of updating the target fouling thermal resistance model based on a particle swarm algorithm to generate a new target fouling thermal resistance model when a thermal resistance difference between any one of the theoretical fouling thermal resistances and a corresponding actual fouling thermal resistance exceeds a preset thermal resistance difference threshold value comprises:
when the thermal resistance difference value between any theoretical dirt thermal resistance and the corresponding actual dirt thermal resistance exceeds a preset thermal resistance difference value threshold, constructing a neighborhood range by taking the first thermal resistance parameter as the center;
generating a particle swarm in the neighborhood range, and initializing the speed and the position of each initial particle in the particles;
evaluating the fitness of each initial particle through a preset fitness function, taking the position of the initial particle with the minimum fitness in the particle swarm as a global optimal position, carrying out iterative updating on the speed and the position of each initial particle through a particle swarm algorithm, and taking the global optimal position when the iterative updating meets the iteration stopping condition as a second thermal resistance parameter;
updating the target dirt thermal resistance model by adopting the second thermal resistance parameter to generate a new target dirt thermal resistance model.
4. The boiler fouling fault prediction method according to claim 1, characterized in that the step of inputting a plurality of simulated fouling resistances associated with the predicted time generated by the target fouling resistance model into a boiler simulation loop, determining a plurality of simulated temperatures, comprises:
acquiring the predicted time of the boiler, sequentially adopting an associated target dirt thermal resistance model to operate according to a preset time step from the predicted time, and sequentially generating a plurality of simulated dirt thermal resistances;
and sequentially inputting the simulated dirt thermal resistance into a boiler simulation loop of the boiler to perform heat exchange simulation, and outputting a plurality of corresponding simulation temperatures.
5. The method for predicting boiler fouling failure according to claim 1, wherein said step of quantifying uncertainty factors for all of said simulated temperatures, constructing a state transition model, comprises:
carrying out data smoothing on all the simulation temperatures by adopting a filter fitting method, and correspondingly generating a plurality of smoothing temperatures;
respectively carrying out difference operation on each simulation temperature and the corresponding smooth temperature to determine a plurality of first temperature differences;
performing variance operation on all the first temperature differences to determine a first variance, and constructing an observed Gaussian uncertainty meeting the first variance;
Performing differential operation on all the smooth temperatures to obtain a plurality of differential temperatures, and performing curve fitting by adopting all the differential temperatures to generate a differential model;
respectively carrying out difference operation on each differential temperature and the differential fitting temperature correspondingly constructed through the differential model, and determining a plurality of second temperature differences;
performing variance operation on the basis of all the second temperature differences to determine a second variance, and constructing a state Gaussian uncertainty meeting the second variance;
and carrying out differential modeling according to the observed Gaussian uncertainty, the differential model and the state Gaussian uncertainty to generate a state transition model.
6. The boiler fouling failure prediction method according to claim 1, wherein the step of performing multi-step prediction according to the state transition model and a preset temperature failure limit value, and determining a failure time of the boiler comprises:
acquiring prior probability distribution corresponding to the state transition model and a preset temperature failure limit value of the boiler, and generating a sampling particle set by the prior probability distribution;
constructing posterior probability distribution by adopting the sampling particle set according to a Monte Carlo method;
carrying out weight normalization and particle resampling on the sampled particle sets to determine target sampled particle sets;
Performing multi-step prediction by adopting the target sampling particle set and the posterior probability distribution, and outputting predicted temperature probability distribution;
calculating a plurality of failure probabilities of the target sample particle set according to the temperature failure limit value and the predicted temperature probability distribution;
and calculating expected values based on all the failure probabilities, and determining the failure moment of the boiler.
7. A boiler fouling fault prediction apparatus, comprising:
the target dirt thermal resistance model construction module is used for constructing a target dirt thermal resistance model according to a least square method by adopting a historical dirt thermal resistance model and an initial dirt thermal resistance model of the boiler, and updating the target dirt thermal resistance model in real time based on a particle swarm algorithm;
the simulated temperature generation module is used for inputting a plurality of simulated dirt thermal resistances related to the predicted time generated by the target dirt thermal resistance model into a boiler simulation loop to determine a plurality of simulated temperatures;
the state transition model construction module is used for quantifying uncertainty factors of all the simulation temperatures and constructing a state transition model;
the failure moment determining module is used for carrying out multi-step prediction according to the state transition model and a preset temperature failure limit value, and determining the failure moment of the boiler;
And the remaining usable life output module is used for carrying out difference value operation on the failure moment and the predicted moment and outputting the remaining usable life of the boiler.
8. The boiler fouling fault prediction device of claim 7, wherein the simulated temperature generation module is specifically configured to:
acquiring the predicted time of the boiler, sequentially adopting an associated target dirt thermal resistance model to operate according to a preset time step from the predicted time, and sequentially generating a plurality of simulated dirt thermal resistances;
and sequentially inputting the simulated dirt thermal resistance into a boiler simulation loop of the boiler to perform heat exchange simulation, and outputting a plurality of corresponding simulation temperatures.
9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the boiler fouling fault prediction method according to any one of claims 1-6.
10. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed implements the boiler fouling fault prediction method according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310854521.2A CN116738862B (en) | 2023-07-12 | 2023-07-12 | Boiler scaling fault prediction method, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310854521.2A CN116738862B (en) | 2023-07-12 | 2023-07-12 | Boiler scaling fault prediction method, device, equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116738862A true CN116738862A (en) | 2023-09-12 |
| CN116738862B CN116738862B (en) | 2024-05-31 |
Family
ID=87901190
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310854521.2A Active CN116738862B (en) | 2023-07-12 | 2023-07-12 | Boiler scaling fault prediction method, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116738862B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117272844A (en) * | 2023-11-22 | 2023-12-22 | 武汉宏联电线电缆有限公司 | Method and system for predicting service life of distribution board |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133460A (en) * | 2017-04-26 | 2017-09-05 | 中国能源建设集团广东省电力设计研究院有限公司 | A kind of online dynamic prediction method of boiler flyash carbon content |
| US20200191380A1 (en) * | 2018-12-17 | 2020-06-18 | Doosan Heavy Industries & Construction Co., Ltd. | System and method for configuring boiler combustion model |
| CN112487694A (en) * | 2020-11-29 | 2021-03-12 | 西北工业大学 | Complex equipment residual life prediction method based on multiple degradation indexes |
| CN113011010A (en) * | 2021-03-01 | 2021-06-22 | 常州英集动力科技有限公司 | Boiler fault diagnosis method and diagnosis system based on structural mechanism and operation data |
| CN114154419A (en) * | 2021-12-06 | 2022-03-08 | 华润电力技术研究院有限公司 | Boiler optimal oxygen amount prediction method and system |
| CN114429003A (en) * | 2021-12-15 | 2022-05-03 | 国家能源集团新能源技术研究院有限公司 | System of boiler four-tube service life prediction method |
| CN114912697A (en) * | 2022-05-26 | 2022-08-16 | 南方电网电力科技股份有限公司 | Boiler slagging degree prediction method based on PSO algorithm and related device |
| CN115291513A (en) * | 2022-07-15 | 2022-11-04 | 国网河北能源技术服务有限公司 | Boiler reheat steam temperature prediction method and device, terminal equipment and storage medium |
| CN116046659A (en) * | 2023-01-28 | 2023-05-02 | 山东蓝想环境科技股份有限公司 | Novel intelligent monitoring heat exchanger and state monitoring method thereof |
-
2023
- 2023-07-12 CN CN202310854521.2A patent/CN116738862B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133460A (en) * | 2017-04-26 | 2017-09-05 | 中国能源建设集团广东省电力设计研究院有限公司 | A kind of online dynamic prediction method of boiler flyash carbon content |
| US20200191380A1 (en) * | 2018-12-17 | 2020-06-18 | Doosan Heavy Industries & Construction Co., Ltd. | System and method for configuring boiler combustion model |
| CN112487694A (en) * | 2020-11-29 | 2021-03-12 | 西北工业大学 | Complex equipment residual life prediction method based on multiple degradation indexes |
| CN113011010A (en) * | 2021-03-01 | 2021-06-22 | 常州英集动力科技有限公司 | Boiler fault diagnosis method and diagnosis system based on structural mechanism and operation data |
| CN114154419A (en) * | 2021-12-06 | 2022-03-08 | 华润电力技术研究院有限公司 | Boiler optimal oxygen amount prediction method and system |
| CN114429003A (en) * | 2021-12-15 | 2022-05-03 | 国家能源集团新能源技术研究院有限公司 | System of boiler four-tube service life prediction method |
| CN114912697A (en) * | 2022-05-26 | 2022-08-16 | 南方电网电力科技股份有限公司 | Boiler slagging degree prediction method based on PSO algorithm and related device |
| CN115291513A (en) * | 2022-07-15 | 2022-11-04 | 国网河北能源技术服务有限公司 | Boiler reheat steam temperature prediction method and device, terminal equipment and storage medium |
| CN116046659A (en) * | 2023-01-28 | 2023-05-02 | 山东蓝想环境科技股份有限公司 | Novel intelligent monitoring heat exchanger and state monitoring method thereof |
Non-Patent Citations (5)
| Title |
|---|
| ZHANG, YIZHUO: "Optimization of Coal-fired Boiler on LS-SVM Model and PSO Algorithms", PROCEEDINGS OF THE 2015 6TH INTERNATIONAL CONFERENCE ON MANUFACTURING SCIENCE AND ENGINEERING, vol. 32, 14 December 2016 (2016-12-14), pages 329 - 334 * |
| 冯殿义;范佳;孙守仁;: "废热锅炉污垢热阻实时预测研究", 热能动力工程, no. 02, 20 March 2009 (2009-03-20), pages 196 - 199 * |
| 刘飞明;张雨飞;: "采用改进混沌粒子群算法的锅炉NO_x排放的LSSVM回归建模", 工业控制计算机, no. 01, 25 January 2017 (2017-01-25), pages 75 - 76 * |
| 李德波等: "改进粒子群算法在燃煤电厂中的应用现状及展望", 环境工程, vol. 41, no. 01, 30 May 2023 (2023-05-30), pages 354 - 362 * |
| 谢吉伟;刘君强;王小磊;: "应用交互式多模型算法的设备剩余寿命预测", 空军工程大学学报(自然科学版), no. 02, 25 April 2016 (2016-04-25), pages 98 - 102 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117272844A (en) * | 2023-11-22 | 2023-12-22 | 武汉宏联电线电缆有限公司 | Method and system for predicting service life of distribution board |
| CN117272844B (en) * | 2023-11-22 | 2024-04-09 | 武汉宏联电线电缆有限公司 | Method and system for predicting service life of distribution board |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116738862B (en) | 2024-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116738862B (en) | Boiler scaling fault prediction method, device, equipment and medium | |
| CN112818595A (en) | Method and system for correcting digital twin model data of evaporation zone of thermal power plant | |
| CN111461551B (en) | Deep learning and SPC criterion-based electric submersible pump fault early warning method | |
| US20150269293A1 (en) | Diagnostic model generating apparatus and method, and abnormality diagnostic apparatus | |
| CN108961460B (en) | Fault prediction method and device based on sparse ESGP (Enterprise service gateway) and multi-objective optimization | |
| CN112116069A (en) | Attention-LSTM-based reinforcement learning Agent knowledge inference method | |
| CN111523727B (en) | Method for predicting remaining life of battery by considering recovery effect based on uncertain process | |
| CN109145373B (en) | Residual life prediction method and device based on improved ESGP and prediction interval | |
| CN116048028A (en) | Technological parameter optimization method based on reinforcement learning | |
| CN108492026B (en) | Soft measurement method based on integrated orthogonal component optimization regression analysis | |
| CN114492675A (en) | Intelligent fault cause diagnosis method for capacitor voltage transformer | |
| CN115421216A (en) | STL-ARIMA-NAR mixed model-based medium-and-long-term monthly rainfall forecasting method | |
| CN114429003A (en) | System of boiler four-tube service life prediction method | |
| CN109242142B (en) | Space-time prediction model parameter optimization method for infrastructure network | |
| CN107688862B (en) | Insulator equivalent salt deposit density accumulation rate prediction method based on BA-GRNN | |
| JP5582487B2 (en) | Process state prediction method | |
| JP4498666B2 (en) | Prediction device, prediction program, and recording medium | |
| CN114648178B (en) | Operation and maintenance strategy optimization method of electric energy metering device based on DDPG algorithm | |
| CN115481568A (en) | Bearing life prediction method based on self-adaptive model particle filter algorithm | |
| CN113743022B (en) | Storage and visualization method for high-precision climate change data | |
| CN113884936B (en) | ISSA coupling DELM-based lithium ion battery health state prediction method | |
| CN111985146B (en) | Method and device for determining measure potential well | |
| JP4461245B2 (en) | Stock price fluctuation analysis system | |
| CN113837443A (en) | Transformer substation line load prediction method based on depth BilSTM | |
| KR101549349B1 (en) | Method for building energy modeling based on data driven model and Method and system for optimal energy management and control |
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 |