CN113198591A - Roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation - Google Patents
Roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation Download PDFInfo
- Publication number
- CN113198591A CN113198591A CN202110532920.8A CN202110532920A CN113198591A CN 113198591 A CN113198591 A CN 113198591A CN 202110532920 A CN202110532920 A CN 202110532920A CN 113198591 A CN113198591 A CN 113198591A
- Authority
- CN
- China
- Prior art keywords
- module
- control
- data
- current
- mpc
- 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
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B02—CRUSHING, PULVERISING, OR DISINTEGRATING; PREPARATORY TREATMENT OF GRAIN FOR MILLING
- B02C—CRUSHING, PULVERISING, OR DISINTEGRATING IN GENERAL; MILLING GRAIN
- B02C15/00—Disintegrating by milling members in the form of rollers or balls co-operating with rings or discs
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B02—CRUSHING, PULVERISING, OR DISINTEGRATING; PREPARATORY TREATMENT OF GRAIN FOR MILLING
- B02C—CRUSHING, PULVERISING, OR DISINTEGRATING IN GENERAL; MILLING GRAIN
- B02C25/00—Control arrangements specially adapted for crushing or disintegrating
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Food Science & Technology (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
Abstract
The invention provides a roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation, which comprises an OPC SERVER module, an OPC module, an MPC algorithm module, an HTTP module, a database module and a WEB module, wherein the OPC SERVER module is used for acquiring a rolling time domain signal; the OPC module is responsible for reading and writing DCS data in the OPC SERVER, filtering the DCS data and outputting the DCS data to the MPC algorithm module, and the OPC module also receives the data output by the algorithm module and writes the data into the OPC SERVER module to finish the issuing operation of the control quantity; the MPC algorithm module processes the data after acquiring the data, and splices and transmits the input quantity and the output quantity after processing; after receiving the data of the MPC algorithm module, the HTTP module sends the data to the HTTP SERVER module, and the HTTP SERVER module calls a database interface to store the data in a time sequence database; and the WEB module acquires the numerical value of the time sequence database from time to time and displays the numerical value. The invention has the advantages that: the production line is integrally stable, and the machine hour yield and the product quality are obviously improved; the unit energy consumption is obviously reduced.
Description
Technical Field
The invention relates to a control method of a roller type vertical mill, in particular to a self-adaptive prediction control system of the roller type vertical mill.
Background
The roller vertical mill is composed of a set of milling devices (i.e. a milling roller and a grinding disc), and materials are milled into powder between the milling roller and the grinding disc. The grinding device is driven by the grinding disc to rotate and drive the grinding rollers correspondingly, and the grinding pressure is mainly used for pressurizing the grinding disc materials by a hydraulic device except for the self weight of the grinding rollers. The ground material is separated by a matched powder selecting machine, the coarse powder returns to a grinding disc for grinding again, and the qualified finished product is sent to a finished product storage bin by a dust collecting and conveying system. The vertical roller mill integrates fine crushing, drying, grinding, powder selecting and conveying, has the advantages of high grinding efficiency, large drying capacity, easy adjustment of product fineness, low noise, low power consumption, simple process flow, low abrasion, low operating cost and the like, and is widely applied to the grinding field of large-scale industrial production lines such as building materials, metallurgy and the like.
A middle speed vertical mill for grinding petroleum coke as disclosed in patent application No. 201410056083.6, including grinder, grinder includes grinding roller subassembly and mill dish subassembly, the grinding roller subassembly includes grinding roller and grinding roller shell, the mill dish subassembly is equipped with the welt, wherein: the width of the grinding roller sleeve is increased by at least 10mm on the original basis, the thickness of the lining plate is increased by at least 20mm, and the wrap angle of the grinding roller is reduced to at least 38.45 degrees. According to the invention, on the basis of a conventional ZGM medium-speed coal mill, a grinding curve is optimized, the grinding efficiency is improved, a hydraulic oil pipeline system and a rotary separator are optimized through a hydraulic variable loading system, and a water spraying device is arranged, so that the operation stability of a medium-speed vertical mill for grinding petroleum coke is ensured.
The roller type vertical mill grinding system is a multivariable strong coupling system. The grinding process of the roller type vertical mill is a pure physical (mechanical) process, and key process parameters such as pressure of a grinding roller, pressure difference of the mill, thickness of a material layer, circulating air quantity, outlet temperature, current of a main motor, current of an external circulation elevator, vibration of the mill and the like exist in an optimal operation interval. The roller type vertical mill can only be operated in an optimal operation range, unit power consumption and equipment loss can be minimized, and the operation efficiency of the mill can be maximized.
However, at present, although the industrial production line has widely adopted the DCS/PLC control system, it is limited to centralized monitoring of the production process and does not exert the potential of computer automatic control, the operation is still maintained in a manual operation state, and it is necessary to rely on human experience and ability, there are phenomena that the operation methods are difficult to unify and standardize, the operation results of the operators are uneven, the operation level fluctuation of each shift is large, the excellent operators often cannot keep the operation within the optimal operation interval for a long time, the power consumption of the system unit is still large different from the international advanced level, and the improvement and improvement of economic indicators such as product quality and energy consumption are realized.
The roller type vertical mill system is a complex system with characteristics of multivariable, large hysteresis, large inertia, nonlinearity, time variation and the like, and the traditional control means cannot effectively solve the control problems due to the limitation of the traditional control means. In general, model predictive control uses a linear time-invariant dynamic model for prediction, however, in practice, due to the nonlinearity of an actuator, it is difficult to accurately predict the linear time-invariant model.
The system control variables are: feeding amount, water spraying amount, grinding roller pressure, circulating air quantity, rotation number of the powder concentrator, opening degree of a pipeline valve and the like; the controlled variables are: the temperature of the outlet of the mill, the fineness of the product, the thickness of a material layer, the pressure difference of the mill, the external circulation quantity, the temperature of the inlet of the mill and the like, and the core controlled parameter is the current of a main mill motor.
The roller type vertical mill system has strong coupling: the feeding amount directly influences the thickness of a material layer, the pressure difference of the mill, the current of a main motor and the current of an external circulation hoister; the water spraying quantity directly influences the outlet temperature of the mill, the thickness of a material layer, the current of a main motor and the pressure difference of the mill; the rotation speed of the powder concentrator and the air volume of the system directly influence the pressure difference of the mill, the fineness of products, the vibration of the mill, the current of an external circulation hoister and the like.
Industry experts and traditional PID control methods: (1) the feeding amount of the mill is adjusted, the material layer of the mill is kept stable, and the current of the mill, the current of a circulating bucket elevator, the vibration of the mill and the differential pressure of the mill are controlled within a set range. (2) The water spraying and grinding pressure of the mill are adjusted, and the stability of the vibration of the mill is ensured. (3) The rotation speed of the powder concentrator and the rotation speed of the exhaust fan are adjusted to ensure the optimization of the quality (fineness) of the raw materials discharged from the mill. (4) The cold and hot air valves and the circulating air valve of the mill are adjusted to ensure the stable temperature of the outlet of the mill and keep the grinding negative pressure to fluctuate within a certain range. Because the PID automatic control strategy and method are rough and simple and are single-input single-output control, various difficulties are encountered when the automatic control is implemented, the automatic control operation rate, maintainability and transportability are still unsatisfactory, and the automatic control cannot be realized.
The patent with the application number of 201910619419.8 discloses a cement raw material vertical mill raw material fineness index prediction method based on a convolutional neural network, which is characterized by comprising the following steps of: step 1: selecting 8 input variables related to the fineness of the raw materials, carrying out normalization processing on the selected variable data, constructing 8 variable time sequence input layers, and simultaneously carrying out time sequence processing on the normalized variable data; step 2: performing convolution pooling and full-connection operation on input variable data, performing convolution operation on the input data, pooling the input data subjected to the convolution operation, performing full-connection operation on output data after multiple times of convolution pooling, and performing Droupout layer processing to complete a forward training process of a convolutional neural network prediction model; and step 3: the convolutional neural network model updates weight parameters by adopting a back propagation technology to improve the prediction precision of the fineness of the raw material, independent adaptive learning rates are designed for different parameters by calculating the first moment estimation and the second moment estimation of the gradient, the weight and the offset of the convolutional layer are updated, fine adjustment of the parameters of the network is completed, the prediction error of the fineness of the raw material model is smaller than a set threshold value, and training of the convolutional neural network model is completed; and 4, step 4: and (4) predicting the fineness index of the raw material of the cement raw material vertical mill in real time by using the trained CNN model. The technical scheme of the patent application is based on a convolutional neural network, the controlled target quantity is the fineness of raw materials, the neural network algorithm needs to be modeled based on a large amount of data, the technical scheme focuses on the realization of the neural network algorithm, and the control of a grinding host machine is not provided at all.
Disclosure of Invention
The technical problem to be solved by the invention is how to improve the prediction accuracy of the roller type vertical mill by realizing stable automatic control.
The invention solves the technical problems through the following technical means: the self-adaptive prediction control system of the roller vertical mill based on the rolling time domain estimation comprises an OPC SERVER module, an OPC module, an MPC algorithm module, an HTTP module, a database module and a WEB module;
the OPC module is responsible for reading and writing OPC SERVER data, reading DCS data on the OPC SERVER, performing filtering processing on the DCS data and outputting the DCS data to the MPC algorithm module, and meanwhile, the OPC module also receives the data output by the algorithm module and writes the data into the OPC SERVER module to finish the issuing operation of the control quantity;
after the MPC algorithm module acquires data, firstly, judging which mode the MPC algorithm module is in at present, if the MPC algorithm module is in a manual mode, the MPC algorithm module does not analyze and process the data, and the algorithm module splices and sends the data input by a mill operator and key technical parameters of the roller type vertical mill; if the mode is the automatic mode, the MPC algorithm module firstly judges whether the current of the hoist is out of limit, if the current of the hoist is out of limit, the MPC algorithm module enters a current control algorithm of the hoist for processing, and simultaneously judges whether the pressure difference is out of limit, if the pressure difference is out of limit, the MPC algorithm module enters a pressure difference control algorithm for processing; if the values of the current and the pressure difference of the hoister are normal, entering a mill host current control algorithm and a water spray flow control algorithm, realizing self-adaptive prediction control by using an MPC algorithm module by using a rolling time domain estimation-based method, optimizing and adjusting parameters and states by using dynamic optimization and backward time measurement vision by using rolling time domain estimation, carrying out numerical convergence on the dynamic optimization problem by using a nonlinear programming solver, updating object model parameters by using a self-adaptive MPC controller in each control interval, keeping the model parameters unchanged in a prediction range once the model parameters are updated, and splicing and sending input quantity and output quantity after algorithm processing;
after receiving the data of the MPC algorithm module, the HTTP module sends the data to the HTTP SERVER module, and the HTTP SERVER module calls a database interface to store the data in a time sequence database;
and the WEB module can acquire the numerical value of the time sequence database and display the numerical value.
As a further optimized technical solution, the DCS data includes: the method comprises the following steps of (1) grinding host current feedback, a grinding host current target value, lifting machine current feedback, a lifting machine current threshold value, a manual-automatic mark, feeding amount setting, automatic water spraying flow setting, water spraying control valve feedback, automatic water spraying valve setting, water spraying flow feedback, a water-material ratio, a lifting machine current dead zone, a lifting machine current feedback value, a raw material grinding feeding amount automatic setting value, a lifting machine current upper limit, feeding amount feedback, a current-located differential pressure interval, a differential pressure dead zone, differential pressure different interval limit values, a powder concentrator rotating speed automatic setting value, a powder concentrator reference value and a grinding machine pressure difference value;
the data written into the OPC SERVER module comprises: heartbeat data and feeding quantity set values.
As a further optimized technical scheme, the OPC module encapsulates an OPC CLIENT, a subscribe function block and a PUBLISH function block, wherein any number of read-write data can be configured in the OPC CLIENT, the OPC CLIENT reads DCS data from the OPC SERVER at regular time and sends the DCS data to the algorithm module for processing, in the data configured in the OPC CLIENT, the read-write operation of the OPC module is triggered as long as the data changes, after the OPC CLIENT reads the data, the DCS CLIENT sends the data to the next module through the PUBLISH function block for processing, and the subscribe function block receives the data output by the algorithm module and writes the data into the OPC SERVER module, thereby completing the issuing operation of the control quantity.
As a further optimized technical scheme, the model predictive control algorithm comprises the following steps:
step 1: starting;
step 2: waiting for a periodic event trigger;
and step 3: estimating and optimizing model parameters in a rolling time domain;
and 4, step 4: updating MPC controller model parameters;
and 5: MPC control solving an optimal solution;
step 6: the optimal solution is delivered to an actuator, whether the cycle times reach the preset times is judged, if not, the model collection step 2 is carried out, otherwise, the step 7 is carried out;
and 7: and (6) ending.
As a further optimization technical scheme, a model prediction algorithm target function of the MPC algorithm module is an L1 norm, a set value of a controlled quantity designates an upper limit SPHI and a lower limit SPLO, a middle range of the upper limit and the lower limit is called a dead zone, and the formula is as follows:
minυ＝whiehi+wloelo+wyy+wuu+wΔuΔu
in which upsilon is the objective function, whiTo penalize the weight for the upper bound, ehiIs an upper bound error value, wloTo penalize the weight lower, eloThe lower error value, wyIs the controlled quantity control weight, y is the predicted value of the controlled quantity model, wuIs a control weight of the control quantity, u is the control quantity, wΔuChanging penalty weight for the control quantity, and delta u is the change of the control quantity;
increase whiCan inhibit the controlled quantity from exceeding the target value upper limit and increase wloCan inhibit the controlled quantity from falling below the lower limit of the target value, wyIn order to be positive to suppress the change in the controlled quantity and negative to promote the change in the controlled quantity, wuIn order to positively suppress the change in the controlled variable and to negatively promote the change in the controlled variable, wΔuThe increase can suppress the variation of the control amount in a single control period, and the objective function can individually control each amount in the model by adjusting the weight.
As a further optimized technical scheme, the grinding host machine current control algorithm strategy is as follows:
the feeding quantity and the current of the mill host have obvious positive correlation, the control quantity is the feeding quantity, the controlled quantity is the current of the mill host, step response is carried out, a response curve of data is checked, corresponding K, TAU and THETA are calculated, model parameters are assigned to initial values in an MPC algorithm module, the current of the mill host can be continuously changed along with working conditions in the production process of the roller type vertical mill, the MPC algorithm module calculates the feeding quantity in real time according to the feedback value of the current of the mill host and sends control to realize the stability of the current of the mill host, and in the control process of the MPC algorithm module, the automatic feeding quantity set value, feeding quantity feedback, mill host current feedback and mill host current set value splicing and sending to an HTTP module are carried out, so that subsequent warehousing, display and debugging are facilitated.
As a further optimized technical scheme, the water spraying flow control algorithm strategy is as follows:
the water spray volume can be changed proportionally along with the change of the feeding volume, the size of the water spray volume is adjusted by controlling the opening degree of a water spray valve, step response is carried out, the response curve of data is checked, corresponding K, TAU and THETA are calculated, initial values are given to model parameters in an MPC algorithm module, the automatic water spray volume setting quantity, the water spray control valve feedback, the water spray valve automatic setting, the water spray volume feedback and the water-material ratio value are spliced and sent to an HTTP module in the control process of the algorithm module, and subsequent storage and display debugging are facilitated.
As a further optimized technical scheme, the elevator current control algorithm strategy is as follows:
when the current of the hoister exceeds a set threshold value, the current control module of the main grinder and the water spray flow control module are stopped, the feeding amount is directly reduced, then the numerical value change of the current of the hoister is continuously monitored until the current value of the hoister is restored to a normal value, and then the current control module of the main grinder and the water spray flow control module are started.
As a further optimized technical scheme, the control strategy of the pressure difference control algorithm is as follows:
the pressure difference control strategy is to divide the pressure difference value into 4 intervals, directly quit the automatic control and switch to the manual mode when the pressure difference is in the zone1, and carry out the emergency treatment by the manual intervention of the milling operator, when the pressure difference is in the zone2, the rotating speed of the powder machine is reduced by 2 steps on the basis of the default rotating speed, when the pressure difference is in the zone3, the rotating speed of the powder machine is reduced by 1 step on the basis of the default rotating speed, and when the pressure difference is in the zone4, the rotating speed of the powder selecting machine is set as the default value. The pressure difference control module monitors the pressure difference value in real time, and adjusts the rotating speed of the powder concentrator in time according to the control strategy when the pressure difference changes, so that the intelligent control of the pressure difference is realized.
As a further optimized technical solution, a plurality of function blocks are edited and packaged in the MPC algorithm module, wherein SUBSCRIBE is used for receiving data of the OPC module, a _ M _ SWITCH is used for determining a manual/automatic state, MODE _ select is used for determining whether a current pressure difference of the hoist is in an out-of-limit state, an HRG _ MPC module is used for controlling a current of the main grinding machine and controlling a water spray flow, the left side of the function block is input, the right side of the function block is output, the upper side of the function block is event, the bottom of the function block is data, the function block firstly needs to trigger an initialization event, then after the parameters and the data are configured, the module can be used by triggering an MPC event, MVOut is a value of a control quantity calculated by the algorithm, MPC0 is a related event, the function block can modify related parameters online and take effect in real time, HRT _ EMERG is used for controlling the current of the hoist and the pressure difference, and OPC _ MV is used for issuing a calculation result of the current control module of the hoist, the method comprises the steps that HTTP _ EMERG _ DATA is used for sending DATA of a library to be stored, OPC _ MPC _ MV is used for sending a calculation result of a main engine current control module, and HTTP _ MPC _ DATA and HTTP _ MPC _ CONFIG are used for sending the DATA of the library to be stored.
The invention has the advantages that: the control quantity of the technical scheme of the invention is the mill main machine current, wherein the algorithm modeling only needs to do step response once, and the invention focuses on realizing the stable control of the mill main machine current through an MPC algorithm and various combined control methods, and the control quantity is as follows:
and (3) stabilizing the production state: because the roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation is real-time online control, system disturbance is effectively inhibited in time through multivariable coordination operation, important control parameters of the system can be well kept stable, the fluctuation of controlled variables is greatly reduced, and the key controlled variables approach a control target infinitely.
A combined control strategy is designed for ensuring stable and efficient operation of the raw mill, the combined control strategy comprising: the system comprises a main grinding machine current control strategy, a water spraying flow control strategy, a hoister current control strategy and a pressure difference control strategy.
The production line is kept stable as a whole, and the machine hour yield and the product quality are obviously improved through the standard operation behavior controlled by the computer; the unit energy consumption is obviously reduced.
The labor productivity is improved; the service life of the equipment is prolonged.
Drawings
FIG. 1 is a diagram of the major functional blocks of an advanced control method;
FIG. 2 is an editing interface of the OPC module;
FIG. 3 is an editing interface of the algorithm module;
FIG. 4 is a manual control interface for a milling operation;
FIG. 5 is an editing interface of the HTTP module;
FIG. 6 is a Web-side page of autonomic development;
FIG. 7 is a dynamic debug interface.
FIG. 8 is a flow chart of a model predictive control algorithm execution.
Fig. 9 is a graph of the effect of the control of the present invention compared to a purely manual control.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the rolling mill adaptive prediction Control system based on rolling time domain estimation of the present invention includes an OPC (OLE for Process Control for object linking and embedding for Process Control) SERVER module (OPC SERVER module), an OPC module, an MPC algorithm module, an HTTP module, a database module, and a WEB module.
As shown in fig. 2, the OPC module encapsulates an OPC CLIENT, a subscribe function block and a PUBLISH function block, wherein the OPC CLIENT is responsible for reading and writing OPC SERVER data, and reading DCS data on the OPC SERVER, and the DCS data include: the method comprises the steps of grinding host machine current feedback, grinding host machine current target value, lifting machine current feedback, lifting machine current threshold value, manual and automatic mark, feeding amount setting, automatic water spraying flow setting amount, water spraying control valve feedback, automatic water spraying valve setting, water spraying flow feedback, water-material ratio, lifting machine current dead zone, lifting machine current feedback value, raw material grinding feeding amount automatic setting value, lifting machine current upper limit, feeding amount feedback, current-located pressure difference interval, pressure difference dead zone, pressure difference different interval limit values, automatic powder concentrator rotating speed setting value, powder concentrator reference value and grinding machine pressure difference value, filtering the DCS data and outputting the DCS data to an MPC algorithm module.
Preferably, any number of read-write data can be configured in the OPC CLIENT, and the OPC CLIENT reads DCS data from the OPC SERVER at regular time and sends the DCS data to the algorithm module for processing. In the data configured by the OPC CLIENT, the read-write operation of the OPC CLIENT is triggered as long as the data changes, and the data configured by the OPC CLIENT can set parameters such as dead zone read frequency.
Meanwhile, after the OPC CLIENT in the OPC module reads data, the data can be sent to the next module for processing through the PUBLISH function block. The SUBSCRIB functional block also receives data output by the algorithm module and writes the data into the OPC SERVER module, so that the issuing operation of the control quantity is completed, and the data written into the OPC SERVER module comprises the following steps: heartbeat data and feeding quantity set values.
And the MPC algorithm module is responsible for processing and calculating data. As shown in FIG. 3, the MPC algorithm module acquires data from the OPC module, and performs filtering processing and analysis calculation. And simultaneously splicing and sending the database data. The specific implementation process of the MPC algorithm module is as follows:
after the MPC algorithm module acquires data, firstly, the MPC algorithm module judges which mode the MPC algorithm module is currently in, if the MPC algorithm module is in the manual mode, the MPC algorithm module does not analyze and process the data, because the whole system is completely controlled by a grinding operator at the moment, an interface shown in FIG. 4 is a manual control interface of the grinding operator, and the algorithm module splices and sends the data input by the grinding operator and key technical parameters of the roller type vertical mill, so that subsequent storage and WEB module display are facilitated; if the elevator is in the automatic mode, the MPC algorithm module firstly judges whether the current of the elevator is out of limit, if so, the MPC algorithm module enters the elevator current control algorithm for processing, and simultaneously judges whether the differential pressure is out of limit, and if so, the MPC algorithm module enters the differential pressure control algorithm for processing; and if the values of the current and the pressure difference of the hoister are normal, entering a mill host current control algorithm and a water spraying flow control algorithm.
The control algorithm of the MPC algorithm module adopts Adaptive predictive MPC (Adaptive MPC). In general, model predictive control uses a linear time-invariant dynamic model for prediction, however, in practice, due to the nonlinearity of an actuator, it is difficult to accurately predict the linear time-invariant model. In order to improve the robustness of the controller, the MPC algorithm module uses a method based on rolling horizon estimation (MHE) to realize self-adaptive prediction control. The rolling horizon estimation uses dynamic optimization and a backward time measurement horizon to optimize tuning parameters and states. The data may include noise (random fluctuations), drift (gradual deviations from true values), outliers (sudden and temporary deviations from true values), or other inaccuracies, and the dynamic optimization problem is numerically converged using a nonlinear programming solver. For example, in practical application, the current of the mill main machine is controlled by the feeding amount, a step test is firstly carried out, namely, the feeding amount is greatly changed, the change of the current of the mill main machine is observed, and initial values of model parameters K (model gain), TAU (time constant) and THETA (time delay) are calculated. And then after the automatic operation, fitting the optimal model parameters by combining the latest historical data in each control period through a rolling time domain estimation algorithm. By updating the model parameters K, TAU and THETA in real time, the control effect is more accurate. The model predictive control algorithm execution flow is shown in fig. 8 and includes the following steps:
step 1: starting;
step 2: waiting for a periodic event trigger;
and step 3: estimating and optimizing model parameters K, TAU and THETA in a rolling time domain;
and 4, step 4: updating MPC controller model parameters K, TAU and THETA;
and 5: MPC control solving an optimal solution;
step 6: the optimal solution is delivered to an actuator, whether the cycle times reach the preset times is judged, if not, the model collection step 2 is carried out, otherwise, the step 7 is carried out;
and 7: and (6) ending.
The model prediction algorithm objective function of the MPC algorithm module is L1 Norm (L1-Norm), the set value of the controlled quantity can specify an upper limit (SPHI) and a lower limit (SPLO), and the middle range of the upper limit and the lower limit is called Dead Band (Dead Band). The formula is as follows:
minυ＝whiehi+wloelo+wyy+wuu+wΔuΔu
in which upsilon is the objective function, whiTo penalize the weight for the upper bound, ehiIs an upper bound error value, wloTo penalize the weight lower, eloThe lower error value, wyIs the controlled quantity control weight, y is the predicted value of the controlled quantity model, wuIs a control weight of the control quantity, u is the control quantity, wΔuFor the control quantity change penalty weight, Δ u is the change in the control quantity.
Increase whiCan inhibitThe controlled quantity exceeds the upper limit of the target value, and w is increasedloThe controlled amount can be suppressed from falling below the target value lower limit. w is ayA positive can inhibit a change in the controlled amount and a negative can promote a change in the controlled amount. w is auThe change in the control amount may be suppressed to be positive, and the change in the control amount may be promoted to be negative. w is aΔuThe increase can suppress variation in the control amount within a single control cycle. This objective function enables individual control of the quantities in the model by adjusting the weights.
After all the algorithms are processed, the input quantity and the output quantity are spliced and sent, and subsequent storage and WEB display analysis are facilitated. The following describes several control strategies involved in a vertical roller mill:
the current control of the mill main machine is realized, and in order to keep the roller type vertical mill running with high efficiency and high stability, the stable running of the mill can be ensured by ensuring the stability of the current of the mill main machine through the on-site analysis and research findings of a process expert, a control expert and a mill operation expert. The feeding amount and the current of the mill main machine have obvious positive correlation, the controlled amount is the feeding amount, and the controlled amount is the current of the mill main machine. We do step response, look at the response curve of the data, and calculate the corresponding K, TAU, THETA. In the MPC algorithm module, model parameters are initialized as follows:
model.k ═ 1.2 (model gain)
Tau 160.0 (time constant)
model, theta is 50.0 (time delay)
model.pre _ ho ═ 60.0 (prediction step size)
Ctl _ ho ═ 60.0 (control step size)
Ctl _ intvl ═ 2.0 (control period)
model. y0 ═ 0.0 (initial value of controlled quantity)
model.u0 ═ 0.0 (initial value of controlled variable)
model u _ upper 100.0 (upper limit of control quantity)
model u _ lower is 0.0 (lower limit of control quantity)
U _ dmax ═ 20.0 (maximum value of control quantity change per cycle)
U _ dcost is 0.2 (control quantity change penalty)Weight WΔu)
U _ cost is 0.0 (control weight w of control amount)u)
model. y _ tau ═ 20.0 (controlled variable time constant)
model.y _ cost ═ 0.0 (controlled quantity control weight w)y)
model, sphi ═ 35.5 (upper limit of set value)
model, splo ═ 34.5 (lower limit of set point)
model, wsphi, 20.0 (upper bound penalty weight w)hi)
model.wsplo ═ 20.0 (lower penalty weight w)lo)
The model is characterized in that the model is a target value range of the mill main machine current, the target value range of the mill main machine current is assigned with an initial value of 34.5-35.5, and the initial value can be assigned when an algorithm module is initialized or the target value can be modified at any time in the operation process. In the production process of the roller type vertical mill, the current of the mill main machine can change along with the working condition, and the MPC algorithm module calculates the feeding amount in real time according to the feedback value of the current of the mill main machine and issues control, so that the stability of the current of the mill main machine is realized. In the control process of the algorithm module, the feeding amount is automatically set, the feeding amount is fed back, the current of the grinding host is fed back, and the current set value of the grinding host is spliced and sent to the HTTP module, so that the subsequent storage, display and debugging are facilitated.
The control of the water spray flow is found by the field analysis and research of a process expert, a control expert and a grinding operation expert, the fixed proportion of the water spray flow and the feeding amount exists, and the proportion can be manually set by the grinding operation personnel according to the condition of the stone on the field. The amount of water sprayed varies proportionally with the amount of feed. The water spraying flow is adjusted by controlling the opening degree of the water spraying valve. We do step response, look at the response curve of the data, and calculate the corresponding K, TAU, THETA. In the MPC algorithm module, model parameters are initialized as follows:
model.k ═ 1.2 (model gain)
Tau 160.0 (time constant)
model, theta is 50.0 (time delay)
model.pre _ ho ═ 60.0 (prediction step size)
Ctl _ ho ═ 60.0 (control step size)
Ctl _ intvl ═ 2.0 (control period)
model. y0 ═ 0.0 (initial value of controlled quantity)
model.u0 ═ 0.0 (initial value of controlled variable)
model u _ upper 100.0 (upper limit of control quantity)
model u _ lower is 0.0 (lower limit of control quantity)
U _ dmax ═ 20.0 (maximum value of control quantity change per cycle)
U _ dcost is 0.2 (control quantity change penalty weight w)Δu)
U _ cost is 0.0 (control weight w of control amount)u)
model. y _ tau ═ 20.0 (controlled variable time constant)
model.y _ cost ═ 0.0 (controlled quantity control weight W)y)
model, sphi ═ 35.5 (upper limit of set value)
model, splo ═ 34.5 (lower limit of set point)
model, wsphi, 20.0 (upper bound penalty weight w)hi)
model.wsplo ═ 20.0 (lower penalty weight w)lo)
Where model, spihi and model, splo are target ranges for water injection flow rate, which is obtained by multiplying the feed amount by a fixed ratio, and varies with the change in feed amount during operation. In the algorithm module control process, the automatic setting quantity of the water spraying flow, the feedback of the water spraying control valve, the automatic setting of the water spraying valve, the feedback of the water spraying flow and the splicing of the water-material ratio are realized and sent to the HTTP module, so that the subsequent storage, display and debugging are facilitated.
The current control of the elevator is discovered through field analysis and research of process experts, control experts and grinding operation experts, the current of the elevator is an important index of stable operation of the roller type vertical mill, and the stable operation of the roller type vertical mill is influenced by overhigh current of the elevator, so that the current of the elevator of the roller type vertical mill needs to be monitored in real time, and a control strategy is adjusted to perform automatic intervention when the current of the elevator is overhigh. When the current of the hoister exceeds a set threshold value, the current control module of the main grinder and the water spray flow control module are stopped, the feeding amount is directly reduced by 20t (the value can be dynamically modified by a grinder operator), then the numerical change of the current of the hoister is continuously monitored until the current value of the hoister is restored to a normal value, and then the current control module of the main grinder and the water spray flow control module are started. In the control process of the MPC algorithm module, the current dead zone of the hoister, the current feedback value of the hoister, the automatic set value of the feeding amount of the raw material mill, the current upper limit of the hoister and the feeding amount feedback are spliced and sent to the HTTP module, so that the subsequent warehousing, display and debugging are facilitated.
Pressure difference control, which is discovered through field analysis and research of process experts, control experts and mill operation experts, wherein mill pressure difference is an important index of stable operation of the roller-type vertical mill, and when the mill pressure difference is too high, mill blockage can be caused to influence the stable operation of the roller-type vertical mill, so that the mill pressure difference of the roller-type vertical mill is monitored in real time, and a control strategy is timely adjusted to automatically intervene to reduce the pressure difference. The pressure difference control strategy is to divide the pressure difference value into 4 intervals, directly quit the automatic control and switch to the manual mode when the pressure difference is in the zone1, and carry out the emergency treatment by the manual intervention of the milling operator, when the pressure difference is in the zone2, the rotating speed of the powder machine is reduced by 2 steps on the basis of the default rotating speed, when the pressure difference is in the zone3, the rotating speed of the powder machine is reduced by 1 step on the basis of the default rotating speed, and when the pressure difference is in the zone4, the rotating speed of the powder selecting machine is set as the default value. And the pressure difference control module monitors the pressure difference value in real time, and adjusts the rotating speed of the powder concentrator in time according to the control strategy when the pressure difference changes, so that the intelligent control of the pressure difference is realized. In the algorithm module control process, the current interval of the pressure difference, the pressure difference dead zone, the limit values of different intervals of the pressure difference, the automatic set value of the rotating speed of the powder concentrator, the reference value of the powder concentrator and the pressure difference value of the mill can be spliced and sent to the HTTP module, so that the subsequent storage, display and debugging are facilitated.
The HTTP module is responsible for warehousing the relevant data. As shown in fig. 5, after receiving the data of the MPC algorithm module, the HTTP module sends the data to the HTTP SERVER module, and the HTTP SERVER module calls the database interface to store the data in the time-series database. The database table is designed as follows:
and the WEB module is responsible for data display. As shown in fig. 6, the WEB module may obtain the value of the time sequence database in real time and display the value. The displayed contents comprise MPC algorithm model parameter values, real-time data curves, parameter statistical information and the like. The WEB module can self-define the displayed data, can display the curve of any time period, and can perform operations such as amplification, reduction, stretching and the like on the curve of any time period. The Y-axis range of the curve can be customized, and the contents such as curve refreshing frequency and the like can be set.
The invention designs a plurality of control strategies aiming at the intelligent control of the roller type vertical mill, and the control strategies can be edited through an Integrated Development Environment (IDE) to dynamically modify the control logic. Specific functions may be encapsulated and multiplexed. As shown in fig. 3, a plurality of function blocks are edited and encapsulated in the MPC algorithm module, where SUBSCRIBE is used to receive data of the OPC module, after receiving the OPC data, the a _ M _ SWITCH function block is used to determine the manual/automatic state, and then the determination result is output to the MODE _ SELECTOR module, the MODE _ SELECTOR module simultaneously determines whether the current and pressure difference of the elevator is in the out-of-limit state, and if not, the MODE _ SELECTOR module enters the HRG _ MPC module to perform the current control and the water spray flow control of the main unit. The left side of the function block of the HRG _ MPC module is input, and the right side is output. Events are at the top and data at the bottom. Using a function block requires first triggering an initialization (INIT) event, then configuring parameters and data, and then using the module by triggering an MPC event. MVOut is the value of the control quantity calculated by the algorithm, and MPC0 is the associated event. This function block can modify relevant parameters online and take effect in real time. If the MODE _ SELECTOR module judges that the current of the elevator is in an out-of-limit state, the current of the elevator enters an HRT _ EMERG function block to control the current and the differential pressure of the elevator, a calculation result of the elevator current control module is issued through an OPC _ EMERG _ MV function block after the primary control is finished, DATA of a library to be stored is sent through an HTTP _ EMERG _ DATA function block, a calculation result of the main motor current control module is issued through an OPC _ MPC _ MV function block, and the DATA of the library to be stored is sent through the HTTP _ MPC _ DATA function block and an HTTP _ MPC _ CONFIG function block. As shown in fig. 7, dynamic debugging can be performed in the field deployment process by using the IDE, visualization can be achieved for data flow in the debugging process, breakpoint debugging can be set, and local debugging and assignment can be performed. Fig. 9 is a graph showing the effect of the control according to the present invention compared with the effect of the purely manual control. It can be seen from fig. 9 that the control curve of the present invention is significantly smoother and less fluctuating than the data curve of the manual control.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present 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 solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. Roll-type vertical mill self-adaptation prediction control system based on rolling time domain estimation is characterized in that: the system comprises an OPC SERVER module, an OPC module, an MPC algorithm module, an HTTP module, a database module and a WEB module;
the OPC module is responsible for reading and writing OPC SERVER data, reading DCS data on the OPC SERVER, performing filtering processing on the DCS data and outputting the DCS data to the MPC algorithm module, and meanwhile, the OPC module also receives the data output by the algorithm module and writes the data into the OPC SERVER module to finish the issuing operation of the control quantity;
after the MPC algorithm module acquires data, firstly, judging which mode the MPC algorithm module is in at present, if the MPC algorithm module is in a manual mode, the MPC algorithm module does not analyze and process the data, and the algorithm module splices and sends the data input by a mill operator and key technical parameters of the roller type vertical mill; if the mode is the automatic mode, the MPC algorithm module firstly judges whether the current of the hoist is out of limit, if the current of the hoist is out of limit, the MPC algorithm module enters a current control algorithm of the hoist for processing, and simultaneously judges whether the pressure difference is out of limit, if the pressure difference is out of limit, the MPC algorithm module enters a pressure difference control algorithm for processing; if the values of the current and the pressure difference of the hoister are normal, entering a mill host current control algorithm and a water spray flow control algorithm, realizing self-adaptive prediction control by using an MPC algorithm module by using a rolling time domain estimation-based method, optimizing and adjusting parameters and states by using dynamic optimization and backward time measurement vision by using rolling time domain estimation, carrying out numerical convergence on the dynamic optimization problem by using a nonlinear programming solver, updating object model parameters by using a self-adaptive MPC controller in each control interval, keeping the model parameters unchanged in a prediction range once the model parameters are updated, and splicing and sending input quantity and output quantity after algorithm processing;
after receiving the data of the MPC algorithm module, the HTTP module sends the data to the HTTP SERVER module, and the HTTP SERVER module calls a database interface to store the data in a time sequence database;
and the WEB module can acquire the numerical value of the time sequence database and display the numerical value.
2. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the DCS data comprises: the method comprises the following steps of (1) grinding host current feedback, a grinding host current target value, lifting machine current feedback, a lifting machine current threshold value, a manual-automatic mark, feeding amount setting, automatic water spraying flow setting, water spraying control valve feedback, automatic water spraying valve setting, water spraying flow feedback, a water-material ratio, a lifting machine current dead zone, a lifting machine current feedback value, a raw material grinding feeding amount automatic setting value, a lifting machine current upper limit, feeding amount feedback, a current-located differential pressure interval, a differential pressure dead zone, differential pressure different interval limit values, a powder concentrator rotating speed automatic setting value, a powder concentrator reference value and a grinding machine pressure difference value;
the data written into the OPC SERVER module comprises: heartbeat data and feeding quantity set values.
3. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the OPC module encapsulates an OPC CLIENT, a SUBSCRIB function block and a PUBLISH function block, any number of read-write data can be configured in the OPC CLIENT, the OPC CLIENT reads DCS data from the OPC SERVER end at regular time and sends the DCS data to the algorithm module for processing, in the data configured by the OPC CLIENT, the read-write operation of the OPC module can be triggered as long as the data are changed, after the OPC CLIENT reads the data, the data are sent to the next module for processing through the PUBLISH function block, the SUBSCRIB function block receives the data output by the algorithm module and writes the data into the OPC SERVER module, and therefore the control quantity issuing operation is completed.
4. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the model predictive control algorithm includes the steps of:
step 1: starting;
step 2: waiting for a periodic event trigger;
and step 3: estimating and optimizing model parameters in a rolling time domain;
and 4, step 4: updating MPC controller model parameters;
and 5: MPC control solving an optimal solution;
step 6: the optimal solution is delivered to an actuator, whether the cycle times reach the preset times is judged, if not, the model collection step 2 is carried out, otherwise, the step 7 is carried out;
and 7: and (6) ending.
5. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the model prediction algorithm target function of the MPC algorithm module is L1 norm, the set value of the controlled variable designates an upper limit SPHI and a lower limit SPLO, the middle range of the upper limit and the lower limit is called a dead zone, and the formula is as follows:
minθ＝whiehi+wloelo+wyy+wuu+wΔuΔu
where θ is the objective function, whiTo penalize the weight for the upper bound, ehiIs an upper bound error value, wloTo penalize the weight lower, eloThe lower error value, wyIs the controlled quantity control weight, y is the predicted value of the controlled quantity model, wuIs a control weight of the control quantity, u is the control quantity, wΔuChanging penalty weight for the control quantity, and delta u is the change of the control quantity;
increase whiCan inhibit the controlled quantity from exceeding the target value upper limit and increase wloCan inhibit the controlled quantity from falling below the lower limit of the target value, wyIn order to be positive to suppress the change in the controlled quantity and negative to promote the change in the controlled quantity, wuIn order to positively suppress the change in the controlled variable and to negatively promote the change in the controlled variable, wΔuThe increase can suppress the variation of the control amount in a single control period, and the objective function can individually control each amount in the model by adjusting the weight.
6. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the grinding host current control algorithm strategy is as follows:
the feeding quantity and the current of the mill host have obvious positive correlation, the control quantity is the feeding quantity, the controlled quantity is the current of the mill host, step response is carried out, a response curve of data is checked, corresponding K, TAU and THETA are calculated, model parameters are assigned to initial values in an MPC algorithm module, the current of the mill host can be continuously changed along with working conditions in the production process of the roller type vertical mill, the MPC algorithm module calculates the feeding quantity in real time according to the feedback value of the current of the mill host and sends control to realize the stability of the current of the mill host, and in the control process of the MPC algorithm module, the automatic feeding quantity set value, feeding quantity feedback, mill host current feedback and mill host current set value splicing and sending to an HTTP module are carried out, so that subsequent warehousing, display and debugging are facilitated.
7. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the water spray flow control algorithm strategy is as follows:
the water spray volume can be changed proportionally along with the change of the feeding volume, the size of the water spray volume is adjusted by controlling the opening degree of a water spray valve, step response is carried out, the response curve of data is checked, corresponding K, TAU and THETA are calculated, initial values are given to model parameters in an MPC algorithm module, the automatic water spray volume setting quantity, the water spray control valve feedback, the water spray valve automatic setting, the water spray volume feedback and the water-material ratio value are spliced and sent to an HTTP module in the control process of the algorithm module, and subsequent storage and display debugging are facilitated.
8. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the elevator current control algorithm strategy is as follows:
when the current of the hoister exceeds a set threshold value, the current control module of the main grinder and the water spray flow control module are stopped, the feeding amount is directly reduced, then the numerical value change of the current of the hoister is continuously monitored until the current value of the hoister is restored to a normal value, and then the current control module of the main grinder and the water spray flow control module are started.
9. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the control strategy of the pressure difference control algorithm is as follows:
the pressure difference control strategy is to divide the pressure difference value into 4 intervals, directly quit the automatic control and switch to the manual mode when the pressure difference is in the zone1, and carry out the emergency treatment by the manual intervention of the milling operator, when the pressure difference is in the zone2, the rotating speed of the powder machine is reduced by 2 steps on the basis of the default rotating speed, when the pressure difference is in the zone3, the rotating speed of the powder machine is reduced by 1 step on the basis of the default rotating speed, and when the pressure difference is in the zone4, the rotating speed of the powder selecting machine is set as the default value. The pressure difference control module monitors the pressure difference value in real time, and adjusts the rotating speed of the powder concentrator in time according to the control strategy when the pressure difference changes, so that the intelligent control of the pressure difference is realized.
10. The rolling mill adaptive predictive control system based on rolling time domain estimation of claim 1, wherein: the MPC algorithm module is edited and packaged with a plurality of function blocks, wherein SUBSCRIBE is used for receiving DATA of an OPC module, A _ M _ SWITCH is used for judging a manual-automatic state, MODE _ SELECTOR is used for judging whether a current pressure difference and the like of a hoist are in an out-of-limit state, an HRG _ MPC module is used for grinding current control and water spray flow control of a host, the left side of the function block is input, the right side of the function block is output, the upper side of the function block is an event, the bottom of the function block is DATA, the function block firstly needs to trigger an initialization event and then is configured with parameters and DATA, the module can be used by triggering the MPC event, MVOut is a value of a control quantity calculated by the algorithm, MPC0 is a related event, the function block can modify related parameters on line and take effect in real time, HRT _ EMERG is used for promoting current control and pressure difference control, OPC _ EMERG _ MV is used for issuing a calculation result of the hoist current control module, HTTP _ EMERG DATA _ is used for sending DATA to be stored, and the OPC _ MPC _ MV is used for sending the calculation result of the current control module of the main grinding machine, and the HTTP _ MPC _ DATA and the HTTP _ MPC _ CONFIG are used for sending the DATA of the library to be stored.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110532920.8A CN113198591B (en) | 2021-05-17 | 2021-05-17 | Roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110532920.8A CN113198591B (en) | 2021-05-17 | 2021-05-17 | Roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113198591A true CN113198591A (en) | 2021-08-03 |
| CN113198591B CN113198591B (en) | 2022-06-07 |
Family
ID=77031506
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110532920.8A Active CN113198591B (en) | 2021-05-17 | 2021-05-17 | Roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113198591B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115007303A (en) * | 2022-06-21 | 2022-09-06 | 合肥水泥研究设计院有限公司 | Predictive control method and storage medium for pre-grinding system of roller press |
| CN116351546A (en) * | 2022-12-19 | 2023-06-30 | 深圳市润丰智慧科技有限公司 | Automatic control method of combined grinding system of roller press and ball mill |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU5944299A (en) * | 1999-02-25 | 2000-08-31 | Tata Consultancy Services Limited | Method and apparatus of manufacturing cement |
| US20110125293A1 (en) * | 2009-11-25 | 2011-05-26 | Honeywell International Inc. | Fast algorithm for model predictive control |
| CN102151605A (en) * | 2011-03-17 | 2011-08-17 | 浙江大学 | Advanced control method and system for vertical mill based on model identification and predictive control |
| CN202097023U (en) * | 2011-03-17 | 2012-01-04 | 浙江大学 | Advanced vertical mill control system based on model identification and predictive control |
| CN102319612A (en) * | 2011-07-05 | 2012-01-18 | 中国科学院沈阳自动化研究所 | Method for intelligently controlling pressure difference of cement raw meal vertical mill |
| US20130030573A1 (en) * | 2010-03-24 | 2013-01-31 | Abb Research Ltd | Computer-based method and device for automatically providing control parameters for a plurality of coal mills supplying coal powder to a plant |
| WO2013087973A1 (en) * | 2011-12-16 | 2013-06-20 | Metso Automation Oy | Method of tuning a process controller |
| CN111701698A (en) * | 2020-05-26 | 2020-09-25 | 上海万澄环保科技有限公司 | Cement mill system and automatic optimization control system and method thereof |
| CN112517220A (en) * | 2020-11-19 | 2021-03-19 | 中材邦业(杭州)智能技术有限公司 | Optimized control system and method based on slag grinding system |
-
2021
- 2021-05-17 CN CN202110532920.8A patent/CN113198591B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU5944299A (en) * | 1999-02-25 | 2000-08-31 | Tata Consultancy Services Limited | Method and apparatus of manufacturing cement |
| US20110125293A1 (en) * | 2009-11-25 | 2011-05-26 | Honeywell International Inc. | Fast algorithm for model predictive control |
| US20130030573A1 (en) * | 2010-03-24 | 2013-01-31 | Abb Research Ltd | Computer-based method and device for automatically providing control parameters for a plurality of coal mills supplying coal powder to a plant |
| CN103097969A (en) * | 2010-03-24 | 2013-05-08 | Abb研究有限公司 | Computer-based method and device for automatically providing control parameters for a plurality of coal mills supplying coal powder to a plant |
| CN102151605A (en) * | 2011-03-17 | 2011-08-17 | 浙江大学 | Advanced control method and system for vertical mill based on model identification and predictive control |
| CN202097023U (en) * | 2011-03-17 | 2012-01-04 | 浙江大学 | Advanced vertical mill control system based on model identification and predictive control |
| CN102319612A (en) * | 2011-07-05 | 2012-01-18 | 中国科学院沈阳自动化研究所 | Method for intelligently controlling pressure difference of cement raw meal vertical mill |
| WO2013087973A1 (en) * | 2011-12-16 | 2013-06-20 | Metso Automation Oy | Method of tuning a process controller |
| CN111701698A (en) * | 2020-05-26 | 2020-09-25 | 上海万澄环保科技有限公司 | Cement mill system and automatic optimization control system and method thereof |
| CN112517220A (en) * | 2020-11-19 | 2021-03-19 | 中材邦业(杭州)智能技术有限公司 | Optimized control system and method based on slag grinding system |
Non-Patent Citations (3)
| Title |
|---|
| 宁艳艳等: "水泥生料立磨压差的广义预测PID控制", 《信息与控制》, no. 03, 15 June 2012 (2012-06-15) * |
| 陶伟光等: "基于模型预测控制的磨机优化***", 《有色金属(选矿部分)》, 15 September 2011 (2011-09-15) * |
| 颜文俊等: "水泥立磨流程的建模和控制优化", 《控制工程》, no. 06, 20 November 2012 (2012-11-20) * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115007303A (en) * | 2022-06-21 | 2022-09-06 | 合肥水泥研究设计院有限公司 | Predictive control method and storage medium for pre-grinding system of roller press |
| CN115007303B (en) * | 2022-06-21 | 2023-10-20 | 合肥水泥研究设计院有限公司 | Predictive control method and storage medium for roller press pre-grinding system |
| CN116351546A (en) * | 2022-12-19 | 2023-06-30 | 深圳市润丰智慧科技有限公司 | Automatic control method of combined grinding system of roller press and ball mill |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113198591B (en) | 2022-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113198591B (en) | Roller type vertical mill self-adaptive prediction control system based on rolling time domain estimation | |
| CN106990763B (en) | A kind of Vertical Mill operation regulator control system and method based on data mining | |
| CN100465825C (en) | Variable structural nonlinear model predictor controller | |
| CN1244032C (en) | Control system for E. G. primary-industry or manufacturing-industry facilities | |
| CN102151605A (en) | Advanced control method and system for vertical mill based on model identification and predictive control | |
| CN202097023U (en) | Advanced vertical mill control system based on model identification and predictive control | |
| CN102147613B (en) | Model predictive control method and system for moisture and temperature of tobacco redryer | |
| Haber et al. | Intelligent process supervision for predicting tool wear in machining processes | |
| CN112517220B (en) | Optimized control system and method based on slag grinding system | |
| CN107239066B (en) | A kind of Vertical Mill operation closed-loop control device and method based on data mining | |
| Chen et al. | Expert system based adaptive dynamic matrix control for ball mill grinding circuit | |
| EP3104240A1 (en) | Device and method for optimizing a working point for the operation of an installation | |
| CN111443597A (en) | Device and method for controlling granularity of vertical mill mineral powder | |
| CN101329575B (en) | Computer intelligent control method of sheet bar continuous casting and rolling | |
| Chai | Optimal operational control for complex industrial processes | |
| CN102284505A (en) | System for controlling thickness of fuzzy PI (Proportional Integral) based on ARM (Advanced RISC machines) | |
| EP1769292B1 (en) | Process-related systems and methods | |
| CN113769880B (en) | Industrial big data-based optimization method for control index of cement production raw material grinding system | |
| Saif | A frame work for the integration of statistical process control and engineering process control | |
| Hoang et al. | Self-adjusting on-line cutting condition for high-speed milling process | |
| CN117519042B (en) | Intelligent control method, system and storage medium for cement production based on AI technology | |
| JP2019184575A (en) | Measurement operation parameter adjustment device, machine learning device, and system | |
| CN106527149A (en) | Reconstruction point decision-making method for reconfigurable manufacture unit | |
| CN118011780B (en) | Control method and system of lithium battery roll slitting machine based on PID | |
| Jovan et al. | Utilization of key performance indicators in production 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 |