CN113759708A - System optimization control method and device and electronic equipment - Google Patents

System optimization control method and device and electronic equipment Download PDF

Info

Publication number
CN113759708A
CN113759708A CN202110181933.5A CN202110181933A CN113759708A CN 113759708 A CN113759708 A CN 113759708A CN 202110181933 A CN202110181933 A CN 202110181933A CN 113759708 A CN113759708 A CN 113759708A
Authority
CN
China
Prior art keywords
historical
model
controlled
control
data
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
Application number
CN202110181933.5A
Other languages
Chinese (zh)
Other versions
CN113759708B (en
Inventor
朱翔宇
詹仙园
霍雨森
张玥
殷宏磊
郑宇�
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jingdong City Beijing Digital Technology Co Ltd
Original Assignee
Jingdong City Beijing Digital Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jingdong City Beijing Digital Technology Co Ltd filed Critical Jingdong City Beijing Digital Technology Co Ltd
Priority to CN202110181933.5A priority Critical patent/CN113759708B/en
Publication of CN113759708A publication Critical patent/CN113759708A/en
Application granted granted Critical
Publication of CN113759708B publication Critical patent/CN113759708B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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

Landscapes

  • Engineering & Computer Science (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 application provides an optimization control method and device of a system and electronic equipment, wherein the optimization control method comprises the following steps: acquiring real-time operation data of a system to be controlled; obtaining a pre-constructed operation optimization model of a system to be controlled, wherein the operation optimization model is constructed according to historical operation data of the system to be controlled and a model prediction control algorithm; and generating a recommended value of the control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm. Therefore, the operation optimization model can be constructed by utilizing the historical operation data of the system to be controlled in an off-line mode, time and labor cost are saved, the influence of the real-time operation data of the system to be controlled on the recommended values of the control parameters can be fully considered, the model prediction control on-line solving algorithm can be utilized for carrying out on-line solving, the solving real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.

Description

System optimization control method and device and electronic equipment
Technical Field
The present application relates to the field of system control technologies, and in particular, to a method and an apparatus for optimizing and controlling a system, an electronic device, and a storage medium.
Background
At present, with the continuous development of technologies such as industry, energy, intelligent manufacturing and the like, the requirement on the optimization control of the system is higher and higher. However, most of the system optimization control methods in the related art adopt a traditional PID control method, mainly perform single-step control adjustment, and have a certain hysteresis in operation, or adopt a control method based on a test, which needs to acquire a large amount of test data, consumes a large amount of manpower and material resources, and has a high cost, or adopt a control method based on a model, which has a high difficulty in modeling a complex system and a low accuracy.
Disclosure of Invention
The method and the device aim at solving one of the technical problems that the optimization control of the system in the related technology has hysteresis, consumes larger manpower and material resources and has lower accuracy to at least a certain extent.
Therefore, the embodiment of the first aspect of the present application provides an optimization control method for a system, which can utilize historical operating data of a system to be controlled to construct an operation optimization model offline, and is helpful for saving time and labor cost, and can utilize a model predictive control algorithm to construct the operation optimization model offline, so that the problem of operation hysteresis of a control method for the system in the related art can be solved, and the stable operation of the system to be controlled can be ensured.
The embodiment of the second aspect of the present application provides an optimization control device of a system.
The embodiment of the third aspect of the application provides an electronic device.
An embodiment of a fourth aspect of the present application provides a computer-readable storage medium.
An embodiment of a first aspect of the present application provides an optimization control method for a system, including: acquiring real-time operation data of the system to be controlled; obtaining a pre-constructed operation optimization model of the system to be controlled, wherein the operation optimization model is constructed according to historical operation data of the system to be controlled and a model prediction control algorithm; and generating a recommended value of the control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and a model predictive control on-line solving algorithm.
According to the optimization control method of the system, the real-time operation data of the system to be controlled are obtained, the operation optimization model of the system to be controlled which is constructed in advance is obtained, the operation optimization model is constructed according to the historical operation data of the system to be controlled and a model prediction control algorithm, and the recommended value of the control parameter of the system to be controlled is generated according to the real-time operation data, the operation optimization model and the model prediction control online solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
In addition, the optimization control method of the system according to the above embodiment of the present application may further have the following additional technical features:
in one embodiment of the present application, further comprising: acquiring historical operating data of the system to be controlled; training according to the historical operating data to obtain a plurality of characteristic submodels of the system to be controlled; and constructing the operation optimization model according to the characteristic submodels and the model predictive control algorithm.
In an embodiment of the application, the system to be controlled is an air cooling system of a thermal power generating unit, and the control parameter is an average rotating speed of each row of fans.
In an embodiment of the present application, the training to obtain a plurality of characteristic submodels of the system to be controlled according to the historical operating data includes: training to obtain a power consumption characteristic sub-model of each fan in the system to be controlled according to the historical fan rotating speed and the historical fan current in the historical operating data; training to obtain a turbine micro-power-increasing characteristic model of the system to be controlled according to historical boiler load, historical generator power, historical steam exhaust pressure and preset first historical state characteristic data related to historical generator power in the historical operation data; and training to obtain an exhaust pressure change characteristic model of the system to be controlled according to second historical state characteristic data and historical fan rotating speeds in the historical operating data.
In an embodiment of the present application, the power consumption characteristic submodel is:
I＝a1*x3+a2*x2+a3*x+b；
Figure BDA0002941691150000021
wherein I is the historical fan current;
the x is the historical fan rotating speed;
a is a1，a2，a3And b is a parameter to be trained;
the P isfThe power consumption of the fan is;
the above-mentioned
Figure BDA0002941691150000022
Setting a preset power factor value;
and U is the rated voltage of the fan.
In one embodiment of the present application, the turbine micro power augmentation characteristic model is:
Figure BDA0002941691150000034
wherein, the PdiffIs the difference between the historical boiler load and the historical generator power;
f ispIs a 2 nd order polynomial regression function to be trained;
said p isbThe historical exhaust steam pressure is obtained;
the above-mentioned
Figure BDA0002941691150000035
The first historical state characteristic data.
In one embodiment of the present application, the exhaust pressure variation characteristic model is:
st＝fbp((s，a)0～t-1)；
wherein, said stThe second historical state characteristic data at the time t;
f isbpPredicting a function for a time series to be trained;
the (s, a)0～t-1The second historical state characteristic data and the historical average rotating speed of each row of fans corresponding to t time steps before t time are obtained.
In one embodiment of the present application, the operation optimization model is:
Figure BDA0002941691150000031
Ct＝Pf；t+Pdiff；t
st+1＝f(st，at)；
s1＝sinit；
wherein the argmin is a minimization function;
s istThe second historical state characteristic data at the time t;
a is atThe historical average rotating speed of each row of fans at the time t is obtained;
the above-mentioned
Figure BDA0002941691150000032
A constraint condition for the second historical state feature data;
the above-mentioned
Figure BDA0002941691150000033
The constraint condition is the historical average rotating speed of each row of fans;
the T is a preset future time step number;
the P isf；tObtaining the historical power consumption of the fan at the time t according to the power consumption characteristic submodel;
the P isdiff；tObtaining a difference value between the historical boiler load and the historical generator power at the time t according to the steam turbine micro-power-increasing characteristic submodel;
f is a dynamic model obtained according to the change of the exhaust steam pressure change characteristic model;
s isinitThe second historical state characteristic data at the initial moment.
In an embodiment of the present application, the generating a recommended value of a control parameter of the system to be controlled according to the real-time operation data, the operation optimization model, and a model predictive control online solving algorithm includes:
generating a recommended control sequence of the average rotating speed of each row of fans by adopting the following preset formula according to the real-time operation data and the operation optimization model;
determining a first value in the recommended control sequence as a recommended value of the average rotating speed of each row of fans at the current moment;
wherein the preset formula is as follows:
Figure BDA0002941691150000041
μt＝η*avae+(1-η)*μ′t+1；
μ′T+1＝μ′T；
Figure BDA0002941691150000042
Figure BDA0002941691150000043
Figure BDA0002941691150000044
wherein, the mu'tIs the t value in the recommended control sequence;
the mutIs the t value in the initial action sequence;
n is the number of action sequences generated in each solving process;
the i is the serial number of the generated action sequence;
the gamma is a preset rewarding weighting factor;
the R isiThe accumulated reward value corresponding to the ith action sequence;
the above-mentioned
Figure BDA0002941691150000045
The average rotating speed of each row of fans at the t moment corresponding to the ith action sequence is obtained;
the eta is a preset weighting coefficient;
the T is a preset future time step number;
a is avaeThe control action is calculated according to a historical control strategy model obtained by pre-training;
the above-mentioned
Figure BDA0002941691150000046
A noise sample at the t moment corresponding to the ith action sequence;
the beta is a filter coefficient.
In one embodiment of the present application, further comprising: and training to obtain the historical control strategy model according to the second historical state characteristic data and the historical fan rotating speed in the historical operating data.
The embodiment of the second aspect of the present application provides an optimization control device for a system, including: the first acquisition module is used for acquiring real-time operation data of the system to be controlled; the second acquisition module is used for acquiring a pre-constructed operation optimization model of the system to be controlled, and the operation optimization model is constructed according to historical operation data of the system to be controlled and a model prediction control algorithm; and the generation module is used for generating a recommended value of the control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm.
The optimization control device of the system of the embodiment of the application obtains real-time operation data of the system to be controlled, obtains a pre-constructed operation optimization model of the system to be controlled, constructs the operation optimization model according to historical operation data of the system to be controlled and a model predictive control algorithm, and generates a recommended value of a control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
In addition, the optimization control device of the system according to the above embodiment of the present application may further have the following additional technical features:
in one embodiment of the present application, the optimization control device of the system further includes: a model building module to: acquiring historical operating data of the system to be controlled; training according to the historical operating data to obtain a plurality of characteristic submodels of the system to be controlled; and constructing the operation optimization model according to the characteristic submodels and the model predictive control algorithm.
In an embodiment of the application, the system to be controlled is an air cooling system of a thermal power generating unit, and the control parameter is an average rotating speed of each row of fans.
In an embodiment of the application, the model building module is specifically configured to: training to obtain a power consumption characteristic sub-model of each fan in the system to be controlled according to the historical fan rotating speed and the historical fan current in the historical operating data; training to obtain a turbine micro-power-increasing characteristic model of the system to be controlled according to historical boiler load, historical generator power, historical steam exhaust pressure and preset first historical state characteristic data related to historical generator power in the historical operation data; and training to obtain an exhaust pressure change characteristic model of the system to be controlled according to second historical state characteristic data and historical fan rotating speeds in the historical operating data.
In an embodiment of the present application, the power consumption characteristic submodel is:
I＝a1*x3+a2*x2+a3*x+b；
Figure BDA0002941691150000051
wherein I is the historical fan current;
the x is the historical fan rotating speed;
a is a1，a2，a3And b is a parameter to be trained;
the P isfThe power consumption of the fan is;
the above-mentioned
Figure BDA0002941691150000052
Setting a preset power factor value;
and U is the rated voltage of the fan.
In one embodiment of the present application, the turbine micro power augmentation characteristic model is:
Figure BDA0002941691150000053
wherein, the PdiffIs the difference between the historical boiler load and the historical generator power;
f ispIs a 2 nd order polynomial regression function to be trained;
said p isbThe historical exhaust steam pressure is obtained;
the above-mentioned
Figure BDA0002941691150000054
The first historical state characteristic data.
In one embodiment of the present application, the exhaust pressure variation characteristic model is:
st＝fbp((s，a)0～t-1)；
wherein, said stThe second historical state characteristic data at the time t;
f isbpPredicting a function for a time series to be trained;
the (s, a)0～t-1The second historical state characteristic data and the historical average rotating speed of each row of fans corresponding to t time steps before t time are obtained.
In one embodiment of the present application, the operation optimization model is:
Figure BDA0002941691150000061
Ct＝Pf；t+Pdiff；t
st+1＝f(st，at)；
s1＝Sinit；
wherein the argmin is a minimization function;
s istThe second historical state characteristic data at the time t;
a is atThe historical average rotating speed of each row of fans at the time t is obtained;
the above-mentioned
Figure BDA0002941691150000062
A constraint condition for the second historical state feature data;
the above-mentioned
Figure BDA0002941691150000063
The constraint condition is the historical average rotating speed of each row of fans;
the T is a preset future time step number;
the P isf；tObtaining the historical power consumption of the fan at the time t according to the power consumption characteristic submodel;
the P isdiff；tObtaining a difference value between the historical boiler load and the historical generator power at the time t according to the steam turbine micro-power-increasing characteristic submodel;
f is a dynamic model obtained according to the change of the exhaust steam pressure change characteristic model;
s isinitThe second historical state characteristic data at the initial moment.
In an embodiment of the application, the generating module is specifically configured to:
generating a recommended control sequence of the average rotating speed of each row of fans by adopting the following preset formula according to the real-time operation data and the operation optimization model;
determining a first value in the recommended control sequence as a recommended value of the average rotating speed of each row of fans at the current moment;
wherein the preset formula is as follows:
Figure BDA0002941691150000064
μt＝η*avae+(1-η)*μ′t+1；
μ′T+1＝μ′T；
Figure BDA0002941691150000065
Figure BDA0002941691150000071
Figure BDA0002941691150000072
wherein, the mu'tIs the t value in the recommended control sequence;
the mutIs the t value in the initial action sequence;
n is the number of action sequences generated in each solving process;
the i is the serial number of the generated action sequence; the gamma is a preset rewarding weighting factor;
the R isiThe accumulated reward value corresponding to the ith action sequence;
the above-mentioned
Figure BDA0002941691150000073
The average rotating speed of each row of fans at the t moment corresponding to the ith action sequence is obtained;
the eta is a preset weighting coefficient;
the T is a preset future time step number;
a is avaeThe control action is calculated according to a historical control strategy model obtained by pre-training;
the above-mentioned
Figure BDA0002941691150000074
A noise sample at the t moment corresponding to the ith action sequence;
the beta is a filter coefficient.
In one embodiment of the present application, the optimization control device of the system further includes: a model training module to: and training to obtain the historical control strategy model according to the second historical state characteristic data and the historical fan rotating speed in the historical operating data.
An embodiment of a third aspect of the present application provides an electronic device, including: a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method for optimizing control of a system as described in the foregoing embodiments of the first aspect when executing the program.
The electronic equipment of the embodiment of the application executes a computer program stored on a memory through a processor to obtain real-time operation data of a system to be controlled and obtain a pre-constructed operation optimization model of the system to be controlled, the operation optimization model is constructed according to historical operation data of the system to be controlled and a model predictive control algorithm, and a recommended value of a control parameter of the system to be controlled is generated according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
An embodiment of a fourth aspect of the present application provides a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the optimization control method of the system according to the embodiment of the first aspect.
The computer-readable storage medium of the embodiment of the application, by storing a computer program and executing the computer program by a processor, acquires real-time operation data of a system to be controlled, acquires a pre-constructed operation optimization model of the system to be controlled, the operation optimization model is constructed according to historical operation data of the system to be controlled and a model predictive control algorithm, and generates a recommended value of a control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control online solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a schematic flow chart of a method for optimizing control of a system according to one embodiment of the present application;
FIG. 2 is a schematic flow chart illustrating a process of constructing an operation optimization model of a system to be controlled in an optimization control method of the system according to an embodiment of the present application;
FIG. 3 is a schematic flow chart illustrating a process of generating recommended values of control parameters of a system to be controlled in an optimization control method of the system according to an embodiment of the present application;
FIG. 4 is a schematic diagram of an optimization control method of a system according to one embodiment of the present application;
FIG. 5 is a schematic block diagram of an optimization control apparatus of a system according to an embodiment of the present application;
FIG. 6 is a schematic block diagram of an optimization control device of a system according to another embodiment of the present application; and
fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining the present application and should not be construed as limiting the present application.
An optimization control method, an apparatus, an electronic device, and a storage medium of a system according to an embodiment of the present application are described below with reference to the drawings.
Fig. 1 is a flow chart illustrating an optimization control method of a system according to an embodiment of the present application.
As shown in fig. 1, the optimization control method of the system according to the embodiment of the present application includes:
and S101, acquiring real-time operation data of a system to be controlled.
It should be noted that the main body of the system optimization control method according to the embodiment of the present application may be an optimization control device of the system, and the optimization control device of the system according to the embodiment of the present application may be configured in any electronic device, so that the electronic device may execute the system optimization control method according to the embodiment of the present application. The electronic device may be a Personal Computer (PC), a cloud device, a mobile device, and the like, and the mobile device may be a hardware device having various operating systems, touch screens, and/or display screens, such as a mobile phone, a tablet Computer, a Personal digital assistant, a wearable device, and an in-vehicle device.
In the embodiment of the application, the real-time operation data of the system to be controlled can be acquired. It can be understood that the system to be controlled can correspond to various real-time operation data, and different systems to be controlled can correspond to different real-time operation data.
It should be noted that, in the embodiment of the present application, the type of the system to be controlled is not limited. For example, the system to be controlled includes, but is not limited to, systems to be controlled in the fields of industry, energy, smart manufacturing, and the like.
Optionally, the system to be controlled may be an air cooling system of the thermal power generating unit. It should be noted that the air cooling system of the thermal power generating unit is a steam cooling system of the thermal power generating unit (a cold end system of the thermal power generating unit), and includes an air cooling fan unit array, where the air cooling fan unit array is composed of a large number of air cooling fans.
When the system to be controlled is an air cooling system of the thermal power generating unit, the corresponding real-time operation data includes, but is not limited to, boiler load, main steam temperature, main steam pressure, reheat steam temperature, reheat steam pressure, ambient temperature, exhaust steam pressure, exhaust steam temperature, generator power, fan current, and the like.
S102, obtaining a pre-constructed operation optimization model of the system to be controlled, and constructing the operation optimization model according to historical operation data of the system to be controlled and a model prediction control algorithm.
In the embodiment of the application, a pre-constructed operation optimization Model of the system to be controlled can be obtained, and the operation optimization Model is constructed according to historical operation data of the system to be controlled and a Model Predictive Control (MPC) algorithm.
In the embodiment of the application, the type of the running optimization model is not limited too much. Alternatively, the running optimization model may be a Machine Learning (ML) model, for example, a Deep Neural Networks (DNN) model.
In the embodiment of the application, historical operating data of a system to be controlled can be acquired. It can be understood that the system to be controlled can correspond to various historical operating data, and different systems to be controlled can correspond to different historical operating data. For example, when the system to be controlled is an air cooling system of a thermal power generating unit, the corresponding historical operating data includes, but is not limited to, boiler load, main steam temperature, main steam pressure, reheat steam temperature, reheat steam pressure, ambient temperature, exhaust steam pressure, exhaust steam temperature, generator power, fan current, and the like.
Optionally, the obtaining of the historical operating data of the system to be controlled may include performing data processing on the collected historical operating data, so that the obtained historical operating data is more accurate. For example, abnormal data and/or sparse data in the collected historical operating data may be deleted, missing value padding may be performed on the collected historical operating data, and the like.
It can be understood that, in the embodiment of the present application, the operation optimization model of the system to be controlled may be pre-constructed according to the historical operation data of the system to be controlled and the model predictive control algorithm. Therefore, the operation optimization model can be constructed off line by utilizing the historical operation data of the system to be controlled, and time and labor cost are saved. The model predictive control algorithm can be used for constructing the operation optimization model in an off-line mode, the model predictive control algorithm is suitable for the time sequence control system, the operation optimization model can be suitable for multi-time-step control optimization in the time sequence control system, the problem of operation hysteresis of a control method of the system in the related technology can be solved, and stable operation of the system to be controlled can be guaranteed.
And S103, generating a recommended value of the control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm.
In the embodiment of the application, the recommended value of the control parameter of the system to be controlled can be generated according to the real-time operation data, the operation optimization model and the model predictive control online solving algorithm, so that the control parameter of the system to be controlled can be optimally controlled according to the recommended value of the control parameter of the system to be controlled.
For example, the real-time operation data can be input into the operation optimization model, and the real-time operation data is calculated by the operation optimization model through a model predictive control online solving algorithm to obtain a recommended value of the control parameter of the system to be controlled.
Optionally, the model predictive control online solving algorithm may be set according to actual conditions. For example, the Model Predictive control online solution algorithm may be a Model Predictive control solution-variant Auto-Encoders (MPPI-VAE) algorithm. Compared with a common optimization algorithm, the model predictive control solving-variational self-encoder algorithm can greatly improve the optimization solving efficiency, has better solving real-time performance and improves the optimization control real-time performance of a system to be controlled, and can more fully utilize historical operating data and has less dependence on a simulation environment, so that the recommended value of the control parameter obtained by solving better conforms to the real operating condition of the system to be controlled and can better meet the stable operating requirement of the system to be controlled.
In the embodiment of the application, the system to be controlled can correspond to one or more control parameters, and different systems to be controlled can correspond to different control parameters. For example, when the system to be controlled is an air cooling system of a thermal power generating unit, the corresponding control parameter is the average rotating speed of each row of fans.
It can be understood that, in the embodiment of the application, the recommended value of the control parameter of the system to be controlled can be generated in real time by using the real-time operation data, the operation optimization model and the model predictive control online solving algorithm, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, the model predictive control online solving algorithm can be used for performing online solving, the solving real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
In summary, according to the system optimization control method of the embodiment of the application, the real-time operation data of the system to be controlled is obtained, the operation optimization model of the system to be controlled which is constructed in advance is obtained, the operation optimization model is constructed according to the historical operation data of the system to be controlled and the model predictive control algorithm, and the recommended value of the control parameter of the system to be controlled is generated according to the real-time operation data, the operation optimization model and the model predictive control online solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
On the basis of any of the above embodiments, as shown in fig. 2, constructing an operation optimization model of the system to be controlled may include:
s201, obtaining historical operation data of a system to be controlled.
The relevant content of step S201 can be referred to the above embodiments, and is not described herein again.
And S202, training according to historical operating data to obtain a plurality of characteristic submodels of the system to be controlled.
In the embodiment of the application, a plurality of characteristic submodels of the system to be controlled can be obtained according to historical operating data training. Therefore, the plurality of characteristic submodels of the system to be controlled can be obtained by utilizing the historical operating data for off-line training, and the influence of the historical operating data on the plurality of characteristic submodels can be fully considered.
It is understood that, in the embodiments of the present application, the system to be controlled may correspond to a plurality of characteristic submodels, and different systems to be controlled may correspond to different characteristic submodels. For example, when the system to be controlled is an air cooling system of a thermal power generating unit, the corresponding multiple characteristic submodels include, but are not limited to, a power consumption characteristic submodel of each fan, a turbine micro-power-increasing characteristic model of the system to be controlled, and an exhaust pressure variation characteristic model of the system to be controlled.
In the embodiment of the present application, the types of the plurality of characteristic submodels are not limited too much. Alternatively, the running optimization model may be a Machine Learning (ML) model, for example, a Deep Neural Networks (DNN) model.
S203, constructing an operation optimization model according to the characteristic submodels and the model predictive control algorithm.
It is understood that in the embodiment of the present application, the operation optimization model may be constructed according to a plurality of characteristic submodels and a model predictive control algorithm. Therefore, the influence of the characteristic submodels on the operation optimization model can be comprehensively considered, so that the operation optimization model is more accurate.
Therefore, the method obtains the historical operation data of the system to be controlled, obtains a plurality of characteristic submodels of the system to be controlled according to the historical operation data, constructs the operation optimization model according to the plurality of characteristic submodels and the model predictive control algorithm, and can comprehensively consider the influence of the plurality of characteristic submodels on the operation optimization model, so that the operation optimization model is more accurate.
On the basis of any of the above embodiments, when the system to be controlled is an air cooling system of a thermal power generating unit, the corresponding multiple characteristic submodels include, but are not limited to, a power consumption characteristic submodel for each fan, a turbine micro-power-increasing characteristic model for the system to be controlled, and an exhaust pressure variation characteristic model for the system to be controlled.
Optionally, in step S202, a plurality of characteristic submodels of the system to be controlled are obtained through training according to the historical operating data, which may include obtaining a power consumption characteristic submodel of each fan in the system to be controlled through training according to the historical fan rotation speed and the historical fan current in the historical operating data. That is to say, the power consumption characteristic submodel of each fan is obtained by training according to the historical fan rotating speed and the historical fan current of each fan, and different fans can correspond to different power consumption characteristic submodels.
It is understood that the power consumption characteristic submodel may be set according to actual situations.
Optionally, the power consumption characteristic submodel is:
I＝a1*x3+a2*x2+a3*x+b；
Figure BDA0002941691150000121
wherein I is historical fan current, x is historical fan rotating speed, and a1，a2，a3B is the parameter to be trained, PfIs the power consumption of the fan,
Figure BDA0002941691150000122
and U is the rated voltage of the fan for a preset power factor fixed value.
For example,
Figure BDA0002941691150000123
may be set to 0.88 and U may be set to 380V.
Therefore, according to the method, the power consumption power characteristic submodel of each fan in the system to be controlled can be obtained through training according to the historical fan rotating speed and the historical fan current in the historical operating data, the influence of the historical fan rotating speed and the historical fan current on the power consumption power characteristic submodel of each fan can be comprehensively considered, different fans can correspond to different power consumption power characteristic submodels, and the flexibility is high.
Optionally, in step S202, a plurality of characteristic submodels of the system to be controlled are obtained through training according to the historical operating data, and a turbine micro-augmentation power characteristic model of the system to be controlled is obtained through training according to the historical boiler load, the historical generator power, the historical exhaust steam pressure, and the preset first historical state characteristic data related to the historical generator power in the historical operating data.
It will be appreciated that the first historical state characteristic data relating to historical generator power may be set as a function of the actual situation. Optionally, the first historical state feature data related to the historical generator power may be determined by a correlation analysis algorithm or a multi-round recursive feature elimination algorithm in feature engineering. It should be noted that the first historical state characteristic data does not include historical exhaust pressure. For example, the first historical state characteristic data includes, but is not limited to, historical boiler load, historical main steam pressure, historical main steam temperature, historical stage extraction steam temperature, and the like.
It can be understood that the turbine micro power characteristic model can be set according to actual conditions.
Optionally, the turbine micro-power-increasing characteristic model is as follows:
Figure BDA0002941691150000124
wherein, PdiffAs a difference between the historical boiler load and the historical generator power, fpFor a 2 nd order polynomial regression function to be trained, pbIn order to obtain the historical exhaust steam pressure,
Figure BDA0002941691150000125
is the first historical state characteristic data.
In this embodiment, the difference between the historical boiler load and the historical generator power may be used to represent the loss of steam during the process of performing work in the steam turbine, and when the historical boiler load is fixed, the smaller the difference between the historical boiler load and the historical generator power is, the larger the historical generator power is, and conversely, the larger the difference between the historical boiler load and the historical generator power is, the smaller the historical generator power is.
In this embodiment, the turbine incremental power characteristic model is a 2 nd order polynomial regression model.
Therefore, according to the historical boiler load, the historical generator power, the historical exhaust steam pressure and the preset first historical state characteristic data related to the historical generator power in the historical operation data, the turbine micro-power-increasing characteristic model of the system to be controlled can be obtained through training, and the influence of the historical boiler load, the historical generator power, the historical exhaust steam pressure and the preset first historical state characteristic data related to the historical generator power on the turbine micro-power-increasing characteristic model of the system to be controlled can be comprehensively considered.
Optionally, in step S202, a plurality of characteristic submodels of the system to be controlled are obtained through training according to the historical operating data, and the method may further include obtaining an exhaust pressure change characteristic model of the system to be controlled through training according to second historical state characteristic data and historical fan rotation speed in the historical operating data.
It is understood that the second historical state characteristic data in the historical operating data can be set according to actual conditions. Optionally, the second historical state feature data may be determined by a correlation analysis algorithm or a multiple-round recursive feature elimination algorithm in feature engineering. For example, the second historical state characteristic data includes, but is not limited to, historical exhaust steam pressure, historical boiler load, historical main steam pressure, historical ambient temperature, historical exhaust steam temperature, historical condensate temperature, and the like.
It is understood that the exhaust steam pressure change characteristic model may be set according to actual conditions. For example, the exhaust steam pressure variation characteristic model may be a Long Short-Term Memory (LSTM) network model.
Optionally, the exhaust steam pressure variation characteristic model is as follows:
st＝fbp((s，a)0～t-1)；
wherein s istIs the second historical state characteristic data at time t, fbpFor the time series prediction function to be trained, (s, a)0～t-1And the second historical state characteristic data corresponding to t time steps before the t moment and the historical average rotating speed of each row of fans are paired.
Therefore, according to the method, the exhaust steam pressure change characteristic model of the system to be controlled can be obtained through training according to the second historical state characteristic data and the historical fan rotating speed in the historical operating data, and the influence of the second historical state characteristic data and the historical fan rotating speed on the exhaust steam pressure change characteristic model of the system to be controlled can be comprehensively considered.
On the basis of any of the above embodiments, when the system to be controlled is an air cooling system of a thermal power generating unit, an operation optimization model may be constructed according to the power consumption characteristic submodel of each fan, the turbine micro-power-increasing characteristic model of the system to be controlled, the exhaust pressure variation characteristic model of the system to be controlled, and the model prediction control algorithm.
Optionally, the operation optimization model is:
Figure BDA0002941691150000131
Ct＝Pf；t+Pdiff；t
st+1＝f(st，at)；
s1＝sinit；
wherein argmin is a minimization function, stIs the second historical state characteristic data at time t, atFor the historical average speed of each row of fans at time t,
Figure BDA0002941691150000141
as a constraint of the second historical state feature data,
Figure BDA0002941691150000142
for the constraint condition of the historical average rotating speed of each row of fans, T is a preset future time step number, Pf；tThe historical power consumption of the fan at the time t is obtained according to the power consumption characteristic submodel, Pdiff；tObtaining the difference value between the historical boiler load and the historical generator power at the time t according to the turbine micro-power-increasing characteristic submodel, f is a dynamic model obtained according to the change of the exhaust steam pressure change characteristic model, and sinitThe second historical state characteristic data at the initial moment.
It can be understood that the constraint conditions of the second historical state characteristic data and the constraint conditions of the historical average rotating speed of each row of fans can be set according to actual conditions.
On the basis of any of the above embodiments, as shown in fig. 3, in step S103, generating a recommended value of a control parameter of the system to be controlled according to the real-time operation data, the operation optimization model, and the model predictive control online solving algorithm includes:
and S301, generating a recommended control sequence of the average rotating speed of each row of fans by adopting a preset formula according to the real-time operation data and the operation optimization model.
It is understood that the preset formula can be set according to actual conditions.
Optionally, the preset formula is:
Figure BDA0002941691150000143
μt＝η*avae+(1-η)*μ′t+1；
μ′T+1＝μ′T；
Figure BDA0002941691150000144
Figure BDA0002941691150000145
Figure BDA0002941691150000146
wherein, mu'tRecommending the tth value in the control sequence, wherein the mut is the tth value in the initial action sequence, N is the number of action sequences generated in each solving process, i is the serial number of the generated action sequences, gamma is a preset rewarding weighting factor, and R isiThe jackpot value for the ith action sequence,
Figure BDA0002941691150000147
the average rotating speed of each row of fans at the moment T corresponding to the ith action sequence is shown, eta is a preset weighting coefficient, T is a preset future time step number, avaeFor the control actions calculated from the historical control strategy model trained in advance,
Figure BDA0002941691150000148
and beta is a filter coefficient, and is a noise sample at the t moment corresponding to the ith action sequence.
Optionally, a historical control strategy model can be obtained by training according to second historical state characteristic data and historical fan rotating speed in the historical operating data. For example, a historical control strategy model may be trained based on the second historical state signature data and the historical average speed pair (s, a).
In the embodiment of the application, the type of the historical control strategy model is not limited too much. Alternatively, the historical control strategy model may be a Variational Auto-Encoders (VAE) model.
Optionally, the historical control strategy model may be defined as a second historical state characteristic data and a conditional distribution p (a | s) of the historical fan rotational speed, that is, the historical control strategy model may obtain a probability distribution of the historical fan rotational speed a according to the second historical state characteristic data s.
S302, determining a first value in the recommended control sequence as a recommended value of the average rotating speed of each row of fans at the current moment.
It is understood that, taking the preset formula as an example, the first value in the recommended control sequence is μ'1。
Therefore, according to the method, a recommended control sequence of the average rotating speed of each row of fans can be generated by adopting a preset formula according to real-time operation data and an operation optimization model, and a first value in the recommended control sequence is determined as the recommended value of the average rotating speed of each row of fans at the current moment.
As shown in fig. 4, when the system to be controlled is an air cooling system of a thermal power generating unit, a plurality of characteristic submodels of the system to be controlled can be obtained through training according to historical operating data, the plurality of characteristic submodels include a power consumption characteristic submodel of each fan, a turbine micro-power-increasing characteristic model of the system to be controlled, an exhaust pressure variation characteristic model of the system to be controlled, and a historical control strategy model, an operation optimization model of the system to be controlled can be constructed according to the plurality of characteristic submodels and a model predictive control algorithm, and then an online solving algorithm can be controlled according to real-time operating data, the operation optimization model and the model predictive control algorithm of the system to be controlled, so that a recommended value of the average rotating speed of each row of fans of the system to be controlled is generated.
Corresponding to the system optimization control method provided in the embodiments of fig. 1 to 4, the present disclosure also provides an optimization control device of the system, and since the system optimization control device provided in the embodiments of the present disclosure corresponds to the system optimization control method provided in the embodiments of fig. 1 to 4, the implementation of the system optimization control method is also applicable to the system optimization control device provided in the embodiments of the present disclosure, and will not be described in detail in the embodiments of the present disclosure.
Fig. 5 is a schematic structural diagram of an optimization control device of the system according to an embodiment of the present application.
As shown in fig. 5, the optimization control apparatus 100 of the system according to the embodiment of the present application may include: a first acquisition module 110, a second acquisition module 120, and a generation module 130.
The first obtaining module 110 is configured to obtain real-time operation data of the system to be controlled.
The second obtaining module 120 is configured to obtain a pre-constructed operation optimization model of the system to be controlled, where the operation optimization model is constructed according to historical operation data of the system to be controlled and a model predictive control algorithm.
And the generating module 130 is configured to generate a recommended value of the control parameter of the system to be controlled according to the real-time operation data, the operation optimization model, and a model predictive control online solving algorithm.
In an embodiment of the present application, as shown in fig. 6, the optimization control apparatus 100 of the system further includes: a model building module 140 for: acquiring historical operating data of the system to be controlled; training according to the historical operating data to obtain a plurality of characteristic submodels of the system to be controlled; and constructing the operation optimization model according to the characteristic submodels and the model predictive control algorithm.
In an embodiment of the application, the system to be controlled is an air cooling system of a thermal power generating unit, and the control parameter is an average rotating speed of each row of fans.
In an embodiment of the present application, the model building module 140 is specifically configured to: training to obtain a power consumption characteristic sub-model of each fan in the system to be controlled according to the historical fan rotating speed and the historical fan current in the historical operating data; training to obtain a turbine micro-power-increasing characteristic model of the system to be controlled according to historical boiler load, historical generator power, historical steam exhaust pressure and preset first historical state characteristic data related to historical generator power in the historical operation data; and training to obtain an exhaust pressure change characteristic model of the system to be controlled according to second historical state characteristic data and historical fan rotating speeds in the historical operating data.
In an embodiment of the present application, the power consumption characteristic submodel is:
I＝a1*x3+a2*x2+a3*x+b；
Figure BDA0002941691150000161
wherein I is the historical fan current;
the x is the historical fan rotating speed;
a is a1，a2，a3And b is a parameter to be trained;
the P isfThe power consumption of the fan is;
the above-mentioned
Figure BDA0002941691150000162
Setting a preset power factor value;
and U is the rated voltage of the fan.
In one embodiment of the present application, the turbine micro power augmentation characteristic model is:
Figure BDA0002941691150000163
wherein, the PdiffIs the difference between the historical boiler load and the historical generator power;
f ispIs a 2 nd order polynomial regression function to be trained;
said p isbThe historical exhaust steam pressure is obtained;
the above-mentioned
Figure BDA0002941691150000164
The first historical state characteristic data.
In one embodiment of the present application, the exhaust pressure variation characteristic model is:
st＝fbp((s，a)0～t-1)；
wherein, said stThe second historical state characteristic data at the time t;
f isbpPredicting a function for a time series to be trained;
the (s, a)0～t-1The second historical state characteristic data and the historical average rotating speed of each row of fans corresponding to t time steps before t time are obtained.
In one embodiment of the present application, the operation optimization model is:
Figure BDA0002941691150000165
Ct＝Pf；t+Pdiff；t
st+1＝f(st，at)；
s1＝sinit；
wherein the argmin is a minimization function;
s istThe second historical state characteristic data at the time t;
a is atThe historical average rotating speed of each row of fans at the time t is obtained;
the above-mentioned
Figure BDA0002941691150000171
A constraint condition for the second historical state feature data;
the above-mentioned
Figure BDA0002941691150000172
The constraint condition is the historical average rotating speed of each row of fans;
the T is a preset future time step number;
the P isf；tObtaining the historical power consumption of the fan at the time t according to the power consumption characteristic submodel;
the P isdift；tObtaining a difference value between the historical boiler load and the historical generator power at the time t according to the steam turbine micro-power-increasing characteristic submodel;
f is a dynamic model obtained according to the change of the exhaust steam pressure change characteristic model;
s isinitThe second historical state characteristic data at the initial moment.
In an embodiment of the present application, the generating module 130 is specifically configured to:
generating a recommended control sequence of the average rotating speed of each row of fans by adopting the following preset formula according to the real-time operation data and the operation optimization model;
determining a first value in the recommended control sequence as a recommended value of the average rotating speed of each row of fans at the current moment;
wherein the preset formula is as follows:
Figure BDA0002941691150000173
μt＝η*avae+(1-η)*μ′t+1；
μ′T+1＝μ′T；
Figure BDA0002941691150000174
Figure BDA0002941691150000175
Figure BDA0002941691150000176
wherein, the mu'tIs the t value in the recommended control sequence;
the mutIs the t value in the initial action sequence;
n is the number of action sequences generated in each solving process;
the i is the serial number of the generated action sequence;
the gamma is a preset rewarding weighting factor;
the R isiThe accumulated reward value corresponding to the ith action sequence;
the above-mentioned
Figure BDA0002941691150000177
The average rotating speed of each row of fans at the t moment corresponding to the ith action sequence is obtained;
the eta is a preset weighting coefficient;
the T is a preset future time step number;
a is avaeThe control action is calculated according to a historical control strategy model obtained by pre-training;
the above-mentioned
Figure BDA0002941691150000181
A noise sample at the t moment corresponding to the ith action sequence;
the beta is a filter coefficient.
In an embodiment of the present application, as shown in fig. 6, the optimization control apparatus 100 of the system further includes: a model training module 150 to: and training to obtain the historical control strategy model according to the second historical state characteristic data and the historical fan rotating speed in the historical operating data.
The optimization control device of the system of the embodiment of the application obtains real-time operation data of the system to be controlled, obtains a pre-constructed operation optimization model of the system to be controlled, constructs the operation optimization model according to historical operation data of the system to be controlled and a model predictive control algorithm, and generates a recommended value of a control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
In order to implement the above-mentioned embodiment, as shown in fig. 7, the present application further proposes an electronic device 200, including: the memory 210, the processor 220 and the computer program stored on the memory 210 and capable of running on the processor 220, when the processor 220 executes the program, the optimization control method of the system as proposed in the foregoing embodiments of the present application is implemented.
The electronic equipment of the embodiment of the application executes a computer program stored on a memory through a processor to obtain real-time operation data of a system to be controlled and obtain a pre-constructed operation optimization model of the system to be controlled, the operation optimization model is constructed according to historical operation data of the system to be controlled and a model predictive control algorithm, and a recommended value of a control parameter of the system to be controlled is generated according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
In order to implement the foregoing embodiments, the present application also proposes a computer-readable storage medium on which a computer program is stored, which, when executed by a processor, implements the optimization control method of the system as proposed by the foregoing embodiments of the present application.
The computer-readable storage medium of the embodiment of the application, by storing a computer program and executing the computer program by a processor, acquires real-time operation data of a system to be controlled, acquires a pre-constructed operation optimization model of the system to be controlled, the operation optimization model is constructed according to historical operation data of the system to be controlled and a model predictive control algorithm, and generates a recommended value of a control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control online solving algorithm. Therefore, the operation optimization model can be constructed offline by utilizing the historical operation data of the system to be controlled, time and labor cost are saved, the operation optimization model can be constructed offline by utilizing the model predictive control algorithm, the problem of operation hysteresis of the control method of the system in the related technology can be solved, stable operation of the system to be controlled is ensured, the influence of the real-time operation data of the system to be controlled on the recommended value of the control parameter can be fully considered, online solution can be carried out by utilizing the model predictive control online solution algorithm, the solution real-time performance is good, and the real-time performance of the optimization control of the system to be controlled is improved.
In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.

Claims (22)

1. A method for optimizing control of a system, comprising:
acquiring real-time operation data of the system to be controlled;
obtaining a pre-constructed operation optimization model of the system to be controlled, wherein the operation optimization model is constructed according to historical operation data of the system to be controlled and a model prediction control algorithm;
and generating a recommended value of the control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and a model predictive control on-line solving algorithm.
2. The optimization control method according to claim 1, further comprising:
acquiring historical operating data of the system to be controlled;
training according to the historical operating data to obtain a plurality of characteristic submodels of the system to be controlled;
and constructing the operation optimization model according to the characteristic submodels and the model predictive control algorithm.
3. The optimization control method according to claim 2, wherein the system to be controlled is an air cooling system of a thermal power generating unit, and the control parameter is an average rotating speed of each row of fans.
4. The optimization control method according to claim 3, wherein the training of the plurality of characteristic submodels of the system to be controlled according to the historical operating data comprises:
training to obtain a power consumption characteristic sub-model of each fan in the system to be controlled according to the historical fan rotating speed and the historical fan current in the historical operating data;
training to obtain a turbine micro-power-increasing characteristic model of the system to be controlled according to historical boiler load, historical generator power, historical steam exhaust pressure and preset first historical state characteristic data related to historical generator power in the historical operation data;
and training to obtain an exhaust pressure change characteristic model of the system to be controlled according to second historical state characteristic data and historical fan rotating speeds in the historical operating data.
5. The optimization control method according to claim 4, wherein the power consumption characteristic submodel is:
I＝a1*x3+a2*x2+a3*x+b；
Figure FDA0002941691140000011
wherein I is the historical fan current;
the x is the historical fan rotating speed;
a is a1，a2，a3And b is a parameter to be trained;
the P isfThe power consumption of the fan is;
the above-mentioned
Figure FDA0002941691140000021
Setting a preset power factor value;
and U is the rated voltage of the fan.
6. The optimization control method according to claim 4, wherein the turbine micro power augmentation characteristic model is:
Figure FDA0002941691140000025
wherein, the PdiffIs the difference between the historical boiler load and the historical generator power;
f ispIs a 2 nd order polynomial regression function to be trained;
said p isbThe historical exhaust steam pressure is obtained;
the above-mentioned
Figure FDA0002941691140000026
The first historical state characteristic data.
7. The optimization control method according to claim 4, wherein the exhaust steam pressure variation characteristic model is:
st＝fbp((s,a)0～t-1)；
wherein, said stThe second historical state characteristic data at the time t;
f isbpPredicting a function for a time series to be trained;
the (s, a)0～t-1The second historical state characteristic data and the historical average rotating speed of each row of fans corresponding to t time steps before t time are obtained.
8. The optimization control method according to claim 4, wherein the operation optimization model is:
Figure FDA0002941691140000022
Ct＝Pf；t+Pdiff；t
st+1＝f(st,at)；
s1＝sinit；
wherein the argmin is a minimization function;
s istThe second historical state characteristic data at the time t;
a is atThe historical average rotating speed of each row of fans at the time t is obtained;
the above-mentioned
Figure FDA0002941691140000023
A constraint condition for the second historical state feature data;
the above-mentioned
Figure FDA0002941691140000024
The constraint condition is the historical average rotating speed of each row of fans;
the T is a preset future time step number;
the P isf；tObtaining the historical power consumption of the fan at the time t according to the power consumption characteristic submodel;
the P isdiff；tObtaining a difference value between the historical boiler load and the historical generator power at the time t according to the steam turbine micro-power-increasing characteristic submodel;
f is a dynamic model obtained according to the change of the exhaust steam pressure change characteristic model;
s isinitThe second historical state characteristic data at the initial moment.
9. The optimization control method according to claim 8, wherein the generating of the recommended values of the control parameters of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control online solving algorithm comprises:
generating a recommended control sequence of the average rotating speed of each row of fans by adopting the following preset formula according to the real-time operation data and the operation optimization model;
determining a first value in the recommended control sequence as a recommended value of the average rotating speed of each row of fans at the current moment;
wherein the preset formula is as follows:
Figure FDA0002941691140000031
μt＝η*avae+(1-η)*μ′t+1；
μ′T+1＝μ′T；
Figure FDA0002941691140000032
Figure FDA0002941691140000033
Figure FDA0002941691140000034
wherein, the mu'tIs the t value in the recommended control sequence;
the mutIs the t value in the initial action sequence;
n is the number of action sequences generated in each solving process;
the i is the serial number of the generated action sequence;
the gamma is a preset rewarding weighting factor;
the R isiThe accumulated reward value corresponding to the ith action sequence;
the above-mentioned
Figure FDA0002941691140000035
The average rotating speed of each row of fans at the t moment corresponding to the ith action sequence is obtained;
the eta is a preset weighting coefficient;
the T is a preset future time step number;
a is avaeThe control action is calculated according to a historical control strategy model obtained by pre-training;
the above-mentioned
Figure FDA0002941691140000036
A noise sample at the t moment corresponding to the ith action sequence;
the beta is a filter coefficient.
10. The optimization control method according to claim 9, further comprising:
and training to obtain the historical control strategy model according to the second historical state characteristic data and the historical fan rotating speed in the historical operating data.
11. An optimization control apparatus for a system, comprising:
the first acquisition module is used for acquiring real-time operation data of the system to be controlled;
the second acquisition module is used for acquiring a pre-constructed operation optimization model of the system to be controlled, and the operation optimization model is constructed according to historical operation data of the system to be controlled and a model prediction control algorithm;
and the generation module is used for generating a recommended value of the control parameter of the system to be controlled according to the real-time operation data, the operation optimization model and the model predictive control on-line solving algorithm.
12. The optimizing control apparatus according to claim 11, further comprising: a model building module to:
acquiring historical operating data of the system to be controlled;
training according to the historical operating data to obtain a plurality of characteristic submodels of the system to be controlled;
and constructing the operation optimization model according to the characteristic submodels and the model predictive control algorithm.
13. The optimization control device according to claim 12, wherein the system to be controlled is an air cooling system of a thermal power generating unit, and the control parameter is an average rotating speed of each row of fans.
14. The optimizing control device according to claim 13, wherein the model building module is specifically configured to:
training to obtain a power consumption characteristic sub-model of each fan in the system to be controlled according to the historical fan rotating speed and the historical fan current in the historical operating data;
training to obtain a turbine micro-power-increasing characteristic model of the system to be controlled according to historical boiler load, historical generator power, historical steam exhaust pressure and preset first historical state characteristic data related to historical generator power in the historical operation data;
and training to obtain an exhaust pressure change characteristic model of the system to be controlled according to second historical state characteristic data and historical fan rotating speeds in the historical operating data.
15. The optimization control device according to claim 14, wherein the power consumption characteristic submodel is:
I＝a1*x3+a2*x2+a3*x+b；
Figure FDA0002941691140000051
wherein I is the historical fan current;
the x is the historical fan rotating speed;
a is a1，a2，a3And b is a parameter to be trained;
the P isfThe power consumption of the fan is;
the above-mentioned
Figure FDA0002941691140000052
Setting a preset power factor value;
and U is the rated voltage of the fan.
16. The optimizing control apparatus according to claim 14, wherein the turbine micro power characteristics model is:
Figure FDA0002941691140000053
wherein, the PdiffIs the difference between the historical boiler load and the historical generator power;
f ispIs a 2 nd order polynomial regression function to be trained;
said p isbThe historical exhaust steam pressure is obtained;
the above-mentioned
Figure FDA0002941691140000054
The first historical state characteristic data.
17. The optimization control device according to claim 14, wherein the exhaust pressure variation characteristic model is:
st＝fbp((s,a)0～t-1)；
wherein, said stThe second historical state characteristic data at the time t;
f isbpPredicting a function for a time series to be trained;
the (s, a)0～t-1The second historical state characteristic data and the historical average rotating speed of each row of fans corresponding to t time steps before t time are obtained.
18. The optimizing control apparatus according to claim 14, wherein the operation optimizing model is:
Figure FDA0002941691140000055
Ct＝Pf；t+Pdiff；t
st+1＝f(st,at)；
s1＝sinit；
wherein the argmin is a minimization function;
s istThe second historical state characteristic data at the time t;
a is atThe historical average rotating speed of each row of fans at the time t is obtained;
the above-mentioned
Figure FDA0002941691140000061
A constraint condition for the second historical state feature data;
the above-mentioned
Figure FDA0002941691140000062
The constraint condition is the historical average rotating speed of each row of fans;
the T is a preset future time step number;
the P isf；tObtaining the historical power consumption of the fan at the time t according to the power consumption characteristic submodel;
the P isdiff；tObtaining a difference value between the historical boiler load and the historical generator power at the time t according to the steam turbine micro-power-increasing characteristic submodel;
f is a dynamic model obtained according to the change of the exhaust steam pressure change characteristic model;
s isinitThe second historical state characteristic data at the initial moment.
19. The optimization and control apparatus of claim 18, wherein the generating module is specifically configured to:
generating a recommended control sequence of the average rotating speed of each row of fans by adopting the following preset formula according to the real-time operation data and the operation optimization model;
determining a first value in the recommended control sequence as a recommended value of the average rotating speed of each row of fans at the current moment;
wherein the preset formula is as follows:
Figure FDA0002941691140000063
μt＝η*avae+(1-η)*μ′t+1；
μ′T+1＝μ′T；
Figure FDA0002941691140000064
Figure FDA0002941691140000065
Figure FDA0002941691140000066
wherein, the mu'tIs the t value in the recommended control sequence;
the mutIs the t value in the initial action sequence;
n is the number of action sequences generated in each solving process;
the i is the serial number of the generated action sequence;
the gamma is a preset rewarding weighting factor;
the R isiThe accumulated reward value corresponding to the ith action sequence;
the above-mentioned
Figure FDA0002941691140000067
The average rotating speed of each row of fans at the t moment corresponding to the ith action sequence is obtained;
the eta is a preset weighting coefficient;
the T is a preset future time step number;
a is avaeThe control action is calculated according to a historical control strategy model obtained by pre-training;
the above-mentioned
Figure FDA0002941691140000068
A noise sample at the t moment corresponding to the ith action sequence;
the beta is a filter coefficient.
20. The optimizing control device according to claim 19, further comprising: a model training module to:
and training to obtain the historical control strategy model according to the second historical state characteristic data and the historical fan rotating speed in the historical operating data.
21. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of optimizing control of a system according to any one of claims 1-10 when executing the program.
22. A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, is adapted to carry out a method for optimal control of a system according to any one of claims 1-10.
CN202110181933.5A 2021-02-09 2021-02-09 Optimization control method and device of system and electronic equipment Active CN113759708B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110181933.5A CN113759708B (en) 2021-02-09 2021-02-09 Optimization control method and device of system and electronic equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110181933.5A CN113759708B (en) 2021-02-09 2021-02-09 Optimization control method and device of system and electronic equipment

Publications (2)

Publication Number Publication Date
CN113759708A true CN113759708A (en) 2021-12-07
CN113759708B CN113759708B (en) 2024-07-16

Family

ID=78786624

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110181933.5A Active CN113759708B (en) 2021-02-09 2021-02-09 Optimization control method and device of system and electronic equipment

Country Status (1)

Country Link
CN (1) CN113759708B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114187685A (en) * 2021-12-08 2022-03-15 广东好太太智能家居有限公司 Method, system, equipment and medium for optimizing power consumption of intelligent lock
CN114625091A (en) * 2022-03-21 2022-06-14 京东城市(北京)数字科技有限公司 Optimization control method and device, storage medium and electronic equipment
CN114912830A (en) * 2022-06-06 2022-08-16 京东城市(北京)数字科技有限公司 Method and device for optimizing regulation and control of heating system

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013178045A (en) * 2012-02-29 2013-09-09 Hitachi Ltd Coal-fired plant control device, and the coal-fired plant
WO2014201849A1 (en) * 2013-06-18 2014-12-24 国网辽宁省电力有限公司电力科学研究院 Method for actively optimizing, adjusting and controlling distributed wind power plant provided with energy-storage power station
CN105571343A (en) * 2014-10-31 2016-05-11 王砧 Operation back pressure continuous optimized control method and system for air-cooled generator unit steam turbine
CN107798167A (en) * 2017-09-21 2018-03-13 东南大学 Direct Air-Cooled generating set cold end system modeling and optimization method
CN110345006A (en) * 2019-03-29 2019-10-18 苏州科技大学 A kind of low wind speed area maximal power tracing optimal control method of wind power generating set
CN111338211A (en) * 2020-03-10 2020-06-26 华东理工大学 Waste heat utilization process optimization control method and system
CN111535881A (en) * 2020-05-11 2020-08-14 国电南京电力试验研究有限公司 Steam turbine optimization adjusting method

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013178045A (en) * 2012-02-29 2013-09-09 Hitachi Ltd Coal-fired plant control device, and the coal-fired plant
WO2014201849A1 (en) * 2013-06-18 2014-12-24 国网辽宁省电力有限公司电力科学研究院 Method for actively optimizing, adjusting and controlling distributed wind power plant provided with energy-storage power station
CN105571343A (en) * 2014-10-31 2016-05-11 王砧 Operation back pressure continuous optimized control method and system for air-cooled generator unit steam turbine
CN107798167A (en) * 2017-09-21 2018-03-13 东南大学 Direct Air-Cooled generating set cold end system modeling and optimization method
CN110345006A (en) * 2019-03-29 2019-10-18 苏州科技大学 A kind of low wind speed area maximal power tracing optimal control method of wind power generating set
CN111338211A (en) * 2020-03-10 2020-06-26 华东理工大学 Waste heat utilization process optimization control method and system
CN111535881A (en) * 2020-05-11 2020-08-14 国电南京电力试验研究有限公司 Steam turbine optimization adjusting method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
刘瑞奇: "金山热电厂300MW机组运行热经济性的分析研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, no. 03, pages 03 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114187685A (en) * 2021-12-08 2022-03-15 广东好太太智能家居有限公司 Method, system, equipment and medium for optimizing power consumption of intelligent lock
CN114187685B (en) * 2021-12-08 2024-07-19 广东好太太智能家居有限公司 Method, system, equipment and medium for optimizing power consumption of intelligent lock
CN114625091A (en) * 2022-03-21 2022-06-14 京东城市(北京)数字科技有限公司 Optimization control method and device, storage medium and electronic equipment
CN114912830A (en) * 2022-06-06 2022-08-16 京东城市(北京)数字科技有限公司 Method and device for optimizing regulation and control of heating system

Also Published As

Publication number Publication date
CN113759708B (en) 2024-07-16

Similar Documents

Publication Publication Date Title
CN113759708B (en) Optimization control method and device of system and electronic equipment
Li et al. Adaptive prognostic of fuel cells by implementing ensemble echo state networks in time-varying model space
Kusiak et al. Short-horizon prediction of wind power: A data-driven approach
CN113298257A (en) Method and device for creating a model of a technical system from measurements
CN116306798A (en) Ultra-short time wind speed prediction method and system
Kusiak et al. Virtual models for prediction of wind turbine parameters
CN108959787B (en) Thermal deformation prediction method and system of macro-macro dual-drive system considering actual working conditions
CN113722889A (en) Energy efficiency online analysis system and method based on artificial intelligence
CN114386563A (en) Bayesian context aggregation of neural processes
CN118170004B (en) Control method and system based on Internet of things
CN109063818A (en) A kind of thermal process model on-line identification method and device
JP2008280912A (en) Time-series data output estimation device and air-fuel ratio control device
JP2022142197A (en) Control device, control method, and program
CN117648983A (en) Time sequence cause and effect discovery method, device, equipment and medium based on intervention data
CN117277346A (en) Energy storage frequency modulation method, device and equipment based on multi-agent system
Cai et al. Dual time-scale state-coupled co-estimation of state of charge, state of health and remaining useful life for lithium-ion batteries via Deep Inter and Intra-Cycle Attention Network
CN115934691A (en) Method and device for determining short-term photovoltaic power
US20240160159A1 (en) Control device for controlling a technical system, and method for configuring the control device
CN115419908A (en) Control method of steam-flue gas heat exchanger based on fuzzy neural network
CN116123028A (en) Wind power plant level MPPT prediction model control method and device
Monti et al. Extending polynomial chaos to include interval analysis
Zhang et al. An Improved Soft Actor-Critic-Based Energy Management Strategy of Fuel Cell Hybrid Vehicles with a Nonlinear Fuel Cell Degradation Model
CN111356959B (en) Method for computer-aided control of a technical system
CN117117858B (en) Wind turbine generator power prediction method, device and storage medium
CN113447813A (en) Fault diagnosis method and equipment for offshore wind generating set

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