CN113777923A - Thermal power plant intelligent control module based on GRU neural network and operation method thereof - Google Patents
Thermal power plant intelligent control module based on GRU neural network and operation method thereof Download PDFInfo
- Publication number
- CN113777923A CN113777923A CN202111056919.9A CN202111056919A CN113777923A CN 113777923 A CN113777923 A CN 113777923A CN 202111056919 A CN202111056919 A CN 202111056919A CN 113777923 A CN113777923 A CN 113777923A
- Authority
- CN
- China
- Prior art keywords
- neural network
- data
- unit
- machine learning
- control
- 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.)
- Pending
Links
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 15
- 238000010801 machine learning Methods 0.000 claims abstract description 46
- 238000012545 processing Methods 0.000 claims abstract description 40
- 238000004891 communication Methods 0.000 claims abstract description 24
- 230000008569 process Effects 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 45
- 238000003062 neural network model Methods 0.000 claims description 36
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000012360 testing method Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 7
- 210000002569 neuron Anatomy 0.000 claims description 7
- 230000007246 mechanism Effects 0.000 claims description 6
- 238000010977 unit operation Methods 0.000 claims description 5
- 238000013524 data verification Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 238000004140 cleaning Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 claims description 2
- 230000005856 abnormality Effects 0.000 claims description 2
- 238000013075 data extraction Methods 0.000 claims description 2
- 230000003247 decreasing effect Effects 0.000 claims description 2
- 238000012795 verification Methods 0.000 claims description 2
- 230000008901 benefit Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000009833 condensation Methods 0.000 description 1
- 230000005494 condensation Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Automation & Control Theory (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a thermal power plant intelligent control module based on a GRU neural network and an operation method thereof, wherein the thermal power plant intelligent control module comprises a communication unit, a data processing unit, a storage unit and a machine learning platform; the invention learns and processes the operation data of the heat supply of the power plant by using the GRU neural network, can completely replace integral control and differential control in PID control by specifically updating and resetting the influence degree of the state of the GRU neural network on the state of the power plant at the previous moment, and has more trainable parameters and higher flexibility compared with the integral control and the differential control in the traditional PID control. According to the intelligent control module of the thermal power plant based on the GRU neural network and the operation method thereof, the neural network is trained and updated regularly, and the control logic is kept in an optimal state, so that the optimal control of a unit is achieved, the operation and debugging cost of the thermal power plant is reduced, and the operation benefit of an enterprise is improved.
Description
Technical Field
The invention belongs to the field of automatic control, and particularly relates to an intelligent control module of a thermal power plant based on a GRU neural network and an operation method thereof.
Background
The control logic of the existing thermal power plant control system usually compares the controlled quantity with a set value, and outputs a control signal after the operation of a PID regulator, so that the controlled quantity is equal to or close to the set value finally. However, the debugging and operation of such a control system mainly have the following three problems.
1. After debugging is finished, the general PID control logic is determined, and along with aging or transformation of a unit, the operation quality of the control logic is reduced, so that the response speed and the stability of the controlled quantity are reduced. When the unit is in a long time or has large deviation, the control logic needs to be modified again, so that the operation cost of the power plant is increased, and the operation safety of the unit is negatively influenced;
2. the design and debugging of the control logic need professional automation control personnel to carry out long-time working condition adjustment and test, and the operating working condition must be debugged within a larger range of the set value of the controlled quantity as much as possible so as to cover all possible operating working conditions. Causing a large amount of labor consumption and bringing certain safety risks.
3. The historical data accumulated during long-term operation of the power plant is difficult to utilize in manual debugging, and the PID or fuzzy self-adaptive PID control logic obtained by manual debugging often shows greater instability when the unit operation parameters are dynamically adjusted. In the test process, certain parameters of the unit are stable, and even 1 hour is needed, so that the requirement of an optimization strategy on parameter control is difficult to meet.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an intelligent control module of a thermal power plant based on a GRU neural network and an operation method thereof.
The invention is realized by the following technical scheme:
an intelligent thermal power plant control module based on a GRU neural network comprises a communication unit, a data processing unit, a storage unit and a machine learning platform;
the communication unit is used for acquiring real-time data of the operation of the power plant and sending a control instruction obtained by the calculation of the machine learning platform to the DCS or the execution mechanism;
the data processing unit is used for extracting, cleaning, converting, removing the weight, sorting and eliminating the operation data acquired by the communication module and sending the operation data to the machine learning platform as required;
the storage unit is used for storing the processed operation data and the weight information of the GRU neural network;
and the machine learning platform is used for training and testing the pre-built GRU neural network according to a plan and carrying out optimization control on real-time operation parameters of the unit.
The communication unit acquires the operation data of the unit and sends a control command in three modes, namely a point-to-point hard wiring mode is used, a Modbus communication protocol is used, and the communication unit communicates with an OPC server to acquire and send data.
The invention is further improved in that the data processing unit operates in three states, one is to process the data received in real time, and the content includes: adding a time column, calculating a moving average value of the time column, calculating a fitted curve gradient, judging data abnormality and writing out abnormal information; secondly, carry out data processing when machine learning platform trains and tests, its content includes: data extraction, data sorting, normalization processing, data verification, data scrambling and data set division; and thirdly, when the control command is forwarded, judging the time interval of the control command and the change amplitude of the controlled quantity to realize the locking increasing/decreasing function.
The invention has the further improvement that the storage unit receives the data processed by the data processing unit and the neural network weight data obtained by the training of the machine learning platform and stores the data into the corresponding database, and when a data reading instruction is received, the storage unit extracts the stored data and sends the data to the data processing unit or the machine learning platform.
The invention is further improved in that the machine learning platform is composed of a model training unit, a model updating unit, a planning task unit and a control calculation unit.
The invention has the further improvement that the model training unit is responsible for training, testing and verifying the GRU neural network, the model updating unit is used for comparing and updating the trained GRU neural network with the current GRU neural network, the planning task unit is used for arranging the model training task and the model updating task, the control computing unit is internally provided with a linear control algorithm, and can use the existing GRU neural network model and weight information to compute a control instruction corresponding to a target parameter and send the control instruction to the data processing unit in real time.
The method is further improved in that a pre-constructed GRU neural network model is carried on the machine learning platform, the neural network model comprises three hidden layers, and all neurons of the hidden layers are GRU neurons.
An operation method of an intelligent control module of a thermal power plant based on a GRU neural network comprises the following steps:
step 1: initializing a machine learning platform: initializing a GRU neural network, and marking a machine learning platform state code as 0 xff; initializing a planning task unit, and arranging a neural network model training task and a neural network model updating task, wherein all weights of the GRU neural network model are random values at the moment, and a control instruction is generated by a linear control algorithm built in a control calculation unit;
step 2: the communication unit acquires unit operation data and forwards the unit operation data to the data processing unit in real time;
step 3: after receiving the data forwarded by the communication unit, the data processing unit adds a time mark to the data according to the current time, obtains the running data at a fixed time interval through an interpolation algorithm, calculates a moving average value and calculates a fitted curve gradient, cleans and arranges the running data, and finally sends the calculated data and the original data to the storage unit;
step 4: the storage unit receives the operation data sent by the data processing unit and stores the operation data into the time sequence database;
step 5: when the local time reaches the training task time of the neural network model arranged by the planning task unit, starting the training of the neural network model, at the moment, firstly sending a model information extracting instruction to a storage unit by the machine learning platform to obtain the weight information of the neural network model, then sending a data set request instruction to a data processing unit, extracting the running data stored in the storage unit by the data processing unit, and sending the running data to the machine learning platform after data sorting, normalization processing, data verification, data scrambling and data set division;
step 6: after a model training unit in the machine learning platform obtains a training data set, judging whether the data volume of the data set meets a preset requirement, if so, starting model training, otherwise, ending a neural network model training task;
step 7: when the local time reaches the moment of a neural network model updating task arranged by a planning task unit and the current neural network model training task is completed, starting to update the neural network model;
step 8: after a model updating unit in the machine learning platform obtains a verification data set, testing and comparing the trained model with the existing model, and if the loss function of the trained model is smaller than that of the existing model, updating weight information of the neural network model in a storage unit;
step 9: after the model updating unit finishes one-time GRU neural network model weight updating, the state code of the machine learning platform is set to be 0x 00;
step 10: when the control calculation unit finds that the machine learning platform state code is 0x00, loading a GRU neural network model and weight information thereof, and calculating a control instruction by using the GRU neural network;
step 11: and the data processing unit checks the control instruction obtained by calculation of the machine learning platform, discards the current control instruction when the deviation between the control instruction and the current execution mechanism state exceeds a given threshold value, and otherwise, sends the instruction to the main DCS of the unit or the execution mechanism through the communication unit to complete the control operation.
Compared with the prior art, the invention has at least the following beneficial technical effects:
according to the intelligent control module of the thermal power plant based on the GRU neural network and the operation method thereof, the GRU neural network is used for learning and processing the operation data of the heat supply of the thermal power plant, and the influence degree of the state of the previous moment on the state of the later moment is controlled through specific updating and resetting, so that the integral control and the differential control in PID control can be completely replaced, and compared with the integral control and the differential control in the traditional PID control, the intelligent control module of the thermal power plant based on the GRU neural network has more trainable parameters and has higher flexibility.
According to the intelligent control module of the thermal power plant based on the GRU neural network and the operation method thereof, the neural network is trained and updated regularly, and the control logic is kept in an optimal state, so that the optimal control of a unit is achieved, the heat consumption and the coal consumption of the thermal power plant are reduced, the operation debugging cost of the power plant is reduced, and the operation benefit of an enterprise is improved.
Drawings
FIG. 1 is a schematic structural diagram of an intelligent thermal power plant control module based on a GRU neural network according to the present invention;
FIG. 2 is a schematic structural diagram of a machine learning platform in an intelligent thermal power plant control module based on a GRU neural network;
FIG. 3 is a schematic structural diagram of GRU neurons in an intelligent thermal power plant control module based on a GRU neural network;
FIG. 4 is a schematic diagram of a structure and a flow of an intelligent thermal power plant control module based on a GRU neural network for real-time control;
fig. 5 is a schematic flow chart of a thermal power plant intelligent control module based on a GRU neural network when training and updating a GRU neural network model.
Detailed Description
The present invention will now be described in further detail with reference to specific examples, which are intended to be illustrative, but not limiting, of the invention.
Taking a certain power plant as an example, the total installed capacity of the power plant is 400MW, wherein the No. 1 unit is a 50MW back pressure unit with middle steam extraction, and the No. 2 unit is a supercritical 350MW steam extraction and condensation type unit. The daily steam consumption of a heat user is about 280t/h, and the maximum steam consumption is about 400 t/h. The operation parameters required to be collected by the intelligent control module comprise main steam, temperature and pressure of four-stage steam extraction, real-time reading of site total load, medium-pressure steam supply total amount, low-pressure steam supply total amount and the like, and the controlled amount is the opening of the No. 2 machine low-pressure steam supply hydraulic control butterfly valve.
As shown in fig. 1, the thermal power plant intelligent control module based on the GRU neural network provided by the invention includes a communication unit, a data processing unit, a storage unit and a machine learning platform.
The communication unit is used for acquiring real-time data of the operation of the power plant and sending a control instruction obtained by the calculation of the machine learning platform to the DCS;
the data processing unit is used for extracting, cleaning, converting, removing the weight, sorting and eliminating the operation data acquired by the communication module and sending the operation data to the machine learning platform as required;
the storage unit is used for storing the processed operation data and the weight information of the GRU neural network;
and the machine learning platform is used for training and testing the pre-built GRU neural network according to a plan and carrying out optimization control on real-time operation parameters of the unit.
As shown in fig. 2, the machine learning platform is composed of a model training unit, a model updating unit, a planning task unit, and a control calculating unit. The model training unit is responsible for training, testing and verifying the GRU neural network. The model updating unit compares and updates the trained GRU neural network with the current GRU neural network, the planning task unit arranges a model training task and a model updating task, and the control computing unit is internally provided with a linear control algorithm and can use the existing GRU neural network model and weight information to compute a control instruction corresponding to the target parameter and send the control instruction to the data processing unit in real time.
As shown in fig. 3, which is a schematic structural diagram of a typical GRU neuron, when a control instruction is calculated by the GRU neural network, a calculation formula of an acquired operation parameter in the GRU neuron is:
rt＝σ(Wr·[ht-1,xt])
zt＝σ(Wz·[ht-1,xt])
yt＝σ(Wo·ht)
wherein x and y are normalized data corresponding to input operation parameters and output control instructions, and Wr、Wz、
WoFor the weight matrix at different nodes in the neural network, σ and tanh are activation functions, and subscripts t-1 and t denote the previous time and the current time, respectively.
As shown in fig. 4, when the intelligent control module performs real-time control, the communication unit collects operation data and forwards the operation data to the data processing unit, the data processing unit reads the operation data in a previous period of time and the operation data at the current moment, the operation data is pre-processed by adding a time column, the time column and the like to form calculation data, the calculation data is sent to the control calculation unit in the machine learning platform, the control calculation unit judges whether a machine learning platform state code is 0x00, if so, the current GRU neural network weight information is loaded from the storage unit, and the GRU neural network is used for calculating a control instruction, otherwise, the preset linear control algorithm is used for calculating the control instruction. The control command obtained by the control calculation unit can be sent to the data processing unit in real time, the data processing unit judges the time interval of the control command and the change amplitude of the controlled quantity, and when the change rate of the controlled quantity and the change rate of the control command meet set conditions, the control command is sent to the communication unit, and finally the controlled quantity, namely the opening degree of the No. 2 low-pressure steam supply hydraulic control butterfly valve is controlled. When the controlled quantity change rate and the control instruction change rate do not meet the set conditions, the control instruction is discarded, the forwarding of the next control instruction is waited, and the control instruction maintains the value at the previous moment unchanged in the waiting period.
As shown in fig. 5, when the intelligent control module performs model training, first, the model training unit in the machine learning platform loads current GRU neural network weight information from the storage unit; the data processing unit extracts historical operating data from the storage unit, assembles the historical operating data into a training set and a test data set, sends the training set to the model training unit, and sends the test set to the model updating unit; the model training unit judges the data volume of the training data set, if the data volume does not meet the training requirement, the model training task is quitted, and if the data volume meets the training requirement, the GRU neural network is trained to obtain updated GRU neural network weight and loss function information; and after the training is finished, the model updating unit loads a new model and calculates the loss function of the new model and the current GRU neural network model, if the loss function of the new model is smaller, the weight information of the GRU neural network model is updated, and if not, the model updating task is exited.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all the intelligent control modules formed by equivalent structures made by using the contents of the present specification and the attached drawings are included in the scope of the present invention.
Claims (8)
1. An intelligent thermal power plant control module based on a GRU neural network is characterized by comprising a communication unit, a data processing unit, a storage unit and a machine learning platform;
the communication unit is used for acquiring real-time data of the operation of the power plant and sending a control instruction obtained by the calculation of the machine learning platform to the DCS or the execution mechanism;
the data processing unit is used for extracting, cleaning, converting, removing the weight, sorting and eliminating the operation data acquired by the communication module and sending the operation data to the machine learning platform as required;
the storage unit is used for storing the processed operation data and the weight information of the GRU neural network;
and the machine learning platform is used for training and testing the pre-built GRU neural network according to a plan and carrying out optimization control on real-time operation parameters of the unit.
2. The intelligent thermal power plant control module based on the GRU neural network as claimed in claim 1, wherein the communication unit collects operation data of the thermal power plant and sends control commands in three ways, namely, a point-to-point hard wiring way is used, a Modbus communication protocol is used, and the communication with an OPC server is used for acquiring and sending data.
3. The intelligent thermal power plant control module based on the GRU neural network as claimed in claim 1, wherein the data processing unit operates in three states, one is to process the data received in real time, and its content includes: adding a time column, calculating a moving average value of the time column, calculating a fitted curve gradient, judging data abnormality and writing out abnormal information; secondly, carry out data processing when machine learning platform trains and tests, its content includes: data extraction, data sorting, normalization processing, data verification, data scrambling and data set division; and thirdly, when the control command is forwarded, judging the time interval of the control command and the change amplitude of the controlled quantity to realize the locking increasing/decreasing function.
4. The intelligent thermal power plant control module based on the GRU neural network as claimed in claim 1, wherein the storage unit receives the data processed by the data processing unit and the neural network weight data obtained by training of the machine learning platform, stores the data in a corresponding database, and when a read data instruction is received, the storage unit extracts the stored data and sends the data to the data processing unit or the machine learning platform.
5. The intelligent thermal power plant control module based on the GRU neural network as claimed in claim 1, wherein the machine learning platform is composed of a model training unit, a model updating unit, a planning task unit and a control computing unit.
6. The intelligent thermal power plant control module based on the GRU neural network as claimed in claim 5, wherein the model training unit is responsible for training, testing and verifying the GRU neural network, the model updating unit is used for comparing and updating the trained GRU neural network with the current GRU neural network, the planning task unit is used for arranging a model training task and a model updating task, the control computing unit is internally provided with a linear control algorithm, and can use the existing GRU neural network model and weight information to compute a control instruction corresponding to a target parameter and send the control instruction to the data processing unit in real time.
7. The intelligent thermal power plant control module based on the GRU neural network as claimed in claim 6, wherein a pre-constructed GRU neural network model is loaded on the machine learning platform, the neural network model comprises three hidden layers, and all neurons of the hidden layers are GRU neurons.
8. The method for operating the intelligent control module of the thermal power plant based on the GRU neural network as claimed in claim 7, characterized by comprising the following steps:
step 1: initializing a machine learning platform: initializing a GRU neural network, and marking a machine learning platform state code as 0 xff; initializing a planning task unit, and arranging a neural network model training task and a neural network model updating task, wherein all weights of the GRU neural network model are random values at the moment, and a control instruction is generated by a linear control algorithm built in a control calculation unit;
step 2: the communication unit acquires unit operation data and forwards the unit operation data to the data processing unit in real time;
step 3: after receiving the data forwarded by the communication unit, the data processing unit adds a time mark to the data according to the current time, obtains the running data at a fixed time interval through an interpolation algorithm, calculates a moving average value and calculates a fitted curve gradient, cleans and arranges the running data, and finally sends the calculated data and the original data to the storage unit;
step 4: the storage unit receives the operation data sent by the data processing unit and stores the operation data into the time sequence database;
step 5: when the local time reaches the training task time of the neural network model arranged by the planning task unit, starting the training of the neural network model, at the moment, firstly sending a model information extracting instruction to a storage unit by the machine learning platform to obtain the weight information of the neural network model, then sending a data set request instruction to a data processing unit, extracting the running data stored in the storage unit by the data processing unit, and sending the running data to the machine learning platform after data sorting, normalization processing, data verification, data scrambling and data set division;
step 6: after a model training unit in the machine learning platform obtains a training data set, judging whether the data volume of the data set meets a preset requirement, if so, starting model training, otherwise, ending a neural network model training task;
step 7: when the local time reaches the moment of a neural network model updating task arranged by a planning task unit and the current neural network model training task is completed, starting to update the neural network model;
step 8: after a model updating unit in the machine learning platform obtains a verification data set, testing and comparing the trained model with the existing model, and if the loss function of the trained model is smaller than that of the existing model, updating weight information of the neural network model in a storage unit;
step 9: after the model updating unit finishes one-time GRU neural network model weight updating, the state code of the machine learning platform is set to be 0x 00;
step 10: when the control calculation unit finds that the machine learning platform state code is 0x00, loading a GRU neural network model and weight information thereof, and calculating a control instruction by using the GRU neural network;
step 11: and the data processing unit checks the control instruction obtained by calculation of the machine learning platform, discards the current control instruction when the deviation between the control instruction and the current execution mechanism state exceeds a given threshold value, and otherwise, sends the instruction to the main DCS of the unit or the execution mechanism through the communication unit to complete the control operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111056919.9A CN113777923A (en) | 2021-09-09 | 2021-09-09 | Thermal power plant intelligent control module based on GRU neural network and operation method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111056919.9A CN113777923A (en) | 2021-09-09 | 2021-09-09 | Thermal power plant intelligent control module based on GRU neural network and operation method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113777923A true CN113777923A (en) | 2021-12-10 |
Family
ID=78842155
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111056919.9A Pending CN113777923A (en) | 2021-09-09 | 2021-09-09 | Thermal power plant intelligent control module based on GRU neural network and operation method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113777923A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107578124A (en) * | 2017-08-28 | 2018-01-12 | 国网山东省电力公司电力科学研究院 | The Short-Term Load Forecasting Method of GRU neutral nets is improved based on multilayer |
| CN109185110A (en) * | 2018-08-18 | 2019-01-11 | 天津绿动未来能源管理有限公司 | A kind of generating plant circulation-water optimizing optimizing operation method |
| CN109376848A (en) * | 2018-09-01 | 2019-02-22 | 哈尔滨工程大学 | A kind of door control unit neural network of simplification |
| CN109711620A (en) * | 2018-12-26 | 2019-05-03 | 浙江大学 | A kind of Short-Term Load Forecasting Method based on GRU neural network and transfer learning |
| CN112698572A (en) * | 2020-12-22 | 2021-04-23 | 西安交通大学 | Structural vibration control method, medium and equipment based on reinforcement learning |
| CN113013893A (en) * | 2021-02-08 | 2021-06-22 | 国网河南省电力公司电力科学研究院 | New energy station self-adaptive frequency control method and system |
| CN113107626A (en) * | 2021-05-25 | 2021-07-13 | 浙大城市学院 | Combined cycle generator set load prediction method based on multivariable LSTM |
-
2021
- 2021-09-09 CN CN202111056919.9A patent/CN113777923A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107578124A (en) * | 2017-08-28 | 2018-01-12 | 国网山东省电力公司电力科学研究院 | The Short-Term Load Forecasting Method of GRU neutral nets is improved based on multilayer |
| CN109185110A (en) * | 2018-08-18 | 2019-01-11 | 天津绿动未来能源管理有限公司 | A kind of generating plant circulation-water optimizing optimizing operation method |
| CN109376848A (en) * | 2018-09-01 | 2019-02-22 | 哈尔滨工程大学 | A kind of door control unit neural network of simplification |
| CN109711620A (en) * | 2018-12-26 | 2019-05-03 | 浙江大学 | A kind of Short-Term Load Forecasting Method based on GRU neural network and transfer learning |
| CN112698572A (en) * | 2020-12-22 | 2021-04-23 | 西安交通大学 | Structural vibration control method, medium and equipment based on reinforcement learning |
| CN113013893A (en) * | 2021-02-08 | 2021-06-22 | 国网河南省电力公司电力科学研究院 | New energy station self-adaptive frequency control method and system |
| CN113107626A (en) * | 2021-05-25 | 2021-07-13 | 浙大城市学院 | Combined cycle generator set load prediction method based on multivariable LSTM |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109062053B (en) | Denitration ammonia injection control method based on multivariate correction | |
| Wang et al. | Fault-tolerant controller design for a class of nonlinear MIMO discrete-time systems via online reinforcement learning algorithm | |
| US5305230A (en) | Process control system and power plant process control system | |
| Jin | Decentralized adaptive fuzzy control of robot manipulators | |
| Zhou et al. | Deep reinforcement learning based real-time AC optimal power flow considering uncertainties | |
| CN105955026B (en) | Fuzzy PID control method, device and system | |
| CN114839880B (en) | Self-adaptive control method based on flexible joint mechanical arm | |
| CN114839881B (en) | Intelligent garbage cleaning and environmental parameter big data Internet of things system | |
| Wang et al. | Stochastic configuration network based cascade generalized predictive control of main steam temperature in power plants | |
| Bonassi et al. | An Offset-Free Nonlinear MPC scheme for systems learned by Neural NARX models | |
| Zhang et al. | Decentralized adaptive output feedback dynamic surface control for stochastic nonstrict‐feedback interconnected nonlinear systems with actuator failures and input quantization via command filter | |
| CN111077771A (en) | Self-tuning fuzzy PID control method | |
| Farag et al. | Design and implementation of a variable-structure adaptive fuzzy-logic yaw controller for large wind turbines | |
| Xia et al. | Distributed adaptive cooperative control via command filters for multi-agent systems including input unmodeled dynamics and sensor faults | |
| CN114386320A (en) | Steam turbine valve management method, device, equipment and storage medium | |
| KR102317935B1 (en) | Dualized closed-loop data analytics system | |
| CN113777923A (en) | Thermal power plant intelligent control module based on GRU neural network and operation method thereof | |
| CN116300755A (en) | Double-layer optimal scheduling method and device for heat storage-containing heating system based on MPC | |
| CN114139442A (en) | Method for optimizing steam turbine valve flow characteristic function based on K-means-LSTM | |
| Banderchuk et al. | Combining robust control and machine learning for uncertain nonlinear systems subject to persistent disturbances | |
| Li et al. | Fault‐Tolerant Control of Wind Turbine System Using Linear Parameter‐Varying Model | |
| CN116484974B (en) | Flow control system of jetting valve based on magneto-rheological technology | |
| CN115016275B (en) | Intelligent feeding and livestock house big data Internet of things system | |
| CN115254529B (en) | Dispensing compensation method and system based on time and pressure | |
| CN111459030B (en) | Self-adaptive modeling method for closed-loop combustion optimization of boiler |
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 |