CN118642412A - An adaptive control system for unit sliding pressure optimization operation - Google Patents

Abstract

The invention relates to the technical field of power system control, and discloses a self-adaptive adjustment unit sliding pressure optimizing operation control system, which comprises the following components: the self-adaptive data acquisition module is used for acquiring and preprocessing the operation data of the set in a high-precision and multi-dimensional manner; the real-time state monitoring module is connected with the self-adaptive data acquisition module and is used for dynamically monitoring and estimating the running state of the unit and detecting the abnormality and the change in real time; and the intelligent optimization decision module is connected with the high-efficiency execution control module and is used for making and optimizing a sliding pressure control strategy. The intelligent optimization decision module, the high-efficiency execution control module and the self-learning feedback module cooperate to realize real-time optimization of the sliding pressure control strategy, dynamic adjustment of control parameters and continuous optimization of the control strategy, improve the self-adaptive capacity and optimization effect of the system, and solve the problems of insufficient real-time performance of the optimization decision, insufficient accuracy and dynamic adjustment capacity of the control strategy and limited optimization and adjustment capacity of the control strategy in the prior art.

Description

An adaptive control system for unit sliding pressure optimization operation

Technical Field

The present invention relates to the technical field of power system control, and in particular to a self-adaptively adjusted unit sliding pressure optimization operation control system.

Background Art

With the continuous growth of global energy demand and increasingly stringent environmental regulations, the power industry is facing the dual pressure of improving power generation efficiency and reducing pollution emissions. As one of the main ways of power production, the operating efficiency and emission level of thermal power units directly affect the overall efficiency and environmental impact of the power system. Therefore, optimizing the operation control strategy of thermal power units and improving their economic and environmental performance have become the hot topics of current research.

The traditional sliding pressure operation control system of thermal power units mainly relies on preset fixed parameters and simple PID control to adjust the main steam pressure, so that the turbine can maintain efficient operation under different loads. However, this method has many limitations. First, the real-time and comprehensive nature of the optimization decision is insufficient. Traditional systems usually adopt single-objective optimization, such as minimizing the heat rate, which makes it difficult to comprehensively consider power generation efficiency and emission targets. In addition, optimization decisions are often based on fixed models and cannot be dynamically adjusted according to real-time operating conditions, resulting in unsatisfactory operating results.

In view of the above problems, the present invention provides a self-adaptive adjustment unit sliding pressure optimization operation control system.

Summary of the invention

In view of the shortcomings of the prior art, the present invention provides an adaptively adjusted unit sliding pressure optimization operation control system, which solves the problems that the optimization decision of the sliding pressure operation control system of traditional thermal power units cannot be adjusted in real time and has a single target, the control strategy execution accuracy and dynamic adjustment capability are insufficient, and the control strategy optimization and adjustment capability is limited.

To achieve the above objectives, the present invention is implemented through the following technical solutions: a self-adaptive adjustment unit sliding pressure optimization operation control system, comprising:

Adaptive data acquisition module, which is used to collect and pre-process unit operation data with high precision and multi-dimensionality;

A real-time status monitoring module, which is connected to the adaptive data acquisition module, is used to dynamically monitor and estimate the unit's operating status and detect anomalies and changes in real time;

Intelligent optimization decision module, which is connected with the efficient execution control module, is used to formulate and optimize the sliding pressure control strategy to make the unit operate in the optimal state;

An efficient execution control module, which is connected to the real-time status monitoring module and the self-learning feedback module to implement precise control strategies and dynamically adjust key operating parameters;

The self-learning feedback module is connected to the intelligent optimization decision-making module to continuously monitor and learn the operating conditions, optimize and adjust the control strategy.

Preferably, the adaptive data acquisition module comprises:

Dynamic sampling unit, which is used to dynamically adjust the sampling frequency and range according to the unit operation status and load changes;

Multi-source data fusion unit, which is used to combine multiple sensor data to achieve comprehensive monitoring of key parameters;

Intelligent pre-processing unit for cleaning and processing raw data through adaptive filtering and noise suppression algorithms.

Preferably, the real-time status monitoring module includes:

A state estimation unit, which is used to estimate the real-time operating state of the unit using Kalman filtering and particle filtering methods;

Anomaly detection unit, which is used to detect anomalies in real time using statistical methods and machine learning models;

A change identification unit is used to identify changes in load and operating conditions based on a change point detection algorithm.

Preferably, the intelligent optimization decision module includes:

An adaptive optimization unit, which is used to design an adaptive dynamic programming algorithm to optimize the sliding pressure control strategy in real time;

Multi-objective optimization unit, which is used to simultaneously consider multiple objectives such as heat rate, power generation efficiency and emissions and use evolutionary algorithms for optimization;

Online learning unit, which is used to continuously update the optimization model using real-time data in combination with deep reinforcement learning.

Preferably, the efficient execution control module includes:

An adaptive PID control unit, which is used to design a model-based adaptive PID controller to dynamically adjust control parameters;

A fuzzy logic control unit for optimizing the adjustment of the main steam pressure by combining fuzzy logic to handle nonlinearity and uncertainty;

A coordinated control unit for achieving coordinated optimization of boilers, turbines and generators through a multivariable control strategy;

A blockchain verification unit is used for distributed storage and verification of control instructions through blockchain technology.

Preferably, the self-learning feedback module includes:

A reinforcement learning unit, which is used to continuously optimize the control strategy through analysis and learning of operating data;

A data-driven modeling unit, which is used to build a data-driven unit operation model using big data analysis technology and machine learning algorithms;

Intelligent diagnostic unit, which is used to combine expert system and digital twin technology to intelligently diagnose and process abnormalities and faults.

Preferably, the dynamic sampling frequency adjustment formula in the dynamic sampling unit is:

,in, is the sampling frequency at time t, is the initial sampling frequency, is the sampling frequency adjustment factor, is the change in pressure at time t, is the base pressure value.

Preferably, the adaptive dynamic programming optimization formula in the adaptive optimization unit combines the immediate cost and the future expected cost to optimize the control strategy, and the formula is:

,in, is the optimal value function of the dynamic s, is the instantaneous cost of state s and dynamic u, is the discount factor, After taking action u in state s, the state is transferred to probability.

Preferably, the adaptive PID control parameter adjustment formula in the adaptive PID control unit is:

)

= (1+

),in, , , are the proportional, integral and derivative gains at time t, , , is the initial gain, , , is the gain adjustment coefficient, and e(t) is the error value at time t.

Preferably, the adaptive PID control unit involves a dynamic sliding pressure setting value adjustment formula, which dynamically adjusts the sliding pressure setting value mainly according to the load and operating status to ensure that the unit operates in the best working condition. The formula is as follows:

,in, is the sliding pressure setting value at time t, is the reference sliding pressure value, is the sliding pressure adjustment coefficient, L(t) is the load at time t, is the base load.

Working principle: When the unit starts, the adaptive data acquisition module starts working immediately. This module adjusts the frequency and range of data acquisition according to the current load and operating status through the dynamic sampling unit to ensure the real-time and accuracy of the collected data. The multi-source data fusion unit receives data from multiple sensors and processes them comprehensively to ensure the integrity and accuracy of the data. Then, the intelligent pre-processing unit cleans and processes the raw data through adaptive filtering and noise suppression algorithms to remove noise and outliers, thereby ensuring data quality. These pre-processed data are transmitted to the real-time status monitoring module.

After the real-time status monitoring module receives the processed data, the state estimation unit uses Kalman filtering and particle filtering methods to estimate the operating status of the unit in real time. The anomaly detection unit uses statistical methods and machine learning models to detect anomalies in the data in real time and issue an alarm in time. When an anomaly or state change is detected, the change identification unit identifies and records the changes in load and operating conditions based on the change point detection algorithm. The processed state information and anomaly detection results are transmitted to the intelligent optimization decision module and the efficient execution control module.

After the intelligent optimization decision module receives the status information, the adaptive optimization unit applies the adaptive dynamic programming algorithm based on this information to optimize the sliding pressure control strategy in real time. The multi-objective optimization unit comprehensively considers the heat rate, power generation efficiency and emission targets, and optimizes through the evolutionary algorithm. The online learning unit combines the deep reinforcement learning method and uses real-time data to continuously update and improve the optimization model to ensure the accuracy and real-time performance of the optimization strategy. The generated optimized control strategy is transmitted to the efficient execution control module.

After the efficient execution control module receives the optimized control strategy, the adaptive PID control unit dynamically adjusts the parameters of the PID controller according to the real-time error. The fuzzy logic control unit handles the nonlinearity and uncertainty of the system, performs more refined adjustments, and optimizes the adjustment of the main steam pressure. The coordination control unit uses a multivariable control strategy to collaboratively optimize the operation of the boiler, turbine, and generator to ensure the coordination and optimization of each part. The blockchain verification unit uses blockchain technology to perform distributed storage and verification of control instructions to ensure the security and non-tamperability of data. At the same time, the efficient execution control module dynamically adjusts the sliding pressure set value according to the real-time load and operating status to ensure that the unit operates in the best state under different loads. The execution results and feedback information are transmitted to the self-learning feedback module and the real-time status monitoring module.

After receiving the execution results and feedback information, the self-learning feedback module analyzes and learns these data through the reinforcement learning unit to continuously optimize the control strategy. The data-driven modeling unit uses big data analysis and machine learning technology to build and update the unit operation model. The intelligent diagnosis unit combines expert systems and digital twin technology to intelligently diagnose and handle abnormalities and faults. The optimized model and strategy information are transmitted back to the intelligent optimization decision module to form a closed-loop feedback, thereby ensuring that the system can be continuously optimized under different operating conditions and loads.

The present invention provides a self-adaptive adjustment unit sliding pressure optimization operation control system, which has the following beneficial effects:

1. The present invention uses an intelligent optimization decision module, combined with an adaptive dynamic programming algorithm and a multi-objective optimization algorithm, to optimize the sliding pressure control strategy in real time, and comprehensively considers the heat consumption rate, power generation efficiency and emission targets to ensure the optimal overall operating effect. It solves the problem that the optimization decisions in the prior art often cannot be adjusted in real time, the optimization targets are single, and there is a lack of comprehensive consideration.

2. The present invention uses an efficient execution control module, adaptive PID control and fuzzy logic control, and dynamically adjusts control parameters to achieve precise control and dynamic adjustment of operating parameters, ensuring that the unit operates in the best state, thereby solving the problems of insufficient accuracy and dynamic adjustment capability of control strategy execution in the prior art.

3. The present invention uses a self-learning feedback module, reinforcement learning and data-driven modeling to continuously optimize the control strategy, and combines expert systems and digital twin technology to intelligently diagnose and process anomalies and faults, thereby improving the system's adaptability and optimization effect, and solving the problem in the prior art that the control strategy has limited optimization and adjustment capabilities and is difficult to adapt to complex and changeable operating conditions.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a framework diagram of the system of the present invention;

FIG2 is a framework diagram of an adaptive data acquisition module of the present invention;

FIG3 is a framework diagram of a real-time status monitoring module of the present invention;

FIG4 is a framework diagram of an intelligent optimization decision module of the present invention;

FIG5 is a framework diagram of an efficient execution control module of the present invention;

FIG6 is a framework diagram of the self-learning feedback module of the present invention.

DETAILED DESCRIPTION

The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

Referring to FIG. 1 , an embodiment of the present invention provides a self-adaptive adjustment unit sliding pressure optimization operation control system, comprising:

Adaptive data acquisition module, which is used to collect and pre-process unit operation data with high precision and multi-dimensionality;

A real-time status monitoring module, which is connected to the adaptive data acquisition module, is used to dynamically monitor and estimate the unit's operating status and detect anomalies and changes in real time;

Intelligent optimization decision module, which is connected with the efficient execution control module, is used to formulate and optimize the sliding pressure control strategy to make the unit operate in the optimal state;

An efficient execution control module, which is connected to the real-time status monitoring module and the self-learning feedback module to implement precise control strategies and dynamically adjust key operating parameters;

The self-learning feedback module is connected to the intelligent optimization decision-making module to continuously monitor and learn the operating conditions, optimize and adjust the control strategy.

Please refer to Figure 2, the adaptive data acquisition module includes:

Dynamic sampling unit, which is used to dynamically adjust the sampling frequency and range according to the unit operation status and load changes;

Multi-source data fusion unit, which is used to combine multiple sensor data to achieve comprehensive monitoring of key parameters;

Intelligent pre-processing unit for cleaning and processing raw data through adaptive filtering and noise suppression algorithms.

The dynamic sampling frequency adjustment formula in the dynamic sampling unit is:

,in, is the sampling frequency at time t, is the initial sampling frequency, is the sampling frequency adjustment factor, is the change in pressure at time t, is the base pressure value.

Specifically, the dynamic sampling unit improves the real-time and accuracy of data collection, and can timely reflect the unit operating status and load changes.

It dynamically adjusts the sampling frequency and range according to the real-time operating status and load changes of the unit to ensure the timeliness and accuracy of the collected data.

The unit adjusts the sampling frequency and range in real time by monitoring the unit's operating status and load changes. Specifically, when the unit's load changes significantly or the operating status changes significantly, the dynamic sampling unit will increase the sampling frequency, and vice versa, it will reduce the sampling frequency to save resources and ensure the validity of the data.

Dynamic sampling frequency adjustment formula: ,in, is the sampling frequency at time t, is the initial sampling frequency, is the sampling frequency adjustment factor, is the change in pressure at time t, is the base pressure value.

The multi-source data fusion unit enhances the integrity and reliability of the data, avoids the error of single sensor data, and improves the monitoring accuracy.

It combines data from multiple sensors to comprehensively monitor the key parameters of the unit and provide accurate operating status information.

The unit fuses data from different sensors to eliminate the error of a single data source. The data from each sensor is weighted and fused into a comprehensive data to improve the accuracy of monitoring.

Data fusion formula: X(t)=

in:

X(t) is the fused data at time t, N is the number of sensors, is the data weight of the ith sensor at time t, is the data of the i-th sensor at time t.

The intelligent pre-processing unit improves data quality, reduces the impact of noise and abnormal data, and ensures the accuracy of subsequent analysis and control.

It cleans and processes the collected raw data through adaptive filtering and noise suppression algorithms to remove noise and outliers.

The unit uses an adaptive filtering algorithm to automatically adjust filtering parameters and optimize data smoothing effects according to real-time data changes. At the same time, the noise suppression algorithm can identify and remove noise and outliers in the data, improving the validity and reliability of the data.

Adaptive filtering formula: (t)x(t)+(1-

in: is the filtered data at time tt, α(t) is the adaptive filter coefficient at time t, and x(t) is the original data at time t.

Through the coordinated work of the dynamic sampling unit, multi-source data fusion unit and intelligent pre-processing unit of the adaptive data acquisition module, high-precision, multi-dimensional unit operation data acquisition and pre-processing are achieved, improving the real-time, accuracy and reliability of the data. This provides a solid data foundation for subsequent real-time status monitoring, intelligent optimization decision-making and efficient execution control, ensuring that the system can maintain the optimal operating state under various working conditions.

Please refer to Figure 3, the real-time status monitoring module includes:

A state estimation unit, which is used to estimate the real-time operating state of the unit using Kalman filtering and particle filtering methods;

Anomaly detection unit, which is used to detect anomalies in real time using statistical methods and machine learning models;

A change identification unit is used to identify changes in load and operating conditions based on a change point detection algorithm.

Specifically, the state estimation unit provides high-precision real-time operation state estimation, enhances the monitoring capability of the system, and ensures the stable operation of the unit.

It uses Kalman filtering and particle filtering methods to estimate the operating status of the unit in real time and provide accurate status information for subsequent processing and decision-making.

The sensor data of the unit is processed through Kalman filtering and particle filtering methods to estimate the current operating state of the unit. Kalman filtering is suitable for linear systems, while particle filtering is used to process nonlinear and non-Gaussian systems. The combination of the two can provide more accurate state estimation.

Kalman filter formula:

Prediction Steps:

2. Update steps:

Particle filter formula:

Initialize particles:

Particle Prediction:

3. Particle weighting:

4. Particle resampling:

The abnormality detection unit promptly discovers and warns of abnormal conditions, preventing the expansion of faults and improving the safety and reliability of unit operation.

Utilize statistical methods and machine learning models to detect anomalies in operating data in real time, identify potential faults and anomalies, and issue early warnings in a timely manner.

The anomaly detection unit analyzes real-time data to detect anomalies through statistical methods (such as Z-score) and machine learning models (such as isolation forest, support vector machine, etc.). When the data deviates from the normal range or pattern, the system will issue an alarm.

The change recognition unit accurately identifies changes in load and operating conditions, and adjusts the control strategy in a timely manner to ensure efficient operation of the unit.

Based on the change point detection algorithm, the changes in unit load and operating conditions are identified, and the change information is provided for optimized decision-making and control adjustments.

The change identification unit analyzes the real-time data stream through the change point detection algorithm (such as CUSUM, Pettitt test, etc.), detects the change points of the data, and thus identifies the changes in load and working conditions.

Through the coordinated work of the state estimation unit, anomaly detection unit and change identification unit of the real-time state monitoring module, high-precision estimation of the unit's operating state, real-time anomaly detection and load change identification are achieved. These functions provide accurate state information and early warning capabilities for the system's intelligent optimization decision-making and efficient execution control, ensuring that the unit can maintain efficient and safe operation under various operating conditions.

Please refer to Figure 4, the intelligent optimization decision module includes:

An adaptive optimization unit, which is used to design an adaptive dynamic programming algorithm to optimize the sliding pressure control strategy in real time;

Multi-objective optimization unit, which is used to simultaneously consider multiple objectives such as heat rate, power generation efficiency and emissions and use evolutionary algorithms for optimization;

Online learning unit, which is used to continuously update the optimization model using real-time data in combination with deep reinforcement learning.

The adaptive dynamic programming optimization formula in the adaptive optimization unit combines the immediate cost and the future expected cost to optimize the control strategy. The formula is:

,in, is the optimal value function of the dynamic s, is the instantaneous cost of state s and dynamic u, is the discount factor, After taking action u in state s, the state is transferred to probability.

Specifically, the adaptive optimization unit provides the ability to optimize the sliding pressure control strategy in real time, which can maintain the optimal operating state of the unit under complex and changeable operating conditions.

An adaptive dynamic programming algorithm is designed to optimize the sliding pressure control strategy in real time, combining the immediate cost and future expected cost to achieve optimal control.

The adaptive optimization unit dynamically optimizes the control strategy based on the consideration of immediate cost and future expected cost through the adaptive dynamic programming algorithm. The algorithm can continuously adjust and optimize the control strategy according to real-time data, so that the system always maintains the best state during operation.

Adaptive dynamic programming optimization formula: ,in, is the optimal value function of the dynamic s, is the instantaneous cost of state s and dynamic u, is the discount factor, After taking action u in state s, the state is transferred to probability.

The multi-objective optimization unit realizes multi-objective comprehensive optimization, which can simultaneously consider heat rate, power generation efficiency and emission targets to ensure the optimal overall operating effect.

An evolutionary algorithm is used to comprehensively optimize multiple objectives to achieve a balance among heat rate, power generation efficiency and emission targets.

The multi-objective optimization unit uses an evolutionary algorithm (such as NSGA-II) to comprehensively optimize multiple optimization objectives. The algorithm can handle multiple conflicting objectives and find the optimal balance point to ensure the best overall operation effect of the unit.

Multi-objective optimization formula:

minF(x)=[

Where: F(x) is the objective function vector, is the i-th objective function, and m is the number of objective functions.

The online learning unit improves the system's adaptability and optimization effect, and can continuously learn and improve control strategies to adapt to changing operating conditions.

Combined with deep reinforcement learning, the optimization model is continuously updated using real-time data to enhance the system's adaptability and optimization effect.

The online learning unit continuously updates and optimizes the control model using real-time data through deep reinforcement learning (such as DQN, DDPG, etc.). The unit can continuously improve the control strategy and improve the system's operating performance based on new operating data.

Reinforcement learning Q-Learning formula:

Q(s,a) [r+

in:

Q(s,a) is the Q-value of state s and action a.

α is the learning rate.

r is the immediate reward.

γ is the discount factor.

is the next state.

Through the collaborative work of the adaptive optimization unit, multi-objective optimization unit and online learning unit of the intelligent optimization decision module, real-time optimization of sliding pressure control strategy, multi-objective comprehensive optimization and continuous learning to improve control strategy are realized. The adaptive optimization unit optimizes the control strategy in real time through adaptive dynamic programming algorithm; the multi-objective optimization unit achieves the balance of heat rate, power generation efficiency and emission targets through evolutionary algorithm; the online learning unit continuously updates the optimization model through deep reinforcement learning to improve the system's adaptive ability and optimization effect. These functions ensure that the unit always maintains the optimal operating state under complex and changeable operating conditions.

Please refer to FIG5 , the efficient execution control module includes:

An adaptive PID control unit, which is used to design a model-based adaptive PID controller to dynamically adjust control parameters;

A fuzzy logic control unit for optimizing the adjustment of the main steam pressure by combining fuzzy logic to handle nonlinearity and uncertainty;

A coordinated control unit for achieving coordinated optimization of boilers, turbines and generators through a multivariable control strategy;

A blockchain verification unit is used for distributed storage and verification of control instructions through blockchain technology.

The adaptive PID control parameter adjustment formula in the adaptive PID control unit is:

)

= (1+

),in, , , are the proportional, integral and derivative gains at time t, , , is the initial gain, , , is the gain adjustment coefficient, and e(t) is the error value at time t.

The adaptive PID control unit involves a dynamic sliding pressure set value adjustment formula, which mainly dynamically adjusts the sliding pressure set value according to the load and operating status to ensure that the unit operates in the best working condition. The formula is as follows:

,in, is the sliding pressure setting value at time t, is the reference sliding pressure value, is the sliding pressure adjustment coefficient, L(t) is the load at time t, is the base load.

Specifically, the adaptive PID control unit improves the control accuracy and response speed, can dynamically adapt to changes in operating conditions, and achieves precise control of the main steam pressure and sliding pressure set values.

Design a model-based adaptive PID controller to dynamically adjust control parameters to ensure that the system operates in the best state under different working conditions.

principle:

The adaptive PID control unit monitors the error in real time and dynamically adjusts the PID control parameters according to the current system status to achieve precise regulation of the control system. The parameter adjustment formula of the PID controller is as follows:

)

= (1+

),in, , , are the proportional, integral and derivative gains at time t, , , is the initial gain, , , is the gain adjustment coefficient, and e(t) is the error value at time t.

Dynamic sliding pressure setting value adjustment formula:

,in, is the sliding pressure setting value at time t, is the reference sliding pressure value, is the sliding pressure adjustment coefficient, L(t) is the load at time t, is the base load.

The fuzzy logic control unit can handle nonlinear and uncertain problems, improving the stability and control effect of the system.

Fuzzy logic is combined to handle nonlinearity and uncertainty, optimize the adjustment of main steam pressure and ensure smooth system operation.

principle:

The fuzzy logic control unit uses fuzzy sets and fuzzy rules to perform fuzzy reasoning and decision-making according to the current system status and operating parameters to optimize the control of the main steam pressure.

Fuzzy control rule formula: u=

in:

u is the control output.

is the weight of the ith rule.

is the membership of input x to the i-th rule.

The coordinated control unit realizes the coordinated optimization of boilers, turbines and generators, improving the overall operating efficiency and stability.

Through multivariable control strategies, coordinated control and optimization of boilers, turbines and generators are achieved to ensure that all parts work together.

principle:

The coordinated control unit uses a multivariable control strategy, comprehensively considers the operating status and needs of each part, and realizes coordinated control through an optimization algorithm.

Coordinated control optimization formula:

minJ=

in:

J is the objective function.

is the weight coefficient.

is the operating state function of the ith subsystem.

The blockchain verification unit improves the security and tamper-proofness of control instructions, ensuring the reliability of system operation and the integrity of data.

Blockchain technology is used to distribute and verify control instructions to prevent data tampering and unauthorized modification.

principle:

The blockchain verification unit stores control instructions and operating data in the form of blocks in a distributed ledger, and ensures the security and immutability of the data through consensus mechanisms and encryption technologies.

Blockchain consensus formula: H(

in:

H( is the hash value of block i.

It is the previous block.

It is the data of the current block.

Through the coordinated work of the adaptive PID control unit, fuzzy logic control unit, coordination control unit and blockchain verification unit of the efficient execution control module, precise control and dynamic adjustment of the unit operating parameters are achieved. The adaptive PID control unit ensures that the system operates in the best state by dynamically adjusting the PID parameters and sliding pressure set values; the fuzzy logic control unit handles nonlinearity and uncertainty and optimizes the control of the main steam pressure; the coordination control unit realizes the coordinated optimization of the boiler, turbine and generator; the blockchain verification unit ensures the security and non-tamperability of the control instructions. These functions together ensure the efficient and stable operation of the unit under various operating conditions.

Please refer to Figure 6, the self-learning feedback module includes:

A reinforcement learning unit, which is used to continuously optimize the control strategy through analysis and learning of operating data;

A data-driven modeling unit, which is used to build a data-driven unit operation model using big data analysis technology and machine learning algorithms;

Intelligent diagnostic unit, which is used to combine expert system and digital twin technology to intelligently diagnose and process abnormalities and faults.

Specifically, the reinforcement learning unit continuously optimizes the control strategy, improves the system's adaptability and operating efficiency, and ensures that the system can maintain the best state under different working conditions.

By analyzing and learning the operating data, the control strategy is continuously optimized using reinforcement learning algorithms to improve the system's responsiveness and optimization effects.

principle:

The reinforcement learning unit uses reinforcement learning algorithms (such as Q-learning, deep Q network DQN, etc.) to learn the best control strategy from historical and real-time data. Through the reward and punishment mechanism, the algorithm continuously adjusts the strategy to maximize long-term returns.

The data-driven modeling unit constructs a high-precision unit operation model, improves the accuracy of prediction and optimization, and enhances the system's intelligent and refined management capabilities.

Utilize big data analysis technology and machine learning algorithms to build and update data-driven unit operation models, providing a reliable data foundation for optimized decision-making.

principle:

The data-driven modeling unit collects and analyzes a large amount of historical operating data and builds unit operation models using machine learning algorithms (such as regression analysis, neural networks, etc.). These models can capture the complex dynamic characteristics of the system and provide support for real-time optimization and control.

The intelligent diagnosis unit improves the accuracy and timeliness of fault diagnosis, reduces fault handling time and maintenance costs, and enhances system reliability and safety.

Combining expert systems and digital twin technology, intelligent diagnosis and processing of anomalies and faults can be performed to provide accurate fault location and diagnosis results.

principle:

The intelligent diagnosis unit uses the rules and knowledge base of the expert system and combines digital twin technology to establish a virtual machine group model. By comparing the actual operation data with the virtual model, it can promptly detect anomalies and faults, and perform intelligent diagnosis and processing.

Through the collaborative work of the reinforcement learning unit, data-driven modeling unit and intelligent diagnosis unit of the self-learning feedback module, continuous optimization, accurate modeling and intelligent diagnosis of the system control strategy are achieved. The reinforcement learning unit continuously optimizes the control strategy and improves the system's adaptive ability by analyzing the operating data; the data-driven modeling unit uses big data analysis and machine learning to build a high-precision operating model to provide a reliable prediction and optimization basis; the intelligent diagnosis unit combines expert systems and digital twin technology to intelligently diagnose and process abnormalities and faults, improving the reliability and safety of the system. These functions together ensure the efficient and stable operation of the unit under various operating conditions.

Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims (10)

1. An adaptively adjusted unit sliding pressure optimizing operation control system, comprising:

The self-adaptive data acquisition module is used for acquiring and preprocessing the operation data of the set in a high-precision and multi-dimensional manner;

The real-time state monitoring module is connected with the self-adaptive data acquisition module and is used for dynamically monitoring and estimating the running state of the unit and detecting the abnormality and the change in real time;

the intelligent optimization decision module is connected with the high-efficiency execution control module and is used for making and optimizing a sliding pressure control strategy so that the unit operates in an optimal state;

The high-efficiency execution control module is connected with the real-time state monitoring module and the self-learning feedback module and is used for implementing an accurate control strategy and dynamically adjusting key operation parameters;

And the self-learning feedback module is connected with the intelligent optimization decision module and is used for continuously monitoring and learning the running condition and optimizing and adjusting the control strategy.

2. The adaptively adjusted unit slide optimization run control system of claim 1, wherein the adaptive data acquisition module comprises:

the dynamic sampling unit is used for dynamically adjusting the sampling frequency and the sampling range according to the running state and the load change of the unit;

The multi-source data fusion unit is used for combining various sensor data to realize comprehensive monitoring of key parameters;

And the intelligent preprocessing unit is used for cleaning and processing the original data through an adaptive filtering and noise suppression algorithm.

3. The adaptively adjusted unit skid optimal operation control system as set forth in claim 1, wherein said real-time status monitoring module comprises:

A state estimation unit for estimating a real-time operation state of the machine set using a kalman filtering and particle filtering method;

an anomaly detection unit for detecting anomalies in real time using a statistical method and a machine learning model;

And the change identification unit is used for identifying the change of the load and the working condition based on the change point detection algorithm.

4. The adaptively adjusted unit slide optimization operation control system according to claim 1, wherein the intelligent optimization decision module comprises:

The self-adaptive optimization unit is used for designing a self-adaptive dynamic programming algorithm to optimize a sliding pressure control strategy in real time;

A multi-objective optimization unit for simultaneously considering a heat rate, a power generation efficiency, and a plurality of objectives of emission and optimizing using an evolutionary algorithm;

And the online learning unit is used for continuously updating the optimization model by utilizing real-time data in combination with the deep reinforcement learning.

5. The adaptively adjusted unit slide optimization run control system of claim 1, wherein said high efficiency execution control module comprises:

the self-adaptive PID control unit is used for designing a model-based self-adaptive PID controller to dynamically adjust control parameters;

a fuzzy logic control unit for optimizing the adjustment of the main vapor pressure in combination with the fuzzy logic processing nonlinearity and uncertainty;

a coordination control unit for implementing cooperative optimization of the boiler, the steam turbine and the generator through a multivariable control strategy;

And the block chain verification unit is used for carrying out distributed storage and verification of the control instructions through a block chain technology.

6. The adaptively adjusted unit slide optimization operation control system according to claim 1, wherein the self-learning feedback module comprises:

A reinforcement learning unit for constantly optimizing a control strategy through analysis and learning of the operation data;

A data-driven modeling unit for constructing a data-driven unit operation model using a big data analysis technique and a machine learning algorithm;

and the intelligent diagnosis unit is used for intelligently diagnosing and processing the abnormality and the fault by combining an expert system and a digital twin technology.

7. The adaptively adjusted unit sliding pressure optimizing operation control system according to claim 2, wherein the dynamic sampling frequency adjustment formula in the dynamic sampling unit is:

Wherein, the method comprises the steps of, wherein, Is the sampling frequency of the time t,Is the initial sampling frequency at which the sample is to be taken,Is the sampling frequency adjustment coefficient and,Is the pressure change over the time t and,Is the reference pressure value.

8. The adaptively adjusted unit sliding pressure optimizing operation control system according to claim 4, wherein the adaptive dynamic programming optimizing formula in the adaptive optimizing unit combines the instant cost and the future expected cost to optimize the control strategy, and the formula is:

Wherein, the method comprises the steps of, wherein, Is a function of the optimal value of the dynamic s,Is the instantaneous cost of state s and dynamic u,Is a discount factor that is used to determine the discount,Is to transition to state after action u is taken at state sIs a probability of (2).

9. The adaptively adjusted unit sliding pressure optimizing operation control system according to claim 5, wherein the adaptive PID control parameter adjustment formula in the adaptive PID control unit is:

)

=(1+

) Wherein, the method comprises the steps of, wherein, 、、Proportional, integral and differential gains of time t respectively,、、Is the initial gain of the gain control unit,、、Is the gain adjustment coefficient and e (t) is the error value for time t.

10. The adaptively adjusted unit sliding pressure optimizing operation control system according to claim 5, wherein the adaptive PID control unit relates to a dynamic sliding pressure set value adjusting formula, which mainly dynamically adjusts the sliding pressure set value according to the load and the operation state, so as to ensure that the unit operates under the optimal working condition, and the formula is as follows:

Wherein, the method comprises the steps of, wherein, Is the slide pressure set point at time t,Is the reference sliding pressure value, the sliding pressure value is the reference sliding pressure value,Is the slip pressure adjustment coefficient, L (t) is the load at time t,Is the reference load.