CN-122022158-A - Dry-wet state conversion self-adaptive adjusting system for deep peak regulation of thermal power unit

CN122022158ACN 122022158 ACN122022158 ACN 122022158ACN-122022158-A

Abstract

The present invention relates to the field of automatic control and deep peak shaving operation technology for thermal power units, and discloses a wet dry state conversion adaptive regulation system for deep peak shaving of thermal power units, including a classification data acquisition module, a data preprocessing module, a hybrid dynamic modeling module, a deep learning adaptive decision-making module, a multi model prediction optimization module, a fully automatic execution module, an integrated linkage module, a safety monitoring module, and an operation and maintenance management module. By designing a collaborative working mechanism of fully automatic execution, integrated linkage, real-time safety monitoring, and intelligent operation and maintenance management, a full process closed-loop intelligent control system is constructed from instruction generation to precise execution, from the main system to the auxiliary system, from process control to risk control and efficiency evaluation, ultimately achieving a significant improvement in system automation level, operational safety, and comprehensive economy, The beneficial effect of reducing the operational burden and safety risks for operators.

Inventors

ZHANG JINGJING
LV HENGBIN
ZHAO HUI
YUE HONGLIANG

Assignees

华润电力（常熟）有限公司

Dates

Publication Date: 20260512
Application Date: 20260128

Claims (10)

1. The dry-wet state conversion self-adaptive regulation system for deep peak shaving of the thermal power generating unit is characterized by comprising a classified data acquisition module, a data preprocessing module, a hybrid dynamic modeling module, a deep learning self-adaptive decision module, a multi-model prediction optimization module, a full-automatic execution module, an integrated linkage module, a safety monitoring module and an operation and maintenance management module; The classified data acquisition module is used for carrying out full-dimensional original data acquisition by four classified stations, and acquired data are transmitted to the data preprocessing module in real time; The data preprocessing module is used for preprocessing all-dimensional original data acquired by four stations, generating a high-quality data set and transmitting the high-quality data set to the hybrid dynamic modeling module and the deep learning self-adaptive decision module; The mixed dynamic modeling module builds a dry-wet state conversion dynamic model, and outputs accurate modeling guidance through a dry-wet state characteristic fusion modeling formula; The deep learning self-adaptive decision module is used for realizing self-adaptive decision of dry-wet state switching time and adjusting parameters through a deep learning algorithm based on the preprocessed high-quality data set and the output result of the mixed dynamic model; The multi-model prediction optimization module predicts key parameters in the future 5-10s based on the mixed dynamic model output and the deep learning self-adaptive decision result by introducing a multi-model prediction control algorithm, so as to realize secondary optimization on the adjustment parameter instruction output by the deep learning self-adaptive decision module, generate an optimal control sequence and transmit the optimal control sequence to the full-automatic execution module; The full-automatic execution module receives the optimal control sequence output by the multi-model prediction optimization module, and based on full-automatic sequence control logic, drives a corresponding execution mechanism to finish accurate actions, and realizes full-automatic execution of the dry-wet state conversion process; the integrated linkage module is based on a DCS deep integration architecture, integrates the auxiliary control system and the main control system comprehensively, builds unified linkage control logic, receives action instructions of the full-automatic execution module, and synchronously coordinates the running states of the auxiliary control systems; the safety monitoring module collects the running parameters of the unit and the running state data of each module in real time, monitors the risks of overtemperature, overpressure and hydrodynamic instability in the transition process in real time by setting a safety threshold and a trend early warning curve, and outputs early warning information; The operation and maintenance management module receives the early warning information of the safety monitoring module and the operation data of each module, constructs a unit operation state evaluation model, carries out real-time accounting and trend analysis on equipment loss and operation efficiency, and realizes the instant maintenance of the equipment.
2. The self-adaptive dry-wet state conversion regulating system for deep peak shaving of thermal power generating unit according to claim 1, wherein the classification type data acquisition module comprises four stations, namely a boiler side data acquisition unit, a steam turbine side data acquisition unit, an auxiliary machine side data acquisition unit and an electric side data acquisition unit.
3. The self-adaptive dry-wet state conversion regulating system for deep peak shaving of thermal power generating unit according to claim 2, wherein the boiler side data acquisition unit acquires key parameters of boiler combustion and steam-water system through a distributed temperature/pressure transmitter and a flowmeter.
4. The self-adaptive dry-wet state conversion regulating system for deep peak shaving of thermal power generating unit according to claim 2, wherein the turbine side data acquisition unit acquires the operation parameters of the turbine and the regulating system thereof through a vibration sensor, a displacement sensor and a thermal instrument.
5. The self-adaptive dry-wet state conversion regulating system for deep peak shaving of thermal power generating unit according to claim 2, wherein the auxiliary machine side data acquisition unit acquires main auxiliary machine states and process parameters through the pump/fan intelligent monitoring device and the valve positioner.
6. The self-adaptive dry-wet state conversion regulating system for deep peak shaving of thermal power generating unit according to claim 2, wherein the electric side data acquisition unit acquires parameters of a generator and a station service electrical system through an electric measurement and protection device.
7. The self-adaptive dry-wet state conversion regulating system for deep peak shaving of thermal power generating unit according to claim 1, wherein the hybrid dynamic modeling module realizes accurate modeling guidance under wide load and multiple working conditions through a dry-wet state characteristic fusion modeling formula, and the calculation formula is as follows: , In the formula (i), Represents the comprehensive dynamic characteristic function in the dry-wet state conversion process, The vector of the input parameters is represented as, The time is represented by the time period of the day, Represents the weight coefficient of the dry-wet state, Representing the dynamic characteristic function of the dry state operation, Representing the dynamic characteristic function of the wet state operation, Representing a dynamic correction term of the transition process; , In the formula (i), Representing the rate of change of the system state vector, A system state vector is represented and is used to represent, The control input vector is represented as such, Represents a mechanism model based on physical laws such as mass conservation and energy conservation, Representing an ith fuzzy rule or data driven sub-model, Representing its adaptive weights based on reinforcement learning real-time evaluation, The total number of fuzzy rules or submodels is represented, and the value range is 4-10.
8. The self-adaptive dry-wet state conversion regulating system for deep peak shaving of a thermal power generating unit according to claim 1, wherein the multi-model prediction optimizing module adopts a state prediction model to construct a multi-objective prediction optimizing model of a conversion process, and is specifically expressed as: , In the formula (i), A predicted output vector representing a key parameter at a future time (k + 1), A system state vector representing the current time (k), Representing the current sequence of control inputs to be optimized, And Respectively representing a state matrix and a control matrix obtained by linearizing the hybrid dynamic model.
9. The adaptive adjustment system for converting a dry state to a wet state for deep peak shaving of a thermal power generating unit according to claim 1, wherein the multi-model predictive optimization module constructs a multi-objective predictive optimization model of a conversion process by adopting a multi-objective optimization function, and the model is expressed as: In the formula (i), Representing the objective function value that needs to be minimized, Meaning minimizing the cost summation of all prediction steps in the prediction time domain, Actual predictor vectors representing key controlled parameters output by the hybrid dynamic model, Representing the dynamic set value or safety range reference track of the key controlled parameter in the transition process, Representing the rate of change or delta of the control input vector, A penalty term or time cost term representing the completion of the transition process, 、、 Respectively represent the weighting coefficients for balancing the parameter tracking precision, the control motion stability and the transition speed.
10. The dry-wet state conversion self-adaptive regulating system for deep peak regulation of a thermal power generating unit according to claim 9, wherein the multi-model prediction optimizing module predicts key parameters of main steam temperature, main steam pressure and working medium liquid level in the future 5-10s based on a state prediction model and a multi-objective optimizing function calculation result, secondarily optimizes regulating parameter instructions output by the deep learning self-adaptive decision module, generates an optimal control sequence and transmits the optimal control sequence to the full-automatic execution module; The full-automatic execution module receives the optimal control sequence output by the multi-model prediction optimization module, drives the corresponding water regulating valve, steam valve and circulating water pump to finish accurate actions based on full-automatic sequence control logic, realizes full-automatic execution of the dry-wet state conversion process, replaces manual operation by standardized execution logic, and simultaneously acquires action feedback data of an execution mechanism in real time to form closed-loop control.

Description

Dry-wet state conversion self-adaptive adjusting system for deep peak regulation of thermal power unit Technical Field The invention relates to the technical field of automatic control and deep peak shaving operation of thermal power units, in particular to a dry-wet state conversion self-adaptive regulating system for deep peak shaving of a thermal power unit. Background With the continuous improvement of the power generation duty ratio of new energy, the power grid has increasingly severe requirements on the deep peak regulation capacity of the thermal power generating unit. The once-through boiler and the supercritical unit with the deep peak shaving capability need to frequently switch the dry state operation mode and the wet state operation mode under low load so as to ensure the hydrodynamic safety and the boiler efficiency. However, the existing dry-wet state conversion process mainly relies on experience of operators to perform manual or semi-automatic operation, and has the outstanding problems that firstly, the conversion process is slow in response and cannot track a power grid peak regulation instruction quickly, secondly, an accurate model capable of accurately describing wide load, dry/wet states under multiple conditions and dynamic characteristics of the conversion process of the dry/wet states is lacking, the control strategy is caused to depend on experience parameters, the self-adaptive capacity is poor, thirdly, the main control system is not fully linked with auxiliary control systems such as circulating water, soot blowing and the like, the 'information island' exists, the cooperative efficiency is low, fourthly, the automation degree is low, the operation steps are complex, the safety risk is high, and key parameters (such as main steam temperature, main steam pressure and working medium liquid level) in the conversion process fluctuate greatly, so that the service life and economy of a unit are influenced. Therefore, there is a need for a dry-wet state transition intelligent regulation system that is fully automatic, adaptive, highly accurate, and safe and economical. Disclosure of Invention (One) solving the technical problems Aiming at the defects of the prior art, the invention provides a dry-wet state conversion self-adaptive regulating system for deep peak shaving of a thermal power generating unit, which has the advantages of quick response, accurate modeling, intelligent decision, full-automatic execution, main-auxiliary depth linkage and overall process safety monitoring, and solves the problems of slow response, low modeling precision, poor linkage, low automation degree and poor safety and economy in the prior art. (II) technical scheme In order to achieve the aim, the invention provides the technical scheme that the dry-wet state conversion self-adaptive regulating system for the deep peak shaving of the thermal power generating unit comprises a classified data acquisition module, a data preprocessing module, a hybrid dynamic modeling module, a deep learning self-adaptive decision module, a multi-model prediction optimizing module, a full-automatic execution module, an integrated linkage module, a safety monitoring module and an operation and maintenance management module; The classified data acquisition module is used for carrying out full-dimensional original data acquisition by four classified stations, and acquired data are transmitted to the data preprocessing module in real time; The data preprocessing module is used for preprocessing all-dimensional original data acquired by four stations, generating a high-quality data set and transmitting the high-quality data set to the hybrid dynamic modeling module and the deep learning self-adaptive decision module; The mixed dynamic modeling module builds a dry-wet state conversion dynamic model, and outputs accurate modeling guidance through a dry-wet state characteristic fusion modeling formula; The deep learning self-adaptive decision module is used for realizing self-adaptive decision of dry-wet state switching time and adjusting parameters through a deep learning algorithm based on the preprocessed high-quality data set and the output result of the mixed dynamic model; The multi-model prediction optimization module predicts key parameters in the future 5-10s based on the mixed dynamic model output and the deep learning self-adaptive decision result by introducing a multi-model prediction control algorithm, so as to realize secondary optimization on the adjustment parameter instruction output by the deep learning self-adaptive decision module, generate an optimal control sequence and transmit the optimal control sequence to the full-automatic execution module; The full-automatic execution module receives the optimal control sequence output by the multi-model prediction optimization module, and based on full-automatic sequence control logic, drives a corresponding execution mechanism to finish accurate actions, and realize