CN-122014365-A - Steam turbine lifting rate and steam temperature cooperative control method based on allowable temperature difference feedback

CN122014365ACN 122014365 ACN122014365 ACN 122014365ACN-122014365-A

Abstract

The invention discloses a steam turbine rising rate and steam temperature cooperative control method based on allowable temperature difference feedback, which belongs to the field of steam turbine operation control and comprises the steps of collecting deep peak regulation operation data, cleaning and normalizing, generating thermal state prediction data by using an unsteady heat conduction analysis model, establishing a cyclic fatigue and creep damage superposition model to calculate an accumulated damage value, dynamically correcting an allowable temperature difference limit envelope curve, distributing control vectors between a steam superheat degree loop and a rising rate loop according to stress sensitivity weight to generate a cooperative decision instruction, and performing feedforward compensation by using an extended state observer and issuing a final control signal. The invention adopts the digital twin prediction and multivariable decoupling cooperative technology, can compensate the hysteresis of the executing mechanism in real time and actively regulate and control the thermal stress track, and obviously improves the running safety and the full life cycle economy of the unit under the frequent variable working condition.

Inventors

LIU ZHICHAO
ZHANG CHUNLEI
ZHANG HONGZHI
YIN JIANGUO
LI CHANGZHENG
YANG JINLONG

Assignees

华润电力(沈阳)有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. The method for cooperatively controlling the rise rate and the steam temperature of the steam turbine based on allowable temperature difference feedback is characterized by comprising the following steps of: Collecting operation data of the steam turbine in a deep peak regulation process in real time, and performing cleaning and normalization processing on the operation data through a thermodynamic state equation to form a basic data set for supporting thermal kinetic energy gradient calculation, wherein the operation data comprises main steam pressure, steam flow, a regulating-stage enthalpy value and multi-point measuring point temperatures distributed on the surface of a rotor of the steam turbine; inputting the basic data set into a preset unsteady heat conduction analysis model, and analyzing an equivalent strain field of an internal non-measurable point of the rotor in real time by using a material heat conductivity coefficient, a specific heat capacity and a surface heat exchange coefficient determined by steam flow and pressure of the rotor to generate thermal state prediction data representing a real-time evolution track of the thermal stress of the rotor; The method comprises the steps of calling the start-stop frequency and the load fluctuation amplitude in a rotor historical operation file, establishing a mathematical model of superposition of cycle fatigue damage caused by alternating thermal stress and creep damage accumulated with time under a high-temperature working condition, and calculating to obtain a current accumulated fatigue damage value of the rotor; Based on the accumulated fatigue damage value, nonlinear shrinkage or expansion adjustment is carried out on the standard allowable temperature difference by utilizing a fracture mechanics correction function, and a dynamic allowable temperature difference limit value envelope curve at the current moment is established; the thermal state prediction data is compared with the dynamic allowable temperature difference limit value envelope curve in real time, a stress safety margin is determined, and a control vector is distributed between a steam superheat degree regulating loop and a rising rate regulating loop according to stress sensitivity weight to generate a collaborative decision instruction; According to the vector distribution result in the collaborative decision-making instruction, calculating to obtain a corresponding steam temperature regulation target value and a corresponding rise rate regulation target value, and respectively generating a corresponding PID control instruction; and capturing mechanical hysteresis and system total disturbance of the executing mechanism by using the extended state observer, performing feedforward compensation correction on the PID control instruction, generating a final control signal and transmitting the final control signal to the turbine speed regulating system and the steam temperature reducing water executing mechanism.
2. The cooperative control method of the turbine lifting rate and the steam temperature based on allowable temperature difference feedback according to claim 1, wherein the collecting operation data of the turbine in the deep peak shaving process in real time, and performing cleaning and normalization processing on the operation data through a thermodynamic equation of state comprises: unifying time references of the acquired pressure, flow, enthalpy and temperature of each measuring point, eliminating phase differences of different physical quantity sensors due to inconsistent sampling frequencies by using an interpolation algorithm, and ensuring synchronism of multidimensional operation data on the same time section; Establishing a state association rule among all operation parameters by utilizing the thermodynamic state equation, identifying and eliminating instantaneous mutation data deviating from thermodynamic logic, and carrying out reconstruction filling on missing or abnormal positions based on historical trend or adjacent effective data; Analyzing to obtain nonlinear characteristic parameters reflecting the heat exchange intensity of the rotor surface by combining the time change rate, the real-time pressure and the real-time flow of the temperature of the rotor surface measuring point; and mapping the cleaned operation parameters to a unified per unit interval, eliminating the influence of different physical dimensions on the calculation of the unsteady heat conduction analysis model, and outputting a standardized basic data set.
3. The method according to claim 1, wherein inputting the basic data set into a preset unsteady heat conduction analysis model, and analyzing the equivalent strain field of the non-measurable point inside the rotor in real time by using the material heat conductivity coefficient, specific heat capacity and surface heat exchange coefficient determined by the steam flow rate and pressure of the rotor comprises: Carrying out numerical simulation on a transient temperature field of the rotor under different variable load working conditions by using a three-dimensional finite element model, and obtaining a historical response sample of a rotor internal stress dangerous point; Taking the basic data set as an input characteristic, taking equivalent stress at a corresponding moment as an output label, and pre-training a weight coefficient in the unsteady heat conduction analysis model by optimizing a mean square error loss function; Inputting the currently acquired basic data set as an excitation signal into a pre-trained unsteady heat conduction analysis model in real time, and performing convolution operation mapping based on physical response characteristics on heat flow impact of the rotor surface by utilizing the weight coefficient; and outputting equivalent thermal stress distribution and strain intensity of a preset dangerous area in the rotor in real time by the unsteady thermal conduction analysis model to form thermal state prediction data.
4. The method for collaborative control of turbine lift rate and steam temperature based on allowable temperature differential feedback according to claim 1, wherein the establishing a mathematical model of cyclic fatigue damage caused by alternating thermal stress superimposed with creep damage accumulated over time under high temperature conditions, calculating a current accumulated fatigue damage value of the rotor comprises: Matching a preset material life loss curve according to the start-stop frequency and the corresponding rotating speed change characteristics, and calculating to obtain a cyclic damage component generated by the alternating action of mechanical stress and thermal stress; according to the steady-state operation temperature of the metal corresponding to the load fluctuation amplitude, and in combination with the accumulated time of high-temperature operation in the historical operation file, analyzing to obtain a creep damage component reflecting the degradation of the microstructure in the material; And carrying out weighted superposition on the cyclic damage component and the creep damage component by adopting a linear loss accumulation criterion, and outputting a current accumulated fatigue damage value representing the current service state and the material degradation degree of the rotor.
5. The method of claim 1, wherein establishing a dynamic allowable temperature difference limit envelope at a current time based on the cumulative fatigue damage value using a fracture mechanics correction function to perform nonlinear contraction or expansion adjustment on a standard allowable temperature difference comprises: Establishing a mapping relation of the strength of the rotor material which is degenerated along with the increase of the current accumulated fatigue damage value, and calculating to obtain a resistance attenuation coefficient reflecting the current real-time bearing capacity of the rotor material; Acquiring a standard allowable temperature difference curve of a rotor in an initial service state, and taking the standard allowable temperature difference curve as an initial reference boundary for nonlinear adjustment; And scaling the initial reference boundary by using the resistance attenuation coefficient, and carrying out real-time compensation on the scaled boundary by combining the change rate of temperature in the operation data along with time to generate a dynamic allowable temperature difference limit envelope curve which dynamically changes along with the service life of the rotor and the real-time working condition.
6. The method of claim 1, wherein the comparing the thermal state prediction data with the dynamic allowable temperature difference limit envelope in real time, determining a stress safety margin, and distributing control vectors between the steam superheat adjustment loop and the rise rate adjustment loop according to a stress sensitivity weight, and generating a collaborative decision command comprises: Calculating the deviation between a thermal stress peak value in the thermal state prediction data at the current moment and a boundary value corresponding to the dynamic allowable temperature difference limit value envelope curve to obtain stress safety margin data reflecting a real-time thermal safety space of the rotor; establishing a sensitivity matrix of the influence of the steam superheat degree and the rising rate on the surface heat exchange of the rotor and the stress of the central hole respectively, and analyzing the real-time contribution weight of each adjusting variable to the stress change; Presetting a cooperative regulation priority, when the stress safety margin is larger than a preset safety threshold, performing micro regulation and control by the steam superheat degree regulating loop preferentially, and when the thermal state prediction data reach the boundary of a dynamic allowable temperature difference limit value envelope curve, increasing the intervention proportion of the rising rate regulating loop according to the contribution weight, so as to realize dynamic allocation of control vectors and output a cooperative decision instruction.
7. The method for controlling the rise rate and the steam temperature of the steam turbine based on allowable temperature difference feedback according to claim 1, wherein the capturing mechanical hysteresis and total system disturbance of the actuating mechanism by using the extended state observer, performing feedforward compensation correction on the PID control command, generating a final control signal, and transmitting the final control signal to the speed regulating system of the steam turbine and the steam temperature reducing water actuating mechanism comprises: The actual feedback opening degrees of the steam turbine speed regulating system and the steam temperature reducing water executing mechanism are obtained in real time by utilizing the extended state observer, and are compared with the PID control instruction, and an executing mechanism internal disturbance component consisting of mechanical friction and response delay is obtained through analysis; defining external heat load impact generated by main steam parameter fluctuation as an external disturbance component, and constructing a system total disturbance model by combining the internal disturbance component; Generating a reverse compensation vector according to the total disturbance model of the system, and performing feedforward superposition on output pulses corresponding to the steam temperature regulation target value and the rise rate regulation target value so as to offset the influence of mechanical hysteresis on the control precision of the thermal stress; And synchronously transmitting the corrected final control signals to a speed regulation air valve and a temperature-reducing water regulating valve in an analog quantity or digital bus mode, so that real-time closed-loop control of the thermal stress track is realized.
8. The method for collaborative control of turbine rise rate and steam temperature based on allowable temperature differential feedback according to claim 1 further comprising: acquiring the actual evolution data of the thermal stress of the steam turbine after executing the final control signal in real time, comparing the actual evolution data with the thermal state prediction data, and calculating to obtain a prediction residual error reflecting the model prediction precision; based on the prediction residual error, the weight coefficient in the unsteady heat conduction analysis model is adjusted on line by using an incremental learning algorithm, so that the prediction deviation generated by equipment performance attenuation is reduced; and dynamically correcting the stress sensitivity weight according to the evolution trend of the prediction residual error, and realizing the self-adaptive optimization of the distribution ratio between the steam superheat degree regulating loop and the rising rate regulating loop.
9. The method for collaborative control of turbine rise rate and steam temperature based on allowable temperature differential feedback according to claim 8 wherein the calculating a prediction residual reflecting model prediction accuracy comprises: synchronously acquiring a predicted temperature value of the position of a measuring point of the surface of the corresponding rotor, which is output by the unsteady heat conduction analysis model in the process of calculating internal heat stress; Comparing the predicted temperature value with a real temperature value of a rotor surface measuring point acquired by a sensor to obtain a predicted residual error of a characterization model for response deviation of a thermal boundary condition; When the fluctuation amplitude of the prediction residual or the sliding average value thereof exceeds a preset precision threshold value, calculating the update increment of the weight coefficient according to the direction of the prediction residual by utilizing an online gradient descent algorithm, and carrying out iterative correction on the unsteady heat conduction analysis model.
10. A turbine lift rate and steam temperature cooperative control system based on allowable temperature difference feedback, wherein the system is used for the turbine lift rate and steam temperature cooperative control method based on allowable temperature difference feedback according to any one of claims 1 to 9, the system comprising: The system comprises an operation data acquisition and preprocessing module, a control module and a control module, wherein the operation data acquisition and preprocessing module acquires operation data of the steam turbine in a deep peak regulation process in real time, and the operation data are subjected to cleaning and normalization processing through a thermodynamic equation of state to form a basic data set for supporting thermal kinetic energy gradient calculation, wherein the operation data comprise main steam pressure, steam flow, regulating level enthalpy and multipoint measuring point temperatures distributed on the surface of a rotor of the steam turbine; The thermal state prediction module is used for inputting the basic data set into a preset unsteady thermal conduction analysis model, and analyzing an equivalent strain field of an unreliability point in the rotor in real time by utilizing the material heat conductivity coefficient, specific heat capacity and a surface heat exchange coefficient determined by steam flow and pressure of the rotor to generate thermal state prediction data representing a real-time evolution track of the thermal stress of the rotor; The accumulated fatigue damage calculation module is used for calling the start-stop frequency and the load fluctuation amplitude in the rotor historical operation file, establishing a mathematical model of superposition of cyclic fatigue damage caused by alternating thermal stress and creep damage accumulated with time under a high-temperature working condition, and calculating to obtain the current accumulated fatigue damage value of the rotor; The dynamic allowable temperature difference limit value envelope curve establishing module is used for carrying out nonlinear shrinkage or expansion adjustment on the standard allowable temperature difference by utilizing a fracture mechanics correction function based on the accumulated fatigue damage value, and establishing a dynamic allowable temperature difference limit value envelope curve at the current moment; The collaborative decision instruction generation module is used for comparing the thermal state prediction data with the dynamic allowable temperature difference limit value envelope curve in real time, determining a stress safety margin, distributing a control vector between a steam superheat degree regulating loop and a rising rate regulating loop according to stress sensitivity weight, and generating a collaborative decision instruction; The control instruction generation module is used for calculating a corresponding steam temperature regulation target value and a corresponding rise rate regulation target value according to the vector distribution result in the collaborative decision instruction, and respectively generating corresponding PID control instructions; And executing a dynamic compensation and instruction issuing module, capturing mechanical hysteresis and system total disturbance of an executing mechanism by using an extended state observer, performing feedforward compensation correction on the PID control instruction, generating a final control signal and issuing the final control signal to a turbine speed regulating system and a steam cooling water executing mechanism.

Description

Steam turbine lifting rate and steam temperature cooperative control method based on allowable temperature difference feedback Technical Field The invention relates to the field of turbine operation control, in particular to a turbine lifting rate and steam temperature cooperative control method based on allowable temperature difference feedback. Background The steam turbine is used as core power equipment of a thermal power plant and plays a key role in the deep peak shaving process of an electric power system. In order to ensure the operation safety of the unit in large-amplitude load fluctuation, the thermal stress and service life loss of key high-temperature components such as a rotor and the like need to be closely monitored and controlled. In the related art, the Chinese patent publication No. CN120175433A discloses a turbine operation control method, an intelligent body and electronic equipment. The technology obtains the corresponding predicted stress parameters by obtaining the current operation parameters of the high-temperature part of the steam turbine and respectively inputting the stress prediction models of the start-stop stage and the load adjustment stage, further calculates the predicted values of the low-cycle fatigue crack initiation loss and the accumulated loss, finally determines the target depth peak regulation operation control parameters meeting the constraint conditions of stress and service life and controls the steam turbine. Aiming at the related technology, by utilizing a single-stage stress prediction model and a traditional life loss evaluation method, it is difficult to comprehensively consider the dynamic change of a safety boundary generated by the accumulation of historical service life of a rotor under a complex working condition, and the feedforward compensation on mechanical hysteresis of a mechanism and total disturbance of a system is lacked in an execution level, so that control precision and response quality still have room to be improved, and economic operation of the whole life cycle of a unit is not facilitated. Disclosure of Invention In order to solve the problems, the invention provides a cooperative control method for the rise rate and the steam temperature of a steam turbine based on allowable temperature difference feedback, which adopts a means of analyzing the internal thermal stress of a rotor in real time, dynamically evaluating accumulated damage to correct a safety boundary and cooperatively distributing the rise rate and a steam temperature control vector to perform feedforward compensation control, so that the accurate, safe and efficient control on the thermal stress of the steam turbine in the deep peak shaving process can be realized. In order to achieve the above purpose, the application adopts the following technical scheme: The method comprises the steps of collecting operation data of a steam turbine in a deep peak regulation process in real time, carrying out cleaning and normalization processing on the operation data through a thermodynamic equation to form a basic data set for supporting thermal energy gradient calculation, wherein the operation data comprises main steam pressure, steam flow, a regulating-stage enthalpy value and multi-point measured temperature distributed on the surface of a steam turbine rotor, inputting the basic data set into a preset unsteady thermal conduction analysis model, utilizing a material heat conductivity coefficient, specific heat capacity and a surface heat exchange coefficient determined by steam flow and pressure of the rotor to analyze an equivalent strain field of an internal non-measurable point of the rotor in real time to generate thermal state prediction data representing a real-time evolution track of the rotor thermal stress, calling a starting-stopping frequency and a load fluctuation amplitude in the historical operation file of the rotor, establishing a mathematical model for superposition of cyclic fatigue damage caused by the alternating thermal stress and creep damage accumulated in time under high-temperature conditions, calculating to obtain a current accumulated fatigue damage value of the rotor, utilizing a fracture correction function to carry out linear expansion decision-making a limit value or a limit value, utilizing a thermal expansion coefficient, carrying out a linear decision-making coefficient, and a dynamic decision-making a dynamic state prediction command according to a thermal stress, determining a limit value, and a dynamic state-dependent on a thermal stress distribution command, and a dynamic state-dependent on a thermal expansion command, generating a cooperative decision, and a thermal stress-dependent on a thermal state prediction command, and capturing mechanical hysteresis and system total disturbance of an executing mechanism by using an extended state observer, performing feedforward compensation correction on the PID control instruction, generating a final c