A method and system for multi-variable coordinated control of a unit

Abstract

The application relates to the technical field of industrial automation control, and discloses a unit multivariable coordinated control method and system, the method comprising the following steps: collecting real-time operation data of a unit; inputting the real-time operation data into a preset disturbance observer, and extracting a lumped disturbance sequence representing internal state deviation of the unit; calculating a feedforward compensation matrix for dynamic decoupling based on the lumped disturbance sequence and preset coupling parameters; superimposing the feedforward compensation matrix and a basic control signal generated based on a load instruction to generate a target control instruction; judging whether a coal feeder in an execution mechanism is in a preset nonlinear dead zone friction hindering interval according to the real-time operation data; and if the coal feeder is in the nonlinear dead zone friction hindering interval, calculating a dead zone inverse compensation component based on a mechanical friction transmission model of the coal feeder and superimposing the dead zone inverse compensation component into the target control instruction; and the application can improve the safety and stability of unit operation.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, and in particular to a multivariable coordinated control method and system for a generator set. Background Technology

[0002] With the continuous adjustment of the energy structure, thermal power units and gas-steam combined cycle units need to participate deeply in the peak shaving tasks of the power grid. The load commands of the units are characterized by high frequency and large fluctuation. In the multivariable coordinated control system of the units, the combustion heat generation process on the boiler side and the steam work process on the turbine side constitute two deeply coupled controlled objects.

[0003] In existing technologies, static decoupling control based on empirical formulas or conventional fuzzy proportional-integral-derivative (FID) control schemes are typically used to coordinate the actions of the boiler and turbine. However, these schemes have a hidden but specific underlying computational pain point: the thermodynamic process of the boiler combustion system from coal feeding and pulverization to steam production has a pure time delay of several minutes and a large thermal inertia, while the time constant of the turbine regulating valve action affecting the actual power is only on the order of milliseconds to seconds. When the load command changes frequently and significantly, this severe dynamic response time mismatch causes the static decoupling matrix to completely fail. Traditional algorithms can only provide lag feedback based on the measurement deviation at the current moment, and cannot calculate and compensate for changes in the internal heat accumulation state of the unit in real time, which in turn causes the main steam pressure to fluctuate significantly overshoot during load changes, seriously threatening the safe and stable operation of the unit.

[0004] Therefore, how to maintain the safe and stable operation of the unit has become an urgent problem to be solved. Summary of the Invention

[0005] This invention provides a multivariable coordinated control method and system for generating units to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a multivariable coordinated control method for a generating unit, comprising: S101, Collect real-time operating data of the unit, including load command, main steam pressure, actual power, regulating valve opening and coal feed rate; S102, the real-time operating data is input into a preset disturbance observer to extract the lumped disturbance sequence characterizing the internal state deviation of the unit; wherein, the disturbance observer is constructed based on the state-space equation, the numerical range of the boiler inertial time parameter matrix in the state-space equation is constrained to the physical transmission delay time from coal feeding to steam production in the unit design parameters, and the bandwidth parameter of the disturbance observer is quantized and tuned according to the highest frequency of heat load fluctuation in the historical operating data of the unit. S103, Based on the lumped perturbation sequence and the preset coupling parameters, calculate the feedforward compensation matrix for dynamic decoupling; S104, the feedforward compensation matrix is ​​superimposed with the basic control signal generated based on the load command to generate a target control command; the real-time operating data is used to determine whether the coal feeder in the actuator is in a preset nonlinear dead zone friction resistance range; if it is in the nonlinear dead zone friction resistance range, the dead zone inverse compensation component is calculated based on the mechanical friction transmission model of the coal feeder and superimposed on the target control command to overcome the underlying mechanical delay resistance of the coal feeding action; the superimposed target control command is output to the actuator of the unit.

[0007] In a preferred embodiment, the real-time operating data of the data acquisition unit includes: The raw sensor data of the unit is acquired in a preset control cycle; A preset filtering algorithm is invoked to filter out high-frequency noise from the original sensor data, and the real-time operating data is output; the cutoff frequency of the filtering algorithm is set based on the thermodynamic inertial time constant of the unit.

[0008] In a preferred embodiment, the step of inputting the real-time operating data into a preset disturbance observer and extracting the lumped disturbance sequence characterizing the internal state deviation of the unit includes: A state-space equation is constructed based on the energy conservation mechanism model of the unit, and the real-time operating data is used as the input variable of the state-space equation. The main steam pressure prediction value and the actual power prediction value are obtained by the computer group using the state-space equations; Calculate the first residual between the predicted main steam pressure and the main steam pressure in the real-time operating data, and the second residual between the predicted actual power and the actual power in the real-time operating data; The first residual and the second residual are input into the state update matrix of the disturbance observer for iterative calculation to extract the lumped disturbance sequence.

[0009] In a preferred embodiment, the step of iteratively calculating the state update matrix of the disturbance observer by inputting the first residual and the second residual into the disturbance observer to extract the lumped disturbance sequence includes: The first residual and the second residual are nonlinearly amplified by a preset nonlinear error feedback function. The amplified first residual and second residual are multiplied by the corresponding observer feedback gain coefficients to calculate the state differential estimate. The state differential estimate is integrated to output the lumped disturbance sequence that characterizes the boiler thermal hysteresis disturbance and the turbine flow disturbance.

[0010] In a preferred embodiment, calculating the feedforward compensation matrix for dynamic decoupling based on the lumped perturbation sequence and preset coupling parameters includes: Extract the rate of change of the lumped disturbance sequence between the current control period and the previous control period; The perturbation rate of change is input into a preset parameter regulator to calculate the feedforward gain adaptive adjustment coefficient. The feedforward compensation matrix is ​​calculated by multiplying the adaptive adjustment coefficient of the feedforward gain by the coupling parameter.

[0011] In a preferred embodiment, the step of superimposing the feedforward compensation matrix with the basic control signal generated based on the load command to generate the target control command includes: The load command is input into a preset proportional-integral-derivative controller to calculate the first basic signal of the main steam pressure control loop and the second basic signal of the actual power control loop, which are used as the basic control signals. The boiler-side feedforward component in the feedforward compensation matrix is ​​added to the first basic signal to generate the target coal feed command. The turbine-side feedforward component in the feedforward compensation matrix is ​​added to the second basic signal to generate the target regulating valve opening command; The target coal feed rate command and the target regulating valve opening command are used as the reference values ​​for the target control command.

[0012] In a preferred embodiment, the process of generating the target control command further includes: Calculate the directional consistency between the first basic signal and the boiler-side feedforward component in the feedforward compensation matrix; When the directional consistency is reversed and the absolute value of the boiler-side feedforward component is greater than the absolute value of the first basic signal, the boiler-side feedforward component is attenuated to a preset ratio to resolve the technical conflict between the rapid response requirement under a large load command step and the risk of critical stress rupture of water-cooled wall tubes caused by the oscillation of the boiler's internal thermal energy.

[0013] In a preferred embodiment, before outputting the superimposed target control command to the actuator of the unit, the method further includes: Monitor the main steam pressure in the real-time operating data; When the main steam pressure exceeds the preset safety threshold boundary, the feedforward compensation matrix is ​​forcibly cleared to zero, and the target control command is switched to the preset safety retreat mode for output, terminating the current dynamic decoupling control.

[0014] In a preferred embodiment, an iterative optimization step for control parameters is also included: Real-time acquisition of execution feedback data of target control commands, including actual coal feed rate of coal feeder, actual opening degree of regulating valve, steady-state deviation of main steam pressure and actual power tracking deviation; Based on the execution feedback data, calculate the bandwidth parameter correction amount of the disturbance observer and the gain correction coefficient of the feedforward compensation matrix; The bandwidth parameters of the disturbance observer are tuned according to the bandwidth parameter correction amount, and the feedforward compensation moment is updated according to the gain correction coefficient.

[0015] In a preferred embodiment, claim 2 To address the above problems, the present invention also provides a multivariable coordinated control system for a generator unit, the system comprising: The real-time data acquisition module is used to collect the unit's real-time operating data, which includes load commands, main steam pressure, actual power, regulating valve opening, and coal feed rate. The lumped disturbance extraction module is used to input the real-time operating data into a preset disturbance observer and extract the lumped disturbance sequence characterizing the internal state deviation of the unit. The disturbance observer is constructed based on the state-space equation, and the numerical range of the boiler inertial time parameter matrix in the state-space equation is constrained to the physical transmission delay time from coal feeding to steam production in the unit design parameters. The bandwidth parameter of the disturbance observer is quantized and tuned according to the highest frequency of heat load fluctuation in the historical operating data of the unit. The feedforward compensation calculation module is used to calculate the feedforward compensation matrix for dynamic decoupling based on the lumped perturbation sequence and preset coupling parameters. The control command generation and execution module is used to superimpose the feedforward compensation matrix with the basic control signal generated based on the load command to generate a target control command; determine whether the coal feeder in the actuator is in a preset nonlinear dead zone friction resistance range based on the real-time operating data; if it is in the nonlinear dead zone friction resistance range, calculate the dead zone inverse compensation component based on the mechanical friction transmission model of the coal feeder and superimpose it into the target control command to overcome the underlying mechanical delay resistance of the coal feeding action; and output the superimposed target control command to the actuator of the unit. Compared with the prior art, the present invention has the following beneficial effects: 1. This invention extracts the lumped disturbance sequence by constructing a disturbance observer based on state-space equations, constrains the boiler inertial time parameter matrix to the physical transmission delay time from coal feeding to steam production, and quantifies and tunes the observer bandwidth parameter according to the highest frequency of historical heat load fluctuations. This effectively avoids disturbance identification deviations caused by unreasonable disturbance observer parameter settings, and can accurately capture internal state deviations of the unit such as boiler heat hysteresis and turbine flow fluctuations. It solves the problems of fuzzy and lagging disturbance identification in the prior art, and lays a solid foundation for subsequent dynamic decoupling control. 2. This invention calculates a feedforward compensation matrix for dynamic decoupling based on lumped disturbance sequence and coupling parameters, and adaptively adjusts the feedforward gain in combination with the disturbance change rate. It can specifically offset the coupling interference between multiple variables such as load command, main steam pressure, and actual power, so that the control loops on the boiler side and the turbine side can be matched in coordination. This solves the problems of control response lag and parameter oscillation caused by multivariate coupling in traditional control, significantly improves the unit's tracking response speed to load command, and reduces the steady-state deviation between main steam pressure and actual power. 3. This invention determines whether the coal feeder is in the nonlinear dead zone friction resistance range, calculates the dead zone inverse compensation component based on the mechanical friction transmission model, and superimposes it onto the target control command. This effectively counteracts the underlying mechanical delay and friction resistance of the coal feeding action, solves the problems of coal feeder dead zone causing coal feeder control deviation and action lag, ensures that the coal feeder and regulating valve opening can accurately respond to the target control command, and improves the matching degree of unit fuel supply and energy output. 4. Before the target control command is output, the present invention adds a main steam pressure over-threshold monitoring mechanism. When the main steam pressure exceeds the safety threshold boundary, the feedforward compensation matrix is ​​forcibly cleared and switched to the safety withdrawal mode. This effectively mitigates critical risks such as water-cooled wall tube rupture caused by the oscillation of the boiler's internal energy storage under high load step, while avoiding abnormal unit operation caused by disturbance runaway, ensuring the long-term safe and stable operation of the unit, and improving the safety and robustness of the control method. 5. This invention adds an iterative optimization step for control parameters. By collecting the execution feedback data of the target control command, the bandwidth parameters of the disturbance observer are dynamically tuned and the gain of the feedforward compensation matrix is ​​updated. This solves the problem that traditional control parameters are fixed and cannot adapt to unit load fluctuations and changes in operating conditions. It enables the control parameters to be dynamically adjusted according to the unit's operating status, continuously improving the coordinated control accuracy under different load conditions and broadening the adaptability of the control method. 6. Through precise coordination control and parameter optimization, the unit can achieve accurate matching of main steam pressure and actual power under different load conditions, reducing fuel waste and energy loss, and lowering unit operating costs; at the same time, it avoids frequent unit adjustments due to control deviations, reduces pollutant emissions, and balances economic efficiency with energy conservation and environmental protection requirements, thus having significant industrial application value and practical significance. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a multivariable coordinated control method for a generator unit according to an embodiment of the present invention. Figure 2 This is a functional block diagram of a multivariable coordinated control system for a generator unit provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0018] This application provides a multivariate coordinated control method for generating units. The executing entity of this multivariate coordinated control method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the multivariate coordinated control method for generating units can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0019] Reference Figure 1 The diagram shown is a flowchart illustrating a multivariable coordinated control method for a generator unit according to an embodiment of the present invention.

[0020] In some embodiments, a multivariable coordinated control method for a generating unit provided in this application includes: collecting real-time operating data of the generating unit, the real-time operating data including load command, main steam pressure, actual power, regulating valve opening, and coal feed rate; inputting the real-time operating data into a preset disturbance observer to extract a lumped disturbance sequence characterizing the internal state deviation of the generating unit; wherein, the disturbance observer is constructed based on state-space equations, the numerical range of the boiler inertial time parameter matrix in the state-space equations is constrained to the physical transmission delay time from coal feed to steam production in the generating unit design parameters, and the bandwidth parameter of the disturbance observer is based on the heat load fluctuations in the historical operating data of the generating unit. The highest frequency is quantized and tuned; based on the lumped disturbance sequence and preset coupling parameters, a feedforward compensation matrix for dynamic decoupling is calculated; the feedforward compensation matrix is ​​superimposed with the basic control signal generated based on the load command to generate a target control command; based on the real-time operating data, it is determined whether the coal feeder in the actuator is in a preset nonlinear dead zone friction resistance range; if it is in the nonlinear dead zone friction resistance range, based on the mechanical friction transmission model of the coal feeder, the dead zone inverse compensation component is calculated and superimposed on the target control command to overcome the underlying mechanical delay resistance of the coal feeding action; the superimposed target control command is output to the actuator of the unit. The mechanical friction transmission model polynomial stored in the register is invoked. This polynomial is a nonlinear fitting polynomial that considers the static friction and dynamic viscous friction of the system, and its dead zone inverse compensation component is included. The calculation formula is as follows: ; in, The target speed command for the coal feeder is obtained by converting the target control command; It is a symbolic function; The target speed command for the coal feeder is obtained by converting the target control command; The critical constant for dead zone start-up (characterizing the reference compensation amount required to overcome the maximum static friction of the underlying physical equipment). This is the first-order viscous friction compensation coefficient; These are the second-order nonlinear damping fitting coefficients; the above parameters , , The data was obtained by fitting historical open-loop step test data of the coal feeder under no-load and full-load conditions. Perform an arithmetic accumulation operation; calculate the dead-zone inverse compensation component. The target control command is superimposed on the target control command to pre-boost the target control command; thereby eliminating the blind spot in the underlying communication response caused by the structuring of cross-border physical equipment at the source, and ensuring that the control signal can be executed without delay after being output to the unit actuator through the digital-to-analog conversion interface.

[0021] Specifically, the system collects low-level sensor signals from the data interface of the distributed control system; real-time operating data characterizes the unit's current actual operating position and energy state in the physical space; multi-dimensional heterogeneous physical data is input into the disturbance observer algorithm module running in the controller's memory; the boiler body manufacturing data manual is extracted; the abstract parameter matrix in the control algorithm is forcibly bound to the actual physical transmission delay time; the algorithm topology is ensured to deeply map the physical entity characteristics; the frequency domain analysis results of the fast Fourier transform are collected; the highest frequency point of heat load fluctuation is extracted; and it is directly assigned to the bandwidth parameter of the disturbance observer. Equations; This strong-foundation strategy, which abandons pure mathematical derivation and deeply integrates the algorithm network structure with specific physical characteristics, enables the observer model to have rigorous physical interpretability and absolute convergence guarantee in engineering applications. For example, if the physical coal drop delay from the coal feeder to the coal mill of a 300MW unit is 20 seconds, then the time constant corresponding to the eigenvalue of the boiler inertial time parameter matrix must be no less than 20 seconds; if the highest frequency of the primary frequency regulation fluctuation of the power grid is 0.1Hz, then the bandwidth parameter of the observer is quantized to be 0.3Hz to 0.5Hz to ensure that it can both cover the disturbance frequency band and suppress high-frequency measurement noise. The disturbance observer performs matrix operations to extract the lumped disturbance sequence characterizing the internal state deviation of the unit. The arithmetic logic unit of the central processing unit is called to perform multiplication and addition operations to calculate the feedforward compensation matrix.

[0022] In some embodiments, to further ensure that the target control command calculated by the algorithm can be accurately converted into mechanical actions in the physical world, the feeder speed or inverter feedback current is extracted; state interval comparison logic is executed; the nonlinear dead zone friction obstacle interval quantifies the technical obstacle of static friction causing the command to be sent but the machine does not move; the mechanical friction transmission model polynomial stored in the register is called; the dead zone inverse compensation component that needs to be additionally overcome to overcome static friction is calculated; an arithmetic accumulation operation is performed; and the target control command is pre-pressurized; thereby eliminating the underlying communication response blind spot caused by the cross-border physical equipment structure at the source, and ensuring that the control signal can be executed without delay after being output to the unit actuator through the digital-to-analog conversion interface.

[0023] In some embodiments, the acquisition of real-time operating data of the unit includes: acquiring raw sensor data of the unit at a preset control cycle; calling a preset filtering algorithm to filter out high-frequency noise from the raw sensor data, and outputting the real-time operating data; the cutoff frequency of the filtering algorithm is set based on the thermodynamic inertial time constant of the unit.

[0024] Specifically, the driver hardware analog-to-digital converter periodically reads the level signals on the sensor bus; the control cycle defines the time granularity of the interaction between the computer system and the physical world; the raw sensor data containing electromagnetic interference is sent to the filter function; digital smoothing calculation is performed; and real-time operating data without sudden spikes is output; the physical boundaries of the filtering algorithm are strictly constrained, and its cutoff frequency parameter must match the inherent thermodynamic inertia time constant of the unit; for example, for the coal feed signal with extremely high thermal inertia, a first-order hysteresis filtering algorithm is called to filter out the mechanical vibration noise of the belt scale using its low-pass characteristic; for the control valve opening signal with extremely fast response, a moving average filtering algorithm is called to ensure the real-time performance of the control system using its small time window size and to prevent the introduction of additional artificial calculation delays due to over-filtering.

[0025] In some embodiments, the step of inputting the real-time operating data into a preset disturbance observer to extract a lumped disturbance sequence characterizing the internal state deviation of the unit includes: constructing a state-space equation based on the unit's energy conservation mechanism model, and using the real-time operating data as the input variable of the state-space equation; calculating the predicted value of the main steam pressure and the predicted value of the actual power of the unit through the state-space equation; calculating a first residual between the predicted value of the main steam pressure and the main steam pressure in the real-time operating data, and a second residual between the predicted value of the actual power and the actual power in the real-time operating data; inputting the first residual and the second residual into the state update matrix of the disturbance observer for iterative calculation to extract the lumped disturbance sequence.

[0026] Specifically, memory space is allocated to instantiate the reduced-order thermodynamic model of the unit; the state-space equations define the idealized digital twin boundary of energy flow within the unit, and its discretized mathematical expression is: ,in, This is the system state vector that includes the main steam pressure and actual power. The control vector includes the coal feed rate and the valve opening; the state matrix. Corresponding to the aforementioned boiler inertia time parameter matrix, the control matrix The physical gain of the mapping actuator, For the next control cycle, the system state vector containing the main steam pressure and actual power is generated; real-time operating data is injected; the numerical solution routine of the ordinary differential equation system is triggered; the predicted values ​​of the main steam pressure and actual power of the unit for the next control cycle are output; data bit width alignment and arithmetic subtraction instructions are executed; the first and second residuals between the theoretical predicted values ​​and the actual observed values ​​are calculated; the generation of the residuals directly reveals the unmodeled physical disturbances that are occurring inside the unit and deviating from the ideal state, such as sudden changes in coal calorific value; the first and second residuals are sent into the feedback loop of the state update matrix; the iterative update mechanism is triggered; and the dimensionally unified lumped disturbance sequence is extracted, which is equivalently reconstructed in the underlying calculation link of the black-box internal complex thermodynamic disturbance terms.

[0027] In some embodiments, the step of inputting the first residual and the second residual into the state update matrix of the disturbance observer for iterative calculation to extract the lumped disturbance sequence includes: performing nonlinear gain amplification on the first residual and the second residual using a preset nonlinear error feedback function; multiplying the amplified first residual and the second residual by the corresponding observer feedback gain coefficient to calculate the state differential estimate; performing an integral operation on the state differential estimate to output the lumped disturbance sequence characterizing the boiler thermal hysteresis disturbance and the turbine flow disturbance.

[0028] Specifically, the nonlinear error feedback function is constructed by calling the piecewise function module in the system's mathematical library as follows: The mathematical logic definition is: when the absolute value of the input residual is... When outputting a non-linear power result ;when When outputting linear scaling results ,in Characterizing the first residual or the second residual, For nonlinear factors in the interval (0,1), The linear region width is defined by a nonlinear error feedback function, which delineates the anti-bounce optimization boundary for large errors with small gains and small errors with large gains. The first and second residuals are numerically reshaped; the observer feedback gain coefficients pre-stored in non-volatile memory are extracted; matrix multiplication is performed; the state differential estimate reflecting the rate of change of state variables is calculated; the state differential estimate is fed into an Euler integrator or a Runge-Kutta integrator; discrete-time accumulation is performed as the time step progresses; finally, the lumped disturbance sequence characterizing the boiler thermal hysteresis disturbance and the turbine flow disturbance is output. This mechanism ensures the global asymptotic stability of the observation and calculation process.

[0029] In some embodiments, the preset coupling parameters are not subjectively set constants, but rather physical mapping values ​​obtained through offline calibration based on historical operating data.

[0030] Specifically, historical operating data of the unit under steady-state conditions and preset step disturbance conditions, such as a preset percentage step change in coal feed rate, is extracted; the recursive least squares system identification algorithm is called to fit and calculate the historical operating data; the cross-influence gain of main steam pressure and actual power under different disturbances is obtained; the cross-influence gain is written into non-volatile memory as the preset coupling parameter; this offline identification and calibration mechanism provides a white-box benchmark with physical mechanism support for the calculation of the feedforward compensation matrix for dynamic decoupling, eliminating the implementation obstacles of the core calculation link of dynamic decoupling.

[0031] In some embodiments, calculating the feedforward compensation matrix for dynamic decoupling based on the lumped disturbance sequence and preset coupling parameters includes: extracting the disturbance change rate of the lumped disturbance sequence between the current control cycle and the previous control cycle; inputting the disturbance change rate into a preset parameter regulator to calculate the feedforward gain adaptive adjustment coefficient; and multiplying the feedforward gain adaptive adjustment coefficient by the coupling parameters to calculate the feedforward compensation matrix. The feedforward gain adaptive adjustment coefficient is output through an embedded linear proportional formula. The specific calculation formula for the linear proportional formula is as follows: ; In the formula, This is the adaptive adjustment coefficient for the feedforward gain. The extracted lumped disturbance sequence represents the rate of change of disturbance between the current control period and the previous control period; The preset regulator proportional gain constant; The set noise shielding dead zone threshold is the rate of change of disturbance. The function is designed to maximize the value; the calculation logic defines that when the absolute value of the disturbance rate of change is less than a threshold... When the disturbance changes drastically and exceeds the threshold, the adjustment coefficient is increased linearly with the rate of change of the disturbance. This coefficient gives the static algorithm dynamic time-awareness. The coupling parameters of the static calibration are extracted. Scalar multiplication of the matrix is ​​performed. The feedforward compensation matrix is ​​calculated.

[0032] Specifically, the system reads historical disturbance data cached in the stack; performs first-order difference calculations; the disturbance change rate quantifies the transient evolution trend of the heat supply and demand imbalance within the unit; activates the parameter regulator logic unit; outputs the feedforward gain adaptive adjustment coefficient through an embedded mapping table or linear proportional formula; this coefficient gives the static algorithm dynamic time-awareness; extracts the statically calibrated coupling parameters; performs scalar multiplication of the matrix; calculates the feedforward compensation matrix; through the above data flow reconstruction, changes in the physical environment that were originally unknown to the outside world, such as delayed steam production caused by deterioration in coal quality, are reverse-translated into objective numerical changes in the weights of the compensation matrix within the controller, thus achieving an objective underlying mapping of the experience.

[0033] In some embodiments, superimposing the feedforward compensation matrix with the basic control signal generated based on the load command to generate the target control command includes: inputting the load command into a preset proportional-integral-derivative controller; calculating a first basic signal of the main steam pressure control loop and a second basic signal of the actual power control loop as the basic control signal; adding the boiler-side feedforward component in the feedforward compensation matrix to the first basic signal to generate a target coal feed command; adding the turbine-side feedforward component in the feedforward compensation matrix to the second basic signal to generate a target regulating valve opening command; and using the target coal feed command and the target regulating valve opening command as the reference values ​​for the target control command.

[0034] Specifically, the system reads the power grid load dispatch data frame; the proportional-integral-derivative (PID) controller calculates the hysteresis control action to eliminate static steady-state error; and outputs the first and second basic signals. The system parses the data structure of the feedforward compensation matrix; extracts the boiler-side feedforward component represented by the intersection of the first row and first column of the matrix; performs arithmetic addition of dimensionless data; synthesizes the target coal feed command; extracts the turbine-side feedforward component representing the turbine energy release demand from the matrix; performs arithmetic addition with the second basic signal; synthesizes the target regulating valve opening command; packages the synthesized command data into a standard industrial control protocol message; and after executing subsequent boundary blocking and dead-zone compensation logic, generates the target control command to drive the field electromechanical equipment.

[0035] In some embodiments, the process of generating the target control command further includes: calculating the directional consistency between the first basic signal and the boiler-side feedforward component in the feedforward compensation matrix; when the directional consistency is opposite and the absolute value of the boiler-side feedforward component is greater than the absolute value of the first basic signal, attenuating the boiler-side feedforward component to a preset ratio to resolve the technical conflict between the rapid response requirement under a large load command step and the risk of critical stress rupture of water-cooled wall tubes caused by the oscillation of the boiler's internal thermal energy. Specifically, the following signed product function operation is performed to determine the consistency of the vector direction between the basic feedback action and the feedforward prediction action: ; In the formula, This is the first basic signal (i.e., the basic action quantity calculated by the main steam pressure control loop). The boiler-side feedforward component; sign function The operational logic is defined as follows: when the variable When the value is greater than 0, the output value is 1; when When the value is less than 0, the output value is -1; when When the value is 0, the output value is 0.

[0036] When the calculation result When the result equals 1, it is determined that the directions are consistent; when the calculation result... When the result is -1, it is determined that the direction is reversed; when the result is 0, it is determined that there is no action at the critical point; the direction consistency assessment defines the boundary condition where the system's regulating energy has internal cancellation; the judgment instruction of the logic AND gate is triggered; when the direction consistency is determined to be reversed (i.e., The absolute value of the boiler-side feedforward component is greater than the absolute value of the first base signal (i.e., equal to 1), and the absolute value of the boiler-side feedforward component is greater than the absolute value of the first base signal (i.e., equal to 1). When the dynamic compensation force is too large, the risk of system reverse overshoot may be identified; the shift register or divider is called to perform numerical scaling; the feedforward component on the boiler side is forcibly attenuated to a preset ratio (e.g., 30% to 50% of the original value) to resolve the technical conflict between the rapid response requirements under the step of the large load command and the critical stress risk of water-cooled wall tube rupture caused by the oscillation of the boiler's internal energy storage.

[0037] Specifically, the system performs symbolic function operations; determines the vector directions of the basic feedback action and the feedforward prediction action; and assesses directional consistency to define the boundary conditions where the system's regulating energy may have internal cancellation. It triggers the judgment instructions of the AND gate; identifies the risk of system reverse overshoot that may result from excessive dynamic compensation; calls a shift register or divider for numerical scaling; and forcibly attenuates the boiler-side feedforward component. When the boundary logic rules take effect in this specific scenario, they effectively suppress the high-frequency spike fluctuations of the underlying main steam pressure curve, resolve the core technical conflict between control rate and system internal energy stability in TRIZ theory, and achieve a smooth transition of control commands under extreme conditions while ensuring the safety of physical equipment.

[0038] In some embodiments, before outputting the superimposed target control command to the actuator of the unit, the method further includes: monitoring the main steam pressure in the real-time operating data; when the main steam pressure exceeds a preset safety threshold boundary, forcibly clearing the feedforward compensation matrix to zero, and switching the target control command to a preset safety pullback mode for output, thereby terminating the current dynamic decoupling control.

[0039] Specifically, the system continuously polls the memory address value of the main steam pressure; executes a threshold comparison interrupt routine; defines the maximum physical stress limit that the unit's thermal pipelines can withstand; triggers a high-priority interrupt response mechanism; writes zero-value coefficients to the multiplier; forcibly blocks data flow in the feedforward compensation loop; calls the redundant instruction set of the safety rollback mode; outputs a conservative control sequence; and completely avoids the output of uncontrollable instructions by complex algorithms when measuring instruments fail or extreme anomalies occur, ensuring the absolute survival baseline of industrial field equipment.

[0040] like Figure 2 The diagram shown is a functional block diagram of a multivariable coordinated control system for a generator set provided in an embodiment of the present invention.

[0041] The multivariable coordinated control system for a generating unit described in this invention can be installed in an electronic device. Depending on the functions implemented, the multivariable coordinated control system may include a real-time data acquisition module, a lumped disturbance extraction module, a feedforward compensation calculation module, and a control command generation and execution module. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, and are stored in the memory of the electronic device.

[0042] In this embodiment, the functions of each module / unit are as follows: The real-time data acquisition module is used to collect the real-time operating data of the unit, including load command, main steam pressure, actual power, regulating valve opening and coal feed rate. The lumped disturbance extraction module is used to input the real-time operating data into a preset disturbance observer and extract the lumped disturbance sequence characterizing the internal state deviation of the unit. The disturbance observer is constructed based on the state-space equation, and the numerical range of the boiler inertial time parameter matrix in the state-space equation is constrained to the physical transmission delay time from coal feeding to steam production in the unit design parameters. The bandwidth parameter of the disturbance observer is quantized and tuned according to the highest frequency of heat load fluctuation in the historical operating data of the unit. The feedforward compensation calculation module is used to calculate the feedforward compensation matrix for dynamic decoupling based on the lumped perturbation sequence and preset coupling parameters. The control command generation and execution module is used to superimpose the feedforward compensation matrix with the basic control signal generated based on the load command to generate a target control command; determine whether the coal feeder in the actuator is in a preset nonlinear dead zone friction resistance range based on the real-time operating data; if it is in the nonlinear dead zone friction resistance range, calculate the dead zone inverse compensation component based on the mechanical friction transmission model of the coal feeder and superimpose it into the target control command to overcome the underlying mechanical delay resistance of the coal feeding action; and output the superimposed target control command to the actuator of the unit.

[0043] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0044] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0045] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0046] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0047] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A multivariable coordinated control method for a generator unit, characterized in that, The method includes: S101, Collect real-time operating data of the unit, including load command, main steam pressure, actual power, regulating valve opening and coal feed rate; S102, the real-time operating data is input into a preset disturbance observer to extract the lumped disturbance sequence characterizing the internal state deviation of the unit; wherein, the disturbance observer is constructed based on the state-space equation, the numerical range of the boiler inertial time parameter matrix in the state-space equation is constrained to the physical transmission delay time from coal feeding to steam production in the unit design parameters, and the bandwidth parameter of the disturbance observer is quantized and tuned according to the highest frequency of heat load fluctuation in the historical operating data of the unit. S103, Based on the lumped perturbation sequence and the preset coupling parameters, calculate the feedforward compensation matrix for dynamic decoupling; S104, the feedforward compensation matrix is ​​superimposed with the basic control signal generated based on the load command to generate a target control command; the real-time operating data is used to determine whether the coal feeder in the actuator is in a preset nonlinear dead zone friction resistance range; if it is in the nonlinear dead zone friction resistance range, the dead zone inverse compensation component is calculated based on the mechanical friction transmission model of the coal feeder and superimposed on the target control command to overcome the underlying mechanical delay resistance of the coal feeding action; the superimposed target control command is output to the actuator of the unit.

2. The multivariable coordinated control method for a generating unit according to claim 1, characterized in that, The real-time operating data of the data acquisition unit includes: The raw sensor data of the unit is acquired in a preset control cycle; A preset filtering algorithm is invoked to filter out high-frequency noise from the original sensor data, and the real-time operating data is output; the cutoff frequency of the filtering algorithm is set based on the thermodynamic inertial time constant of the unit.

3. The multivariable coordinated control method for a generating unit according to claim 1, characterized in that, The step of inputting the real-time operating data into a preset disturbance observer and extracting the lumped disturbance sequence characterizing the internal state deviation of the unit includes: A state-space equation is constructed based on the energy conservation mechanism model of the unit, and the real-time operating data is used as the input variable of the state-space equation. The main steam pressure prediction value and the actual power prediction value are obtained by the computer group using the state-space equations; Calculate the first residual between the predicted main steam pressure and the main steam pressure in the real-time operating data, and the second residual between the predicted actual power and the actual power in the real-time operating data; The first residual and the second residual are input into the state update matrix of the disturbance observer for iterative calculation to extract the lumped disturbance sequence.

4. The multivariable coordinated control method for a generator unit according to claim 3, characterized in that, The step of iteratively calculating the lumped disturbance sequence by inputting the first residual and the second residual into the state update matrix of the disturbance observer includes: The first residual and the second residual are nonlinearly amplified by a preset nonlinear error feedback function. The amplified first residual and second residual are multiplied by the corresponding observer feedback gain coefficients to calculate the state differential estimate. The state differential estimate is integrated to output the lumped disturbance sequence that characterizes the boiler thermal hysteresis disturbance and the turbine flow disturbance.

5. The multivariable coordinated control method for a generating unit according to claim 1, characterized in that, The step of calculating the feedforward compensation matrix for dynamic decoupling based on the lumped perturbation sequence and preset coupling parameters includes: Extract the rate of change of the lumped disturbance sequence between the current control period and the previous control period; The perturbation rate of change is input into a preset parameter regulator to calculate the feedforward gain adaptive adjustment coefficient. The feedforward compensation matrix is ​​calculated by multiplying the adaptive adjustment coefficient of the feedforward gain by the coupling parameter.

6. The multivariable coordinated control method for a generating unit according to claim 1, characterized in that, The step of superimposing the feedforward compensation matrix with the basic control signal generated based on the load command to generate the target control command includes: The load command is input into a preset proportional-integral-derivative controller to calculate the first basic signal of the main steam pressure control loop and the second basic signal of the actual power control loop, which are used as the basic control signals. The boiler-side feedforward component in the feedforward compensation matrix is ​​added to the first basic signal to generate the target coal feed command. The turbine-side feedforward component in the feedforward compensation matrix is ​​added to the second basic signal to generate the target regulating valve opening command; The target coal feed rate command and the target regulating valve opening command are used as the reference values ​​for the target control command.

7. The multivariable coordinated control method for a generator unit according to claim 6, characterized in that, The process of generating target control instructions also includes: Calculate the directional consistency between the first basic signal and the boiler-side feedforward component in the feedforward compensation matrix; When the directional consistency is reversed and the absolute value of the boiler-side feedforward component is greater than the absolute value of the first basic signal, the boiler-side feedforward component is attenuated to a preset ratio to resolve the technical conflict between the rapid response requirement under a large load command step and the risk of critical stress rupture of water-cooled wall tubes caused by the oscillation of the boiler's internal thermal energy.

8. The multivariable coordinated control method for a generating unit according to claim 1, characterized in that, Before outputting the superimposed target control command to the unit's actuator, the method further includes: Monitor the main steam pressure in the real-time operating data; When the main steam pressure exceeds the preset safety threshold boundary, the feedforward compensation matrix is ​​forcibly cleared to zero, and the target control command is switched to the preset safety retreat mode for output, terminating the current dynamic decoupling control.

9. The multivariable coordinated control method for a generating unit according to claim 1, characterized in that, It also includes iterative optimization steps for control parameters: Real-time acquisition of execution feedback data of target control commands, including actual coal feed rate of coal feeder, actual opening degree of regulating valve, steady-state deviation of main steam pressure and actual power tracking deviation; Based on the execution feedback data, calculate the bandwidth parameter correction amount of the disturbance observer and the gain correction coefficient of the feedforward compensation matrix; The bandwidth parameters of the disturbance observer are tuned according to the bandwidth parameter correction amount, and the feedforward compensation matrix is ​​updated according to the gain correction coefficient.

10. A multivariable coordinated control system for a generator unit, characterized in that, The system, used to implement the multivariable coordinated control method for a generator unit as described in claim 1, comprises: The real-time data acquisition module is used to collect the real-time operating data of the unit, including load command, main steam pressure, actual power, regulating valve opening and coal feed rate. The lumped disturbance extraction module is used to input the real-time operating data into a preset disturbance observer and extract the lumped disturbance sequence characterizing the internal state deviation of the unit. The disturbance observer is constructed based on the state-space equation, and the numerical range of the boiler inertial time parameter matrix in the state-space equation is constrained to the physical transmission delay time from coal feeding to steam production in the unit design parameters. The bandwidth parameter of the disturbance observer is quantized and tuned according to the highest frequency of heat load fluctuation in the historical operating data of the unit. The feedforward compensation calculation module is used to calculate the feedforward compensation matrix for dynamic decoupling based on the lumped disturbance sequence and preset coupling parameters. The control command generation and execution module is used to superimpose the feedforward compensation matrix with the basic control signal generated based on the load command to generate a target control command; determine whether the coal feeder in the actuator is in a preset nonlinear dead zone friction resistance range based on the real-time operating data; if it is in the nonlinear dead zone friction resistance range, calculate the dead zone inverse compensation component based on the mechanical friction transmission model of the coal feeder and superimpose it into the target control command to overcome the underlying mechanical delay resistance of the coal feeding action; and output the superimposed target control command to the actuator of the unit.

Images