Thermal power generating unit AGC prediction control method and device based on cost optimization

Abstract

The invention discloses a thermal power generating unit AGC prediction control method and device based on cost optimization, and the method comprises the steps: obtaining the actual operation data and AGC load instruction of a thermal power generating unit, and giving a unit dynamic model; determining a steady-state operation target function from the current state to a steady state and considering a cost index, and determining a dynamic operation target function from continuous control from the current state to the steady state and considering the cost index; based on the steady-state operation objective function, in combination with an AGC load instruction and a unit dynamic model, giving an optimal steady-state parameter of the thermal power unit; based on the dynamic operation objective function, in combination with the optimal steady-state parameter, the AGC load instruction, the actual operation data and the unit dynamic model, a control parameter of the thermal power generating unit is given; and controlling the thermal power generating unit through the given control parameters. According to the method, the operation economical efficiency in the AGC load tracking control process of the thermal power generating unit can be remarkably improved, and the flexible adjustment and response supporting capacity of the thermal power generating unit is enhanced.

Description

Technical Field

[0001] This invention belongs to the field of thermal power unit control technology, specifically relating to a cost-optimized AGC predictive control method and device for thermal power units, and more particularly to a cost-optimized AGC predictive control method and device for supercritical coal-fired power generating units. Background Technology

[0002] With the continuous increase in the proportion of renewable energy installed capacity, the power system structure is undergoing profound changes. Traditional thermal power units are gradually shifting from basic power sources to regulating power sources, undertaking increasingly frequent peak-shaving and frequency regulation tasks. Supercritical coal-fired power generating units, due to their high thermal efficiency, large single-unit capacity, and strong load-changing capacity, remain an important supporting power source for the power system. However, frequent and large-scale load fluctuations place higher demands on their operation and control, requiring not only the maintenance of stable key parameters such as main steam pressure and power generation, but also further improvement in operating economy and reduction of operating costs such as coal consumption.

[0003] In actual dispatching, thermal power units often need to respond to AGC (Automatic Generation Control) load commands from the power grid. However, the actual output power of the thermal power units often deviates from the issued AGC load commands. This deviation not only affects the evaluation of the unit's regulation performance but also adversely impacts the grid's frequency regulation quality and the economic benefits of the thermal power units. Furthermore, due to the complex characteristics of thermal power units—multivariable, strongly coupled, nonlinear, and with large inertia—traditional PID control or tracking model predictive control struggles to coordinate various economic indicators during dynamic processes. For example, patent application CN106406080A discloses a remote optimization system and method for the AGC function of a thermal power generating unit. This system includes a PID controller used to control the actual load value of the thermal power generating unit based on the unit load error value and preset PID controller parameters. The unit load error value is the difference between the AGC load setpoint and the actual load value of the unit. A data acquisition device collects the actual load value, AGC load setpoint, and the output value of the PID controller in real time and transmits them to a remote data center server. The remote data center server is used to establish a first-order model of the thermal power generating unit. Based on the first-order model, the preset PID controller parameters are tuned using an internal model control algorithm. Finally, the tuned PID controller parameters are fed back to the PID controller. The PID controller is also used to control the actual load value of the thermal power generating unit based on the tuned PID controller parameters, ultimately realizing remote control of the thermal power generating unit's automatic power generation. Although existing studies have combined real-time optimization with lower-level tracking control, this two-layer structure achieves economic optimization through a hierarchical structure. After the upper layer determines the steady-state optimal setpoint, it directly sends it down to the lower-level controller (MPC) for tracking control. This approach still has limitations such as response lag and unreachable setpoints, making it difficult to adapt to the dynamic process of frequently changing actual operating conditions.

[0004] Therefore, there is an urgent need to propose a predictive control method for thermal power units that can significantly improve the operational economy of thermal power units in the AGC load tracking control process while meeting system constraints and grid regulation requirements, and enhance the flexible regulation and support capabilities of thermal power units in the new power system. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a cost-optimized AGC predictive control method and device for thermal power units, which can significantly improve the operational economy of thermal power units in the AGC load tracking control process and enhance the flexible adjustment and response support capabilities of thermal power units.

[0006] In a first aspect, the present invention provides a cost-optimized AGC predictive control method for thermal power units, comprising: Obtain actual operating data and AGC load commands from thermal power units, and provide a pre-built dynamic model of the units; Determine the steady-state operation objective function of the thermal power unit from the current state to reach a steady state, taking into account cost indicators; determine the dynamic operation objective function of the thermal power unit from the current state to reach a steady state through continuous control, taking into account cost indicators. Based on the steady-state operation objective function, and combined with AGC load commands and unit dynamic models, the optimal steady-state parameters of thermal power units are given; Based on the dynamic operating objective function, and combined with the optimal steady-state parameters, AGC load commands, actual operating data and unit dynamic model, the control parameters of the thermal power unit are given, and the thermal power unit is controlled.

[0007] Furthermore, the pre-construction of the unit dynamic model includes: Determine the state variables, controlled variables, and control variables of different units of the thermal power unit; The energy-mass conversion mechanism under the operating state of thermal power units is analyzed, and the state variables, controlled variables and control variables of different units of thermal power units are combined to construct the controlled variable function and the change function of state variables that are interrelated to different units. The system acquires operating data of thermal power units under different load conditions, identifies the controlled and variable functions that are interrelated between different units, and provides a dynamic model of the unit.

[0008] Furthermore, the thermal power units are supercritical coal-fired power units.

[0009] Furthermore, the state variables include a first state variable, a second state variable, and a third state variable; the controlled variables include a first controlled variable, a second controlled variable, and a third controlled variable; the control variables include a first control variable, a second control variable, and a third control variable; the controlled variable functions include a first controlled variable function, a second controlled variable function, and a third controlled variable function; and the change variable functions include a first change variable function, a second change variable function, and a third change variable function. The unit dynamic model satisfies the following relationship:

[0010] In the formula, u1, u2, and u3 are the first control variable, the second control variable, and the third control variable, respectively; y1, y2, and y3 are the first controlled variable, the second controlled variable, and the third controlled variable, respectively; and x1, x2, and x3 are the first state variable, the second state variable, and the third state variable, respectively. Let x1, x2, and x3 be the changes respectively, F1(·) be the first change function, F2(·) be the second change function, F3(·) be the third change function, H1(·) be the first controlled variable function, H2(·) be the second controlled variable function, and H3(·) be the third controlled variable function.

[0011] Furthermore, the construction of the controlled variable functions that are interconnected among different units includes: The first variable function is determined by using the physical parameters of the pulverizing system of the thermal power unit, combined with the amount of pulverized coal, fuel and main steam entering the thermal power unit; The second variable function is determined by using the first set of physical parameters of the thermal power unit, the amount of pulverized coal entering the boiler, the feedwater flow rate, the steam valve opening, and the main steam pressure, combined with the functional relationship between the main steam flow rate and the main steam pressure, and the functional relationship between the main steam specific enthalpy and the main steam pressure. The first set of physical parameters of the thermal power unit includes the gain of heat absorption in the boiler superheater section, the inertia coefficient of the working fluid pressure change in the separator, the feedwater specific enthalpy, the energy distribution dynamic coefficient of the thermal power unit, and the ratio of the superheater outlet steam specific enthalpy to the separator steam specific enthalpy. The third variable function is determined by considering the second set of physical parameters of the thermal power unit, the amount of pulverized coal entering the boiler, the feedwater flow rate, the steam valve opening, and the main steam pressure, combined with the functional relationship between the main steam flow rate and the main steam pressure, as well as the functional relationship between the main steam specific enthalpy and the main steam pressure. The second set of physical parameters of the thermal power unit includes the gain of heat absorption in the boiler superheater section, the inertia coefficient of the change in the enthalpy of the separator working fluid, the feedwater specific enthalpy, the dynamic coefficient of energy distribution of the thermal power unit, and the ratio of the specific enthalpy of the superheater outlet steam to the specific enthalpy of the separator steam. The first controlled variable function is determined by the separator steam pressure of the thermal power unit and the functional relationship between the superheater pipeline pressure difference and the separator steam pressure of the thermal power unit. The second controlled variable function is determined by the specific enthalpy of the separator steam in the thermal power unit; The third controlled variable function is determined by using the third physical parameter set of the thermal power unit, the main steam valve opening and the main steam pressure, and combining the functional relationship between the main steam flow rate and the main steam pressure of the thermal power unit, as well as the functional relationship between the main steam specific enthalpy and the main steam pressure of the thermal power unit; the third physical parameter set of the thermal power unit includes the turbine gain and feedwater specific enthalpy of the thermal power unit.

[0012] Furthermore, the unit dynamic model satisfies the following relationship:

[0013] In the formula, u is the control variable, u1, u2, and u3 are the first control variable, the second control variable, and the third control variable, respectively, and u = [u1 u2 u3]. T =[μ B D fw μ t ] T T is the transpose, μ B The fuel quantity of the thermal power unit is the first control quantity, D. fwThis indicates that the feedwater flow rate of the thermal power unit is the second control variable, μ. t The opening degree of the main steam valve of the thermal power unit is the third control variable; y is the controlled variable, and y1, y2, and y3 are the first, second, and third controlled variables, respectively, y = [y1 y2 y3] T =[P t h m Ne] T P t This indicates that the main steam pressure of the thermal power unit is the first controlled variable, h. m Let represent the enthalpy of the separator steam in the thermal power unit as the second controlled variable and the third state variable, and Ne represent the power generation as the third controlled variable; x is a state variable, where x1, x2, and x3 are the first, second, and third state variables, respectively, and x = [x1 x2 x3]. T =[r B P m h m ] T , Let r be the derivative of x1, x2, and x3, representing the change in x1, x2, and x3. B This indicates that the amount of pulverized coal entering the boiler of the thermal power unit is the first state quantity, P. m The separator steam pressure of the thermal power unit is represented as a second state quantity; h fw This indicates the specific enthalpy of the feedwater in a thermal power unit. s For the Laplace operator; τ The pure delay time of the pulverizing system in the thermal power unit is given by ; e is the natural constant. c 0 represents the inertia coefficient of the pulverizing system of the thermal power unit; k 0 represents the heat absorption gain of the boiler superheater section in a thermal power unit; c 1 and c 2 represents the inertia coefficients for the changes in working fluid pressure and enthalpy in the separator of the thermal power unit, respectively; d 1 and d 2 represents the dynamic coefficient of energy distribution in thermal power units; l It is the ratio of the specific enthalpy of the superheater outlet steam to the specific enthalpy of the separator steam in a thermal power unit; k 2 represents the turbine gain of the thermal power unit; f (·) represents the functional relationship between the main steam flow rate and the main steam pressure of the thermal power unit; g (·) represents the functional relationship between the superheater pipe pressure difference and the separator steam pressure of a thermal power unit; h (·) represents the functional relationship between the main steam specific enthalpy and the main steam pressure of a thermal power unit.

[0014] Furthermore, the operational data includes steady-state operational data and dynamic operational data; the model parameters of the unit's dynamic model include static parameters, dynamic parameters, and operational functions; Identify the controlled variable functions and variable functions that are interrelated between different units, and provide a dynamic model of the unit, including: Determine the steady-state model of the thermal power unit with respect to static parameters under steady-state operation; Based on the steady-state operating data of thermal power units under different load conditions, and combined with the steady-state model of thermal power units under operating conditions, static parameters are determined. The steady-state operating data of different units of thermal power units under different load conditions were fitted by regression analysis to determine the operating function; Dynamic operating data where the load and main steam pressure of thermal power units fluctuate beyond the threshold are selected from operating data under different load conditions. Based on static parameters, running functions, and dynamic running data, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, the particle swarm optimization algorithm is used to solve and determine the dynamic parameters. Based on static parameters, operating functions, and dynamic parameters, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, a dynamic model of the unit is given.

[0015] Furthermore, the steady-state model of the thermal power unit with respect to static parameters under steady-state operation is determined, including: The first sub-steady-state function is determined by the matching relationship between the first control quantity and the first state quantity under steady-state operation. The second sub-steady-state function is determined using the second control quantity, the third state quantity, the ratio of the superheater outlet steam enthalpy to the separator steam enthalpy, the main steam flow rate, the feedwater enthalpy, and the main steam enthalpy under steady-state operation of the thermal power unit. This includes: determining the heat power carried away from the boiler by the steam using the second control quantity, the third state quantity, and the ratio of the superheater outlet steam enthalpy to the separator steam enthalpy under steady-state operation; determining the energy loss rate using the main steam flow rate, the second control quantity, and the feedwater enthalpy under steady-state operation; determining the heat power carried by the main steam using the main steam flow rate and the main steam enthalpy under steady-state operation; and determining the second sub-steady-state function based on the balance between the heat power carried away from the boiler by the steam, the energy loss rate, and the heat power carried by the main steam. The third sub-steady-state function is determined by considering the feedwater flow rate, feedwater specific enthalpy, amount of pulverized coal entering the boiler, heat gain of the boiler superheater section, second control variable, third state variable, and the ratio of superheater outlet steam specific enthalpy to separator steam specific enthalpy under steady-state operation of the thermal power unit. This includes: determining the total energy carried by the feedwater fed into the boiler using the feedwater flow rate and feedwater specific enthalpy under steady-state operation; determining the total heat flow absorbed by the boiler superheater section using the amount of pulverized coal entering the boiler and heat gain of the boiler superheater section under steady-state operation; and determining the third sub-steady-state function based on the balance between the total energy carried by the feedwater, the heat power carried away by the steam from the boiler, and the total heat flow absorbed by the boiler superheater section. The output power of the thermal power unit is determined by the main steam specific enthalpy, feedwater specific enthalpy, main steam flow rate, and turbine gain under steady-state operation. The fourth sub-steady-state function is determined based on the balance between the output power and the generated power of the thermal power unit. By integrating the first, second, third, and fourth sub-steady-state functions, a steady-state model of a thermal power unit under operating conditions is presented.

[0016] Furthermore, the steady-state model of a thermal power unit under operating conditions satisfies the following relationship:

[0017] In the formula, * represents the steady-state value, u1 and u2 are the first and second control variables, respectively, x1 and x3 are the first and third state variables, respectively, l is the ratio of the specific enthalpy of the superheater outlet steam to the specific enthalpy of the separator steam in the thermal power unit, and D t Indicates the main steam flow rate, h fw h represents the feedwater specific enthalpy of a thermal power unit. t This indicates the specific enthalpy of the main steam. k 0 represents the heat absorption gain of the boiler superheater section in a thermal power unit. k 2 represents the turbine gain of the thermal power unit, and Ne represents the power generation capacity.

[0018] Furthermore, the determination of the steady-state operating objective function for thermal power units, taking into account cost indicators, includes: The constraints of thermal power units under steady-state operation are determined, and the objective function for steady-state operation is given with the goal of minimizing the operating costs generated by controlling the thermal power units to reach steady state.

[0019] Furthermore, the determination of the steady-state operating objective function for thermal power units, taking into account cost indicators, includes: Determine the upper and lower limits of each control variable and each controlled variable, and combine the matching relationship between the AGC load command and the third controlled variable, the change function and controlled variable function of the thermal power unit under steady-state operation, and determine the constraints of steady-state operation. The fuel consumption cost coefficient, water supply cost coefficient, and cost coefficient related to throttling loss are obtained, and then multiplied with the corresponding control variables and integrated to determine the steady-state operating cost. Based on the constraints of steady-state operation, and with the goal of minimizing the operating costs incurred in controlling the thermal power unit to reach steady state, the objective function for steady-state operation is determined, satisfying the following relationship:

[0020] In the formula, L represents the operating cost. s The parameters represent the steady-state values ​​of the parameters under cost-optimal and steady-state operation. Parameters refer to state variables, controlled variables, and control variables; α1 represents the fuel consumption cost coefficient; α2 represents the water supply consumption cost coefficient; α3 represents the cost coefficient related to throttling losses; F(·) is the variation function of different units of the thermal power unit with respect to state variables, including the first variation function, the second variation function, and the third variation function; H(·) is the control variable function of different units of the thermal power unit, including the first control variable function, the second control variable function, and the third control variable function; y3 is the third control variable, y 3r Indicates AGC load command; y max and y min These represent the upper and lower limits of the controlled variable y, respectively; u max and u min These represent the upper and lower limits of the control quantity u, respectively. Furthermore, the determination of the dynamic operating objective function for thermal power units, taking into account cost indicators, includes: Based on the operating costs of controlling thermal power units, the benefits of AGC regulation, and the real-time control coefficients for achieving steady state, the control cost of controlling a thermal power unit in a single operation is determined. The constraints and sampling step size of the thermal power unit under dynamic operation are determined, and the objective function for dynamic operation is determined with the goal of minimizing the total control cost of controlling the thermal power unit within the step size.

[0021] Furthermore, the determination of the dynamic operating objective function for thermal power units, taking into account cost indicators, includes: Determine the upper and lower limits of the change of each control variable, and combine the upper and lower limits of each control variable, the upper and lower limits of each controlled variable, the current state variable measurement value, the controlled variable function of the thermal power unit in the current operating state, and the relationship between the change function and the state variable at the next moment to determine the constraints of dynamic operation. Determine the deviation between the power generation and the AGC load command, the corresponding adjustment accuracy coefficient, the AGC adjustment compensation benefit coefficient, and the AGC adjustment benefit; superimpose the magnitudes of fuel cost, water supply cost, throttling loss cost, and AGC adjustment benefit, and add a regularization term to determine the total control cost of controlling the thermal power unit within the step size. Based on the constraints of dynamic operation, and with the objective of minimizing the total control cost of controlling the thermal power unit within a step size, the dynamic operation objective function is determined, satisfying the following relationship:

[0022] In the formula, N p α represents the sampling step size of the thermal power unit; α4 represents the AGC regulation compensation benefit coefficient; D(t) represents the relationship between y3 and y4 at time t. 3r The deviation, D(t)=|y3(t)-y 3r (t)|;K(t) represents the adjustment accuracy coefficient, K(t)=0.2-|y3(t)-y 3r (t)| / (1%P GN ), P GN The rated power; b1 and b2 represent regularization coefficients. For L2 regularization; x(t), y(t), and u(t) represent the state variable, controlled variable, and control variable at time t, respectively; t k Indicates the current time; x m (t k ) represents the state variable x at the current time. t k The measured value; Δu is the change in the control variable u, Δu(t) = u(t) - u(t-1), Δu max and Δu min These represent the upper and lower limits of Δu, respectively.

[0023] Furthermore, based on the steady-state operation objective function and combined with AGC load commands and the unit dynamic model, the optimal steady-state parameters of the thermal power unit are given, including: The AGC load command is input into the unit's dynamic model, and based on the steady-state operation objective function, it is solved by a numerical optimization solver to obtain the control and state variables when the thermal power unit executes the AGC load command to achieve steady-state operation and minimize costs.

[0024] Furthermore, based on the dynamic operating objective function, and combined with the optimal steady-state parameters, AGC load commands, actual operating data, and the unit's dynamic model, the control parameters of the thermal power unit are given, including: The control and state variables of the thermal power unit when executing AGC load commands to achieve steady-state operation and minimize costs are input into the dynamic operation objective function. The AGC load commands and actual operating data are input into the unit dynamic model, and the numerical optimization solver is used to solve the problem to obtain the control variables of each control within the sampling step of the thermal power unit.

[0025] Secondly, the present invention also provides a cost-optimized AGC predictive control device for thermal power units, employing the above-mentioned predictive control method, the device comprising: The initial condition determination module is used to acquire the actual operating data and AGC load commands of the thermal power unit and provide a pre-built dynamic model of the unit. The objective function determination module is used to determine the steady-state operation objective function of the thermal power unit from the current state to the steady state, taking into account cost indicators, and to determine the dynamic operation objective function of the thermal power unit from the current state to the steady state through continuous control, taking into account cost indicators. The steady-state parameter determination module is used to determine the optimal steady-state parameters of the thermal power unit based on the steady-state operation objective function and in combination with AGC load commands and the unit dynamic model. The parameter determination control module is used to determine the control parameters of the thermal power unit based on the dynamic operating objective function, combined with the optimal steady-state parameters, AGC load commands, actual operating data and unit dynamic model, and to control the thermal power unit.

[0026] Furthermore, the initial condition determination module is used for: Determine the state variables, controlled variables, and control variables of different units of the thermal power unit; The energy-mass conversion mechanism under the operating state of thermal power units is analyzed, and the state variables, controlled variables and control variables of different units of thermal power units are combined to construct the controlled variable function and the change function of state variables that are interrelated to different units. The system acquires operating data of thermal power units under different load conditions, identifies the controlled and variable functions that are interrelated between different units, and provides a dynamic model of the unit.

[0027] Furthermore, the initial condition determination module is also used for: Determine the steady-state model of the thermal power unit with respect to static parameters under steady-state operation; Based on the steady-state operating data of thermal power units under different load conditions, and combined with the steady-state model of thermal power units under operating conditions, static parameters are determined. The steady-state operating data of different units of thermal power units under different load conditions were fitted by regression analysis to determine the operating function; Dynamic operating data where the load and main steam pressure of thermal power units fluctuate beyond the threshold are selected from operating data under different load conditions. Based on static parameters, running functions, and dynamic running data, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, the particle swarm optimization algorithm is used to solve and determine the dynamic parameters. Based on static parameters, operating functions, and dynamic parameters, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, a dynamic model of the unit is given.

[0028] Furthermore, the objective function determination module is used for: The constraints of thermal power units under steady-state operation are determined, and the objective function for steady-state operation is given with the goal of minimizing the operating costs generated by controlling the thermal power units to reach steady state.

[0029] Furthermore, the objective function determination module is also used for: Based on the operating costs of controlling thermal power units, the benefits of AGC regulation, and the real-time control coefficients for achieving steady state, the control cost of controlling a thermal power unit in a single operation is determined. The constraints and sampling step size of the thermal power unit under dynamic operation are determined, and the objective function for dynamic operation is determined with the goal of minimizing the total control cost of controlling the thermal power unit within the step size.

[0030] Furthermore, the steady-state parameter determination module is also used for: The AGC load command is input into the unit's dynamic model, and based on the steady-state operation objective function, it is solved by a numerical optimization solver to obtain the control and state variables when the thermal power unit executes the AGC load command to achieve steady-state operation and minimize costs.

[0031] Furthermore, the parameter determination control module is also used for, including: The control and state variables of the thermal power unit when executing AGC load commands to achieve steady-state operation and minimize costs are input into the dynamic operation objective function. The AGC load commands and actual operating data are input into the unit dynamic model, and the numerical optimization solver is used to solve the problem to obtain the control variables of each control within the sampling step of the thermal power unit.

[0032] The present invention provides a cost-optimized AGC predictive control method and device for thermal power units, which has at least the following beneficial effects: (1) By using the dynamic model of the unit and the dynamic operation objective function that continuously controls the process from the current state to the steady state and takes into account the cost indicators, the operating economy of the thermal power unit in the AGC load tracking control process can be significantly improved under the premise of meeting the operating constraints and regulation performance of the thermal power unit, and the flexible regulation and response support capability of the thermal power unit in the new power system can be enhanced. Attached Figure Description

[0033] Figure 1A flowchart of a cost-optimized AGC predictive control method for thermal power units provided by the present invention; Figure 2 This is a schematic diagram of a thermal power unit control method provided in a certain embodiment of the present invention; Figure 3 A flowchart for constructing a dynamic model of a generating unit provided in one embodiment of the present invention; Figure 4 A flowchart illustrating a dynamic model of a generating unit is provided as an embodiment of the present invention; Figure 5 This is a schematic diagram of a cost-optimized AGC predictive control device for thermal power units provided by the present invention. Detailed Implementation

[0034] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0035] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0036] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0037] like Figure 1 As shown, this invention provides a cost-optimized AGC predictive control method for thermal power units, comprising: Obtain actual operating data and AGC load commands from thermal power units, and provide a pre-built dynamic model of the units; Determine the steady-state operation objective function of the thermal power unit from the current state to reach a steady state, taking into account cost indicators; determine the dynamic operation objective function of the thermal power unit from the current state to reach a steady state through continuous control, taking into account cost indicators. Based on the steady-state operation objective function, and combined with AGC load commands and unit dynamic model, the optimal steady-state parameters of the thermal power unit are given; the optimal steady-state parameters may include state variables and control variables; Based on the dynamic operating objective function, and combined with the optimal steady-state parameters, AGC load commands, actual operating data and unit dynamic model, the control parameters of the thermal power unit are given, and the thermal power unit is controlled.

[0038] The thermal power units are supercritical coal-fired power generating units. Specifically, a supercritical coal-fired power generating unit includes a pulverizing system, boiler, turbine, high-pressure cylinder, intermediate-pressure cylinder, low-pressure cylinder, condenser, feedwater pump, feedwater heater, etc. The boiler includes a furnace, pulverizer, steam-water separator, superheater, spray desuperheater, reheater, and economizer. Fuel (raw coal) is pulverized by the pulverizer and then fed into the furnace for combustion, where the released chemical energy is converted into thermal energy. The feedwater in the boiler absorbs the heat released by combustion and becomes high-temperature, high-pressure steam, realizing the conversion from thermal energy to the internal energy of the steam working fluid. The steam enters the turbine and drives the rotor to rotate, further converting the thermal energy of the working fluid into mechanical energy. Finally, the turbine drives the generator to operate, converting mechanical energy into electrical energy and outputting it to the power grid.

[0039] To improve the economic efficiency of thermal power units (supercritical coal-fired power units) during load changes and enhance their flexible adjustment and support capabilities in new power systems, this invention proposes a predictive control method implemented through a control structure. Predictive control is a control method in the field of thermal power unit control. Its core is to generate control commands by solving the control problem at each sampling time, using the current state as the initial condition for optimization, and implementing only the optimal sequence of control actions. The control structure implements control through a model predictive controller for the thermal power unit, such as... Figure 2 As shown, the model predictive controller of the thermal power unit adopts the aforementioned predictive control method. Specifically, the model controller uses fuel quantity commands, feedwater flow rate, and main steam valve opening to control main steam pressure, separator steam specific enthalpy, and power generation.

[0040] In practical application scenarios, such as Figure 3 As shown, this invention selects key variables as control variables, controlled variables, and state variables, and establishes a control-oriented dynamic model of the unit by combining the process mechanism of the thermal power unit; the pre-construction of the unit dynamic model includes: Determine the state variables, controlled variables, and control variables of different units of the thermal power unit; The energy-mass conversion mechanism under the operating state of thermal power units is analyzed, and the state variables, controlled variables and control variables of different units of thermal power units are combined to construct the controlled variable function and the change function of state variables that are interrelated to different units. The system acquires operating data of thermal power units under different load conditions, identifies the controlled and variable functions that are interrelated between different units, and provides a dynamic model of the unit.

[0041] Furthermore, the state variables include a first state variable, a second state variable, and a third state variable; the controlled variables include a first controlled variable, a second controlled variable, and a third controlled variable; the control variables include a first control variable, a second control variable, and a third control variable; the controlled variable functions include a first controlled variable function, a second controlled variable function, and a third controlled variable function; and the change variable functions include a first change variable function, a second change variable function, and a third change variable function. The unit dynamic model satisfies the following relationship:

[0042] In the formula, u1, u2, and u3 are the first control variable, the second control variable, and the third control variable, respectively; y1, y2, and y3 are the first controlled variable, the second controlled variable, and the third controlled variable, respectively; and x1, x2, and x3 are the first state variable, the second state variable, and the third state variable, respectively. Let x1, x2, and x3 be the changes respectively, F1(·) be the first change function, F2(·) be the second change function, F3(·) be the third change function, H1(·) be the first controlled variable function, H2(·) be the second controlled variable function, and H3(·) be the third controlled variable function.

[0043] The construction of the controlled variable functions that are interconnected among different units can include: The first variable function is determined by using the physical parameters of the pulverizing system of the thermal power unit, combined with the amount of pulverized coal, fuel and main steam entering the thermal power unit; The second variable function is determined by using the first set of physical parameters of the thermal power unit, the amount of pulverized coal entering the boiler, the feedwater flow rate, the steam valve opening, and the main steam pressure, combined with the functional relationship between the main steam flow rate and the main steam pressure, and the functional relationship between the main steam specific enthalpy and the main steam pressure. The first set of physical parameters of the thermal power unit includes the gain of heat absorption in the boiler superheater section, the inertia coefficient of the working fluid pressure change in the separator, the feedwater specific enthalpy, the energy distribution dynamic coefficient of the thermal power unit, and the ratio of the superheater outlet steam specific enthalpy to the separator steam specific enthalpy. The third variable function is determined by considering the second set of physical parameters of the thermal power unit, the amount of pulverized coal entering the boiler, the feedwater flow rate, the steam valve opening, and the main steam pressure, combined with the functional relationship between the main steam flow rate and the main steam pressure, as well as the functional relationship between the main steam specific enthalpy and the main steam pressure. The second set of physical parameters of the thermal power unit includes the gain of heat absorption in the boiler superheater section, the inertia coefficient of the change in the enthalpy of the separator working fluid, the feedwater specific enthalpy, the dynamic coefficient of energy distribution of the thermal power unit, and the ratio of the specific enthalpy of the superheater outlet steam to the specific enthalpy of the separator steam. The first controlled variable function is determined by the separator steam pressure of the thermal power unit and the functional relationship between the superheater pipeline pressure difference and the separator steam pressure of the thermal power unit. The second controlled variable function is determined by the specific enthalpy of the separator steam in the thermal power unit; The third controlled variable function is determined by using the third physical parameter set of the thermal power unit, the main steam valve opening and the main steam pressure, and combining the functional relationship between the main steam flow rate and the main steam pressure of the thermal power unit, as well as the functional relationship between the main steam specific enthalpy and the main steam pressure of the thermal power unit; the third physical parameter set of the thermal power unit includes the turbine gain and feedwater specific enthalpy of the thermal power unit.

[0044] Specifically, the control variables can include fuel quantity commands (i.e., the fuel quantity corresponding to the thermal power unit), feedwater flow rate, and main steam valve opening; the controlled variables include main steam pressure, separator steam specific enthalpy, and power generation; and the state variables include the amount of pulverized coal entering the boiler, separator steam pressure, and separator steam specific enthalpy. Based on the control variables, controlled variables, and state variables, a dynamic model of the unit is constructed, satisfying the following relationships:

[0045] In the formula, u is the control variable, u1, u2, and u3 are the first control variable, the second control variable, and the third control variable, respectively, and u = [u1 u2 u3]. T =[μ B D fw μ t ] T T is the transpose, μ B D represents the fuel quantity of a thermal power unit. fw The feedwater flow rate of a thermal power unit is expressed in μ. t This represents the opening degree of the main steam valve of the thermal power unit; y is the controlled variable, where y1, y2, and y3 are the first, second, and third controlled variables, respectively, and y = [y1 y2 y3]. T =[P t h m Ne] T P t This indicates that the main steam pressure of the thermal power unit is the first controlled variable, h. mLet represent the enthalpy of the separator steam in the thermal power unit as the second controlled variable and the third state variable, and Ne represent the power generation as the third controlled variable; x is a state variable, where x1, x2, and x3 are the first, second, and third state variables, respectively, and x = [x1 x2 x3]. T =[r B P m h m ] T , Let r be the derivative of x1, x2, and x3, representing the change in x1, x2, and x3. B P represents the amount of pulverized coal entering the boiler of a thermal power unit. m The separator steam pressure of the thermal power unit is represented as a second state quantity; h fw This indicates the specific enthalpy of the feedwater in a thermal power unit. s For the Laplace operator; τ The pure delay time of the pulverizing system in the thermal power unit is given by ; e is the natural constant. c 0 represents the inertia coefficient of the pulverizing system of the thermal power unit; k 0 represents the heat absorption gain of the boiler superheater section in a thermal power unit; c 1 and c 2 represents the inertia coefficients for the changes in working fluid pressure and enthalpy in the separator of the thermal power unit, respectively; d 1 and d 2 represents the dynamic coefficient of energy distribution of thermal power units, which characterizes the coupling relationship between separator pressure and enthalpy changes; l It is the ratio of the specific enthalpy of the superheater outlet steam to the specific enthalpy of the separator steam in a thermal power unit; k 2 represents the turbine gain of the thermal power unit; f (·) represents the functional relationship between the main steam flow rate and the main steam pressure of the thermal power unit; g (·) represents the functional relationship between the superheater pipe pressure difference and the separator steam pressure of a thermal power unit; h (·) represents the functional relationship between the main steam specific enthalpy and the main steam pressure of a thermal power unit.

[0046] Operating data for thermal power units can include steady-state operating data and dynamic operating data; the model parameters of the unit's dynamic model can include static parameters, dynamic parameters, and operating functions. For example... Figure 4 As shown, the controlled variable functions and variable functions that are interrelated between different units are identified, and a dynamic model of the unit is given, which may include: Determine the steady-state model of the thermal power unit with respect to static parameters under steady-state operation; Based on the steady-state operating data of thermal power units under different load conditions, and combined with the steady-state model of thermal power units under operating conditions, static parameters are determined. The steady-state operating data of different units of thermal power units under different load conditions were fitted by regression analysis to determine the operating function; Dynamic operating data where the load and main steam pressure of thermal power units exceed thresholds are selected from operating data under different load conditions. The thresholds for load fluctuation of thermal power units include the load change amplitude threshold and the change rate threshold. The load change amplitude threshold is 200MW, and the change rate threshold is 5MW / min. The threshold for main steam pressure fluctuation can be determined according to the specific application scenario. Based on static parameters, running functions, and dynamic running data, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, the particle swarm optimization algorithm is used to solve and determine the dynamic parameters. Based on static parameters, operating functions, and dynamic parameters, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, a dynamic model of the unit is given.

[0047] Specifically, for the identification of static parameters k0, k2, and l, steady-state operating data of multiple thermal power units under different load conditions can be selected, and the parameter values ​​of static parameters k0, k2, and l can be solved by combining them with a steady-state model. Specifically, the steady-state model of a thermal power unit under operating conditions satisfies the following relationship:

[0048] In the formula, * represents the steady-state value, u1 and u2 are the first and second control variables, respectively, x1 and x3 are the first and third state variables, respectively, l is the ratio of the specific enthalpy of the superheater outlet steam to the specific enthalpy of the separator steam in the thermal power unit, and D t Indicates the main steam flow rate, h fw h represents the feedwater specific enthalpy of a thermal power unit. t This indicates the specific enthalpy of the main steam. k 0 represents the heat absorption gain of the boiler superheater section in a thermal power unit. k 2 represents the turbine gain of the thermal power unit, and Ne represents the power generation capacity.

[0049] To identify the undetermined functions (operating functions) f(·), g(·), and h(·), steady-state operating data of multiple thermal power units under different load conditions can be selected, and regression analysis can be used to fit the expression of the undetermined functions. Specifically, for the operating function f(·), data of the thermal power units at multiple different stable load points within the sliding pressure operating load range (usually 50% to 90% of rated load) are collected. Each data point should include: main steam flow rate (D) and main steam pressure (P); the operating function f(·) is obtained by fitting the data points using a polynomial regression model. Specifically, multiple steady-state operating condition (D, P) data pairs are selected from the historical database; main steam flow rate is used as the independent variable (input), and main steam pressure is used as the dependent variable (output); a quadratic polynomial is fitted using the least squares method to obtain the constant term coefficients, linear term coefficients, and quadratic term coefficients; among them, the constant term coefficients are used to set the base pressure, and the linear and quadratic term coefficients together describe the nonlinear relationship between main steam pressure and main steam flow rate; the fitting score and residual plot of the fitted quadratic polynomial are determined and compared with preset conditions to give a quadratic polynomial that meets the requirements. The fitting score is determined by the sum of the squares of the differences between the predicted and actual observed values, and the sum of the squares of the differences between the actual observed values ​​and their average values. For example, the fitting score R = 1 - (ss... res / ss tot ), ss res ss is the sum of squares of the differences between predicted and observed values. tot It is the sum of squares of the differences between the actual observed values ​​and their average values.

[0050] For the operating function g(·), data is collected from the unit at multiple different stable load points. Each data point should include: superheater differential pressure (ΔP), separator pressure (P...). sep) And the main steam flow rate (D). Then, it is fitted using multiple linear / nonlinear regression. Specifically, the three-dimensional data set of the operating function g(·) is collected, including the superheater pressure difference, separator pressure, and main steam flow rate; the natural logarithm is taken for each type of data in the three-dimensional data set to obtain the new variables ln(ΔP), ln(D), and ln(P) for each type of data. sep Using the new variables of separator pressure and main steam flow rate as independent variables and the new variable of superheater pressure difference as dependent variable, a multiple linear regression is performed to obtain a multiple linear regression model, for example, ln(ΔP)=C+Aln(D)+Bln(P) sep The multiple linear regression model was compared with the nonlinear mechanism model of the thermal power unit regarding the superheater pipe pressure difference and separator steam pressure to determine the operating function g(·). The nonlinear mechanism model of the thermal power unit regarding the superheater pipe pressure difference and separator steam pressure is: ln(ΔP) = ln(β0) + 2*ln(D) - ln(Psep), where... The function g(·) is: .

[0051] The operating function h(·) is determined based on thermodynamic principles through regression. This is primarily achieved by collecting data from all steady-state operating conditions of the thermal power unit. Each data point should include: main steam specific enthalpy (H), main steam pressure (P), and main steam temperature (T0). The operating function h(·) can be determined using a lookup table. Alternatively, it can be based on multivariate nonlinear regression, specifically fitting the relationship between main steam pressure and main steam temperature and main steam specific enthalpy. The linear model is: H = β1 + β2P + β3T0; or the polynomial model is: H = β1 + β2P + β3T0 + β4P. 2 +β5T0 2 +β6PT0, where β1, β2, β3, β4, β5, and β6 are polynomial coefficients that can be obtained by fitting actual data. In practical applications, the data triplets (main steam specific enthalpy, main steam pressure, and main steam temperature) of the running function h(·) are collected; with main steam pressure and main steam temperature as independent variables and main steam specific enthalpy as dependent variable, multiple linear regression is performed to obtain a linear model or a polynomial model as the running function h(·).

[0052] For dynamic parameters τ , c 0、 c 1. c 2. d 1 and d For identification in section 2, dynamic operating data segments with significant load and pressure fluctuations in thermal power units can be selected. Combined with the unit's dynamic model, the particle swarm optimization algorithm can be used to solve for the dynamic parameters. τ , c 0、 c 1. c 2. d 1 and dThe parameter values ​​are set to 2. Specifically, dynamic operating data segments with significant load and pressure fluctuations in thermal power units are selected as training data. These data contain the dynamic response characteristics of the system under different operating conditions and can effectively reflect the dynamic characteristics of the model. Secondly, the established unit dynamic model, static parameters identified through steady-state operating data, and regression expressions of undetermined functions are combined to form the initial framework for the particle swarm optimization algorithm. Next, the particle swarm optimization algorithm initializes a set of particles (each particle representing a candidate solution for a set of dynamic parameters) through an iterative optimization process and updates them based on the objective function of the dynamic parameters (minimizing the error between model predictions and actual measurements, such as the root mean square error). Particles gradually approach the optimal dynamic parameter values ​​by adjusting their velocity and position, combining the global best and individual best solutions, ultimately determining the best estimate of the dynamic parameters. Here, position is the particle's coordinate in the search space, directly encoded as a solution vector; velocity is the direction and step size of the particle's movement, also a vector with the same dimension as the solution vector.

[0053] In practical applications, the iterative optimization of the particle swarm optimization algorithm can include: Step S11: Input the static parameters identified through steady-state operation data and the regression expression of the undetermined function into the unit dynamic model; Step S12: Perform step disturbance or pseudo-random sequence (PRBS) tests on the thermal power unit to collect input and output dynamic data; or directly use historical input and output dynamic data of the thermal power unit under different operating conditions; further, the input and output dynamic data may include training set and test set.

[0054] Step S13: Initialize the particle swarm optimization algorithm, including: Step S131, Define particle encoding: The position of each particle in the swarm represents a candidate solution for a set of input-output dynamic parameters.

[0055] Step S132: Set the search space: Define a reasonable range of values ​​for each input and output dynamic parameter; this range can be determined based on prior engineering knowledge to prevent searching in physically unreasonable regions (such as negative time constants). Step S133: Initialize the particle swarm: Within the defined search space, a group of particles (e.g., 50) are randomly generated. The position of each particle is randomly initialized, and its velocity is initialized to 0 or a small random value. Step S14, iterative optimization using the particle swarm optimization algorithm, including: Step S141: Decode the position of each particle into a set of specific dynamic parameter values; substitute each set of dynamic parameter values ​​into the unit dynamic model in step S11 (the unit dynamic model already contains static parameters and undetermined functions); use the input data of the training set to drive the unit dynamic model (the unit dynamic model already contains static parameters and undetermined functions, and substitute the dynamic parameter values) to perform dynamic simulation and obtain the predicted output of driving the unit dynamic model; calculate the error between the predicted output and the actual measured output of the training set, which can be calculated by the fitness function and determined by the sum of mean squares (MSE) or root mean square error (RMSE); with the goal of minimizing the fitness value, find the dynamic parameters that minimize the difference between the predicted output of the unit dynamic model and the actual measured output.

[0056] Step S142: Compare the current fitness of each particle with its individual historical best; if it is better, update the individual historical best with the current position; compare the individual historical best of all particles with the global historical best, and select the training set with the best fitness (i.e. the training set with the smallest error) to update the global historical best.

[0057] Step S143: According to the velocity update formula of the particle swarm optimization algorithm (the conventional velocity update formula of the particle swarm optimization algorithm can be used), guide the particles to fly towards the direction of the individual historical best and the global historical best; at the same time, constrain the velocity and position based on the boundaries of velocity and position. Step S144: Repeat steps S141 to S143 until the termination condition is met (such as reaching the maximum number of iterations, the improvement of the global historical best being less than a certain threshold, or reaching a satisfactory error level). Step S15: After the iteration is completed, the global optimal position is the best dynamic parameter estimate found.

[0058] After outputting the optimal dynamic parameter estimates, the parameter set represented by the global optimal position can be substituted into the unit dynamic model; simulation is performed using test set data that was not used in the training, and the error between the predicted value and the actual measured value of the unit dynamic model is calculated; if the test set error is less than the preset threshold, the unit dynamic model meets the predetermined requirements, that is, the dynamic parameters are successfully identified; otherwise, the parameters of the particle swarm optimization algorithm are adjusted for iterative optimization.

[0059] The determination of the steady-state operation objective function of thermal power units considering cost indicators may include: determining the constraints of the thermal power units under steady-state operation, and giving the steady-state operation objective function with the goal of minimizing the operating costs generated by controlling the thermal power units to reach steady state.

[0060] Based on the steady-state operation objective function, and combined with AGC load commands and the unit dynamic model, the optimal steady-state parameters of the thermal power unit are given. This can include: inputting the AGC load command into the unit dynamic model, and solving it using a numerical optimization solver based on the steady-state operation objective function to obtain the control and state variables when the thermal power unit executes the AGC load command to achieve steady-state operation and minimize costs.

[0061] The determination of the steady-state operating objective function for thermal power units, taking into account cost indicators, may include: Determine the upper and lower limits of each control variable and each controlled variable, and combine the matching relationship between the AGC load command and the third controlled variable, the change function and controlled variable function of the thermal power unit under steady-state operation, and determine the constraints of steady-state operation. The fuel consumption cost coefficient, water supply cost coefficient, and cost coefficient related to throttling loss are obtained, and then multiplied with the corresponding control variables and integrated to determine the steady-state operating cost. Based on the constraints of steady-state operation, and with the goal of minimizing the operating costs incurred by controlling the thermal power unit to reach steady state, the steady-state operation objective function is determined. The steady-state operation objective function satisfies the following relationship:

[0062] In the formula, L represents the operating cost. s This represents the steady-state value of the parameters under cost-optimal and steady-state operation. Parameters refer to state variables, controlled variables, and control variables; that is, the subscript 's' indicates the steady-state value under cost-optimal and steady-state operation; α1 represents the fuel consumption cost coefficient; α2 represents the water supply consumption cost coefficient; α3 represents the cost coefficient related to throttling losses; F(·) is the function of the change in state variables for different units of the thermal power unit, including the first change function, the second change function, and the third change function; H(·) is the function of the controlled variables that are interconnected between different units of the thermal power unit, including the first controlled variable function, the second controlled variable function, and the third controlled variable function; y3 is the third controlled variable, y 3r This indicates the AGC load command, which is the setpoint of the third controlled variable y3; y max and y min These represent the upper and lower limits of the controlled variable y, respectively; u max and u min These represent the upper and lower limits of the control quantity u, respectively; where the cost index of the steady-state operation objective function is fuel cost + water supply cost + throttling loss cost; By setting control parameters for the model controller of thermal power units, the operational economy and safety of AGC load tracking control in thermal power units can be improved. Specifically, control parameters are set for the economic model predictive controller, including the prediction time domain, sampling time regularization coefficient, and constraint conditions. The determination of the dynamic operating objective function of the thermal power unit considering cost indicators may include: Based on the operating costs of controlling thermal power units, the benefits of AGC regulation, and the real-time control coefficients for achieving steady state, the control cost of controlling a thermal power unit in a single operation is determined. The constraints and sampling step size of the thermal power unit under dynamic operation are determined, and the objective function for dynamic operation is determined with the goal of minimizing the total control cost of controlling the thermal power unit within the step size.

[0063] Specifically, the determination of the dynamic operating objective function for thermal power units, taking into account cost indicators, includes: Determine the upper and lower limits of the change of each control variable, and combine the upper and lower limits of each control variable, the upper and lower limits of each controlled variable, the current state variable measurement value, the controlled variable function of the thermal power unit in the current operating state, and the relationship between the change function and the state variable at the next moment to determine the constraints of dynamic operation. Determine the deviation between power generation and AGC load command, the corresponding adjustment accuracy coefficient, the AGC adjustment compensation benefit coefficient, and the AGC adjustment benefit; superimpose the magnitudes of fuel cost, water supply cost, throttling loss cost, and AGC adjustment benefit, and add a regularization term to determine the total control cost of controlling the thermal power unit within the step size; wherein, the regularization term is the superposition of the regularization result of the steady-state operation control quantity and the current control quantity, and the regularization result of the steady-state operation state quantity and the current moment state quantity; Based on the constraints of dynamic operation, and with the objective of minimizing the total control cost of controlling the thermal power unit within a step size, the dynamic operation objective function is determined, satisfying the following relationship:

[0064] In the formula, N p α represents the sampling step size of the thermal power unit; α4 represents the AGC regulation compensation benefit coefficient; D(t) represents the relationship between y3 and y4 at time t. 3r The deviation, D(t)=|y3(t)-y 3r (t)|;K(t) represents the adjustment accuracy coefficient, K(t)=0.2-|y3(t)-y 3r (t)| / (1%P GN ), y3 and y 3r The smaller the deviation, the higher the adjustment accuracy, P GN b1 and b2 represent the rated power; b1 and b2 are regularization coefficients, which are hyperparameters used to ensure the smoothness of the control process. For L2 regularization; x(t), y(t), and u(t) represent the state variable, controlled variable, and control variable at time t, respectively; t k Indicates the current time; x m (t k ) represents the state variable x at the current time. t k The measured value; Δu is the change in the control variable u, Δu(t) = u(t) - u(t-1), Δu max and Δu min These represent the upper and lower limits of Δu, respectively.

[0065] The core objective of the dynamic operation objective function is to minimize the operating cost, i.e., to minimize the total cost within the data acquisition time domain. The first four terms are the actual cost items (i.e., cost indicators, specifically fuel cost + water supply cost + throttling loss cost - AGC adjustment revenue). Specifically, fuel cost is the product of the fuel consumption cost coefficient and the first control variable; water supply cost is the product of the water supply consumption cost coefficient and the second control variable; and throttling loss cost is the product of the cost coefficient related to throttling loss and the third control variable. The last two terms are stability guarantees for the control process (regularization terms) to ensure the smoothness of the control process. The regularization terms for the last two terms are determined by regularizing the steady-state control variable and the current control variable, as well as the steady-state state variable and the current state variable, respectively. This effectively achieves stable control while reducing the complexity of the control itself.

[0066] The thermal power unit model controller uses fuel quantity commands, feedwater flow rates, and main steam valve openings to control main steam pressure, separator steam specific enthalpy, and power generation; that is, the thermal power unit model controller determines the main steam pressure, separator steam specific enthalpy, and power generation that need to be controlled. Specifically, based on the dynamic operating objective function, and combined with optimal steady-state parameters, AGC load commands, actual operating data, and the unit's dynamic model, the control parameters of the thermal power unit are given, including: The control and state variables of the thermal power unit when executing AGC load commands to achieve steady-state operation and minimize costs are input into the dynamic operation objective function. The AGC load commands and actual operating data are input into the unit dynamic model, and the numerical optimization solver is used to solve the problem to obtain the control variables of each control within the sampling step of the thermal power unit.

[0067] The control principle of this invention is as follows: First, model initialization is performed at the current time. t k Acquire real-time measurements of state variables and controlled variables, as well as the economically optimal steady-state operating point obtained during the steady-state optimization process. x s and us Based on the constructed dynamic objective function and its constraints, the numerical optimization solver is called to solve the above problem online to obtain the optimal control quantity, which is then sent to the actuator of the thermal power unit to perform the corresponding control. When the next control cycle arrives, the above steps are repeated for prediction and optimization, thereby realizing closed-loop feedback control.

[0068] The core of this invention's prediction lies in the rolling optimization mechanism of the dynamic objective function, where the current state quantity is measured at each sampling time. x m ( t k Using these as initial conditions, the future is determined based on the established unit dynamic model. N p The dynamic behavior within a sampling step is analyzed, and the control sequence (control input) is optimized across the entire sampling time domain. The dynamic objective function characterizes the prediction and optimization of dynamic economy and thermal power unit behavior at multiple future moments within the sampling step. Although only the optimal control input at the current moment is applied, the control decision is based on the prediction and optimization of future states, thus achieving proactive decision-making and dynamic economic optimization. Furthermore, this invention periodically utilizes real-time operating data to update dynamic parameters online to adapt to changes in the characteristics of the thermal power unit and maintain its long-term control accuracy.

[0069] Furthermore, conventional control methods employ a hierarchical structure to achieve economic optimization. The upper layer calculates the optimal setpoint based on a steady-state model and sends it down to the lower-level controller (MPC) for tracking. In contrast, this invention directly embeds the economic objective into a single-layer dynamic operating objective function framework, performing dynamic economic optimization directly at each sampling time. Moreover, in conventional control methods, the steady-state optimization result is treated as an instruction and strictly executed by the lower layer, often leading to problems such as response lag and unreachable setpoints, making it difficult to adapt to the dynamic process of frequently changing actual operating conditions. In contrast, the steady-state optimization result of this invention serves only as a reference direction for dynamic optimization. The dynamic optimization process is guided by economic optimization rather than simply aiming to minimize tracking deviation, thus ensuring a smooth and stable convergence to the economically optimal steady-state operating point during the transition from dynamic to steady state.

[0070] like Figure 5 As shown, the present invention also provides a cost-optimized AGC predictive control device for thermal power units, employing the above-mentioned predictive control method. The device includes: The initial condition determination module is used to acquire the actual operating data and AGC load commands of the thermal power unit and provide a pre-built dynamic model of the unit. The objective function determination module is used to determine the steady-state operation objective function of the thermal power unit from the current state to the steady state, taking into account cost indicators, and to determine the dynamic operation objective function of the thermal power unit from the current state to the steady state through continuous control, taking into account cost indicators. The steady-state parameter determination module is used to determine the optimal steady-state parameters of the thermal power unit based on the steady-state operation objective function and in combination with AGC load commands and the unit dynamic model. The parameter determination control module is used to determine the control parameters of the thermal power unit based on the dynamic operating objective function, combined with the optimal steady-state parameters, AGC load commands, actual operating data and unit dynamic model, and to control the thermal power unit.

[0071] Furthermore, the initial condition determination module is used for: Determine the state variables, controlled variables, and control variables of different units of the thermal power unit; The energy-mass conversion mechanism under the operating state of thermal power units is analyzed, and the state variables, controlled variables and control variables of different units of thermal power units are combined to construct the controlled variable function and the change function of state variables that are interrelated to different units. The system acquires operating data of thermal power units under different load conditions, identifies the controlled and variable functions that are interrelated between different units, and provides a dynamic model of the unit.

[0072] Furthermore, the initial condition determination module is also used for: Determine the steady-state model of the thermal power unit with respect to static parameters under steady-state operation; Based on the steady-state operating data of thermal power units under different load conditions, and combined with the steady-state model of thermal power units under operating conditions, static parameters are determined. The steady-state operating data of different units of thermal power units under different load conditions were fitted by regression analysis to determine the operating function; Dynamic operating data where the load and main steam pressure of thermal power units fluctuate beyond the threshold are selected from operating data under different load conditions. Based on static parameters, running functions, and dynamic running data, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, the particle swarm optimization algorithm is used to solve and determine the dynamic parameters. Based on static parameters, operating functions, and dynamic parameters, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, a dynamic model of the unit is given.

[0073] Furthermore, the objective function determination module is used for: The constraints of thermal power units under steady-state operation are determined, and the objective function for steady-state operation is given with the goal of minimizing the operating costs generated by controlling the thermal power units to reach steady state.

[0074] Furthermore, the objective function determination module is also used for: Based on the operating costs of controlling thermal power units, the benefits of AGC regulation, and the real-time control coefficients for achieving steady state, the control cost of controlling a thermal power unit in a single operation is determined. The constraints and sampling step size of the thermal power unit under dynamic operation are determined, and the objective function for dynamic operation is determined with the goal of minimizing the total control cost of controlling the thermal power unit within the step size.

[0075] Furthermore, the steady-state parameter determination module is also used for: The AGC load command is input into the unit's dynamic model, and based on the steady-state operation objective function, it is solved by a numerical optimization solver to obtain the control and state variables when the thermal power unit executes the AGC load command to achieve steady-state operation and minimize costs.

[0076] Furthermore, the parameter determination control module is also used for, including: The control and state variables of the thermal power unit when executing AGC load commands to achieve steady-state operation and minimize costs are input into the dynamic operation objective function. The AGC load commands and actual operating data are input into the unit dynamic model, and the numerical optimization solver is used to solve the problem to obtain the control variables of each control within the sampling step of the thermal power unit.

[0077] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A cost-optimized AGC predictive control method for thermal power units, characterized in that, include: Obtain actual operating data and AGC load commands from thermal power units, and provide a pre-built dynamic model of the units; Determine the steady-state operation objective function of the thermal power unit from the current state to reach a steady state, taking into account cost indicators; determine the dynamic operation objective function of the thermal power unit from the current state to reach a steady state through continuous control, taking into account cost indicators. Based on the steady-state operation objective function, and combined with AGC load commands and unit dynamic models, the optimal steady-state parameters of thermal power units are given; Based on the dynamic operating objective function, and combined with the optimal steady-state parameters, AGC load commands, actual operating data and unit dynamic model, the control parameters of the thermal power unit are given, and the thermal power unit is controlled.

2. The predictive control method as described in claim 1, characterized in that, The pre-construction of the unit dynamic model includes: Determine the state variables, controlled variables, and control variables of different units of the thermal power unit; The energy-mass conversion mechanism under the operating state of thermal power units is analyzed, and the state variables, controlled variables and control variables of different units of thermal power units are combined to construct the controlled variable function and the change function of state variables that are interrelated to different units. The system acquires operating data of thermal power units under different load conditions, identifies the controlled and variable functions that are interrelated between different units, and provides a dynamic model of the unit.

3. The predictive control method as described in claim 1, characterized in that, The thermal power units are supercritical coal-fired power generating units.

4. The predictive control method as described in claim 3, characterized in that, The state variables include a first state variable, a second state variable, and a third state variable; the controlled variables include a first controlled variable, a second controlled variable, and a third controlled variable; the control variables include a first control variable, a second control variable, and a third control variable; the controlled variable functions include a first controlled variable function, a second controlled variable function, and a third controlled variable function; and the change variable functions include a first change variable function, a second change variable function, and a third change variable function. The unit dynamic model satisfies the following relationship: ； In the formula, u1, u2, and u3 are the first control variable, the second control variable, and the third control variable, respectively; y1, y2, and y3 are the first controlled variable, the second controlled variable, and the third controlled variable, respectively; and x1, x2, and x3 are the first state variable, the second state variable, and the third state variable, respectively. Let x1, x2, and x3 be the changes respectively, F1(·) be the first change function, F2(·) be the second change function, F3(·) be the third change function, H1(·) be the first controlled variable function, H2(·) be the second controlled variable function, and H3(·) be the third controlled variable function.

5. The predictive control method as described in any one of claims 2 to 4, characterized in that, Operational data includes steady-state operational data and dynamic operational data; the model parameters of the unit's dynamic model include static parameters, dynamic parameters, and operating functions; Identify the controlled variable functions and variable functions that are interrelated between different units, and provide a dynamic model of the unit, including: Determine the steady-state model of the thermal power unit with respect to static parameters under steady-state operation; Based on the steady-state operating data of thermal power units under different load conditions, and combined with the steady-state model of thermal power units under operating conditions, static parameters are determined. The steady-state operating data of different units of thermal power units under different load conditions were fitted by regression analysis to determine the operating function; Dynamic operating data where the load and main steam pressure of thermal power units fluctuate beyond the threshold are selected from operating data under different load conditions. Based on static parameters, running functions, and dynamic running data, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, the particle swarm optimization algorithm is used to solve and determine the dynamic parameters. Based on static parameters, operating functions, and dynamic parameters, and combining the controlled variable functions and the change function of state variables that are interrelated between different units, a dynamic model of the unit is given.

6. The predictive control method as described in claim 5, characterized in that, The steady-state model of a thermal power unit under operating conditions satisfies the following relationship: ； In the formula, * represents the steady-state value, u1 and u2 are the first and second control variables, respectively, x1 and x3 are the first and third state variables, respectively, l is the ratio of the specific enthalpy of the superheater outlet steam to the specific enthalpy of the separator steam in the thermal power unit, and D t Indicates the main steam flow rate, h fw h represents the feedwater specific enthalpy of a thermal power unit. t This indicates the specific enthalpy of the main steam. k 0 represents the heat absorption gain of the boiler superheater section in a thermal power unit. k 2 represents the turbine gain of the thermal power unit, and Ne represents the power generation capacity.

7. The predictive control method as described in claim 1 or 6, characterized in that, The determination of the steady-state operating objective function for thermal power units, considering cost indicators, includes: The constraints of thermal power units under steady-state operation are determined, and the objective function for steady-state operation is given with the goal of minimizing the operating cost generated by controlling the thermal power units to reach steady state. The determination of the dynamic operating objective function for cost indicators of thermal power units includes: Based on the operating costs of controlling thermal power units, the benefits of AGC regulation, and the real-time control coefficients for achieving steady state, the control cost of controlling a thermal power unit in a single operation is determined. The constraints and sampling step size of the thermal power unit under dynamic operation are determined, and the objective function for dynamic operation is determined with the goal of minimizing the total control cost of controlling the thermal power unit within the step size.

8. The predictive control method as described in claim 7, characterized in that, Based on the steady-state operation objective function, and combined with AGC load commands and the unit dynamic model, the optimal steady-state parameters of the thermal power unit are given, including: The AGC load command is input into the unit's dynamic model, and based on the steady-state operation objective function, it is solved by a numerical optimization solver to obtain the control and state variables when the thermal power unit executes the AGC load command to achieve steady-state operation and minimize costs.

9. The predictive control method as described in claim 8, characterized in that, Based on the dynamic operating objective function, and combined with the optimal steady-state parameters, AGC load commands, actual operating data, and the unit's dynamic model, the control parameters of the thermal power unit are given, including: The control and state variables of the thermal power unit when executing AGC load commands to achieve steady-state operation and minimize costs are input into the dynamic operation objective function. The AGC load commands and actual operating data are input into the unit dynamic model, and the numerical optimization solver is used to solve the problem to obtain the control variables of each control within the sampling step of the thermal power unit.

10. A cost-optimized AGC predictive control device for thermal power units, characterized in that, The apparatus employing the predictive control method as described in any one of claims 1 to 9 comprises: The initial condition determination module is used to acquire the actual operating data and AGC load commands of the thermal power unit and provide a pre-built dynamic model of the unit. The objective function determination module is used to determine the steady-state operation objective function of the thermal power unit from the current state to the steady state, taking into account cost indicators, and to determine the dynamic operation objective function of the thermal power unit from the current state to the steady state through continuous control, taking into account cost indicators. The steady-state parameter determination module is used to determine the optimal steady-state parameters of the thermal power unit based on the steady-state operation objective function and in combination with AGC load commands and the unit dynamic model. The parameter determination control module is used to determine the control parameters of the thermal power unit based on the dynamic operating objective function, combined with the optimal steady-state parameters, AGC load commands, actual operating data and unit dynamic model, and to control the thermal power unit.

Images