Interconnected power grid regulation and control method and system based on explicit model predictive control

Abstract

The invention relates to a power system frequency control technology, in particular to an interconnected power grid regulation and control method and system based on explicit model predictive control. The method comprises the following steps: constructing a control model of grid-connected industrial power grid system frequency fluctuation and interconnected power grid tie line power fluctuation in which electrolytic aluminum load and self-contained thermal power generating unit secondary frequency modulation jointly participate; establishing a model prediction controller, constructing an optimization objective function of the grid-connected industrial power grid system model, and converting the optimization objective function into a quadratic programming problem; converting the optimization objective function into a multi-parameter quadratic programming problem and solving the multi-parameter quadratic programming problem to obtain an explicit relationship between a control variable and a system state quantity, and converting the model prediction controller into an explicit model prediction controller; the optimal explicit control law of each grid-connected industrial power grid system is obtained through offline calculation; by calling the optimal explicit control law corresponding to the current state of the system, the corresponding control quantity is calculated, and the method can enable the electrolytic aluminum load and the secondary frequency modulation of the thermal power generating unit to coordinate and stabilize the power fluctuation of the tie line.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system frequency control, and particularly relates to a method and system for regulating and controlling an interconnected power grid based on explicit model predictive control. Background Art

[0002] In a traditional interconnected power system, the active power regulation characteristics of each regional power grid are not very different. When an unbalanced power disturbance occurs, the synchronous generator sets with an automatic generation control (AGC) system within their respective regions can adjust their output power, thereby jointly suppressing the power fluctuations of the inter-regional tie lines. In contrast, industrial power grids with a high penetration of wind power have significant differences in terms of magnitude in terms of regulation means and regulation capacity compared with large power grids. Due to the frequency support of the large power grid, the fluctuations of wind power will be suppressed by the large power grid on the grid-connected side through the tie lines, and the frequency stability problem of the grid-connected industrial power grid can be ensured. However, this mode has two drawbacks: for industrial enterprises, a large tie line capacity will generate huge tie line reserve capacity fees, which is not conducive to the economic operation of the enterprises; for the large power grid side, when a large number of industrial power grids with a high penetration of renewable energy are connected, it will exert great pressure on the safe operation of the large power grid side, and additional fast-regulation generating units need to be built.

[0003] The volatility of wind power is the main reason for the power fluctuations of the tie lines between the grid-connected industrial power grid and the large power grid. Wind power shows different degrees of volatility at different time scales such as hourly, minute-level, and second-level. Among them, the ultra-short-term second-level wind power fluctuations pose higher requirements for the active power regulation rate of the power system, and are one of the technical difficulties in suppressing the tie line power fluctuations at present. However, the coal-fired thermal power units on the traditional power supply side are limited by the regulation rate of their steam turbines and are difficult to effectively track and suppress such ultra-short-term second-level wind power fluctuations. Direct load control has fast power regulation characteristics, so performing power control on the load side of the industrial power grid is an effective method for suppressing tie line power fluctuations.

[0004] Designing a coordinated control system for jointly suppressing the tie line power fluctuations by the electrolytic aluminum load and the secondary frequency modulation of thermal power units can form a control command for solving the model predictive control (MPC) optimization problem, so that the electrolytic aluminum load and the secondary frequency modulation of thermal power units can jointly suppress the tie line power fluctuations. However, due to the characteristics of repeated online optimization calculations of model predictive control, the model predictive control technology can only be applied to occasions with a small problem scale or a slow system dynamic change, and is difficult to be applied to systems with a high sampling frequency and a fast dynamic change. Summary of the Invention

[0005] The present invention provides a control method and system for interconnected power grids based on explicit model predictive control, which is used to solve the defect that model predictive control in the prior art is difficult to be applicable to systems with high sampling frequencies and fast dynamic changes, and to effectively control the fluctuations of the power of the tie lines of the interconnected power grids.

[0006] The present invention provides a control method for interconnected power grids based on explicit model predictive control, including the following steps: Step 1. Construct a control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the tie lines of the interconnected power grids, in which the electrolytic aluminum load and the secondary frequency regulation of the self-provided thermal power units participate together; Step 2. Based on the control model, establish a model predictive controller, including determining the state variables, disturbance variables and control variables of the grid-connected industrial power grid system, constructing an optimization objective function corresponding to the model predictive control optimization problem of the grid-connected industrial power grid system, and converting it into a quadratic programming problem; Step 3. According to the multi-parameter theory, convert the optimization objective function into a multi-parameter quadratic programming problem, and by solving the multi-parameter quadratic programming problem, obtain the explicit relationship between the control variables and the state variables of the grid-connected industrial power grid system, and convert the model predictive controller into an explicit model predictive controller; Step 4. Convert the online rolling optimization process of the model predictive controller into an offline calculation; obtain the optimal explicit control law of each grid-connected industrial power grid system through offline calculation; monitor the current state of the grid-connected industrial power grid system online through the explicit model predictive controller, call the corresponding optimal explicit control law, and calculate the corresponding control quantity.

[0007] According to the control method for interconnected power grids based on explicit model predictive control provided by the present invention, the control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the tie lines of the interconnected power grids includes a first area control model and a second area control model; the first area control model includes the control of self-provided thermal power units, the control of electrolytic aluminum load, the control of self-provided wind power units, and the control of other load disturbances, and the second area control model is the control of the large power grid. The first area control model and the second area control model are connected through tie lines; The frequency change of the grid-connected industrial power grid system is expressed as:

[0008] where M is the system equivalent inertia coefficient, D is the system equivalent damping coefficient, is the governor-turbine dynamic model, R is the primary frequency regulation droop characteristic coefficient, is the secondary frequency regulation power adjustment command of the thermal power unit, is the active power output adjustment command of the electrolytic aluminum load, is the disturbance quantity of other loads; represents the differential of the change in the system frequency to the change in the tie-line power, s is the complex frequency, is the dynamic response time constant of the electrolytic aluminum load; The change in the tie-line power is expressed as: .

[0009] According to an interconnected power grid regulation method based on explicit model predictive control provided by the present invention, the electrolytic aluminum load control realizes the regulation of the active power of the electrolytic aluminum load by adjusting the control current of the electrolytic aluminum load saturable reactor.

[0010] According to an interconnected power grid regulation method based on explicit model predictive control provided by the present invention, the implementation process of establishing a model predictive controller based on the control model includes the following sub-steps: Step 2.1. Determine the implementation methods of the state variables, disturbance variables, and control variables of the grid-connected industrial power grid system as follows. Establish a secondary frequency modulation controller for thermal power units and an electrolytic aluminum load controller, and upload the state variables of the grid-connected industrial power grid system, the wind power disturbance and the disturbance quantity of other loads to the model predictive controller through the wide-area measurement system. The model predictive controller calculates the secondary frequency modulation power regulation command for the thermal power units and the active power output regulation command for the electrolytic aluminum load, and sends them to the electrolytic aluminum load controller and the secondary frequency modulation controller for thermal power units; where is the change in the active power of the boiler, is the change in the active power of the governor, is the change in the system frequency of the grid-connected industrial power grid system, is the change in the tie-line power; Step 2.2. The implementation method of constructing the optimization objective function corresponding to the model predictive control optimization problem of the grid-connected industrial power grid system is as follows. Considering the model predictive controller in the feasible region of the electrolytic aluminum load regulation takes a discrete form, and its discrete state-space equation is obtained. Substitute the t control variables k predicted at the , …, into it to obtain the t state of the grid-connected industrial power grid system predicted at the moment, where , The secondary frequency regulation power regulation command of the thermal power unit and the active output regulation command of the electrolytic aluminum load are used as the control variables of the model predictive controller; the optimization objective function of the model predictive control is constructed with the goal of minimizing the system frequency deviation and the change in the interconnection line power while reducing the electrolytic aluminum load response; Step 2.3. The optimization objective function of model predictive control is transformed into a quadratic programming problem as follows: If there is an optimal variable sequence , U Include variables, satisfying , then the optimization objective function of the model predictive control constructed in step 2.2 can be transformed into a constrained finite-time optimal control function; represents the first control variables, acting on time, for t The state variables of the grid-connected industrial power grid system at time ; Will t Predicted at all times The state variables of the grid-connected industrial power grid system at time Substituting the above constrained finite-time optimal control function for replacement, we obtain a quadratic programming problem for a discrete state space model.

[0011] According to an interconnected power grid control method based on explicit model predictive control provided by the present invention, the conversion of the model predictive controller into the explicit model predictive controller comprises the following steps: Step 3.1. The implementation method of constructing a multi-parameter quadratic programming problem is as follows: The quadratic programming problem of the discrete state space model is transformed into a standard multi-parameter quadratic programming problem by using variable substitution; given a closed parameter polyhedron set X ,definition Represents the feasible parameter domain of the multi-parameter quadratic programming problem, and the cost function on this parameter domain is limited; define the state of any grid-connected industrial power grid system , It represents the minimum value of the cost function of the multi-parameter quadratic programming problem in this state, and the optimal value corresponding to this state is ; Step 3.2. The steps to solve the multi-parameter quadratic programming problem are as follows: Step 3.2.1. Define the constraints as follows: When the system is in i When the state interval is iGroup constraints , , are constant matrices. If , then this constraint is an effective constraint, denoted by , , ; if , then this constraint is a non-effective constraint, denoted by , , ; Let be the set of constraint subscripts, m denote the number of groups of constraint conditions. For any effective constraint subscript set , matrices and respectively represent and submatrices; For a given current state quantity of a grid-connected industrial power grid system, the set represents the critical domain related to the effective constraint subscript set L ; Step 3.2.2. Solve for the initial state The implementation method for the corresponding effective constraint subscript set is as follows, Consider the multi-parameter quadratic programming problem. Let the constant parameter matrix , and the feasible parameter domain is convex; The optimal value function is continuous and convex on the polyhedron, and is affine on each critical domain; The cost function is continuous, convex and piecewise quadratic on the polyhedron; Obtain an initial state of a grid-connected industrial power grid system. Select as the Chebyshev ball center contained in the parameter polyhedron set X . The initial state is located within the parameter polyhedron set X , and makes the multi-parameter quadratic programming problem feasible for ; Solve the multi-parameter quadratic programming problem to obtain the corresponding optimal value . Since , the solution is unique, and an effective constraint subscript set outside the constraints of the multi-parameter quadratic programming problem can be uniquely determined; Step 3.2.3. Determine the affine relationship between and x . The implementation method is as follows, When this step is executed for the first time, let , , consider a set of active constraint subscripts , and assume that the linear independent constraint qualification (LICQ) condition is satisfied; on the set of active constraint subscripts corresponding critical region , solve for the optimal value according to the necessary condition of the first-order optimal solution of the multi-parameter quadratic programming problem, the KKT condition Regarding x explicit affine function of, denoted as the optimal value function ; Step 3.2.4. Determine the implementation method of the remaining critical region as follows, Through the optimal value function obtain the critical region expression corresponding to the initial state, eliminate the redundant inequalities therein, and obtain a set of active constraint subscripts corresponding critical region a compact expression of; the critical region is a polyhedron in the parameter polyhedron set X space, representing the largest set that can keep the set of active constraint subscripts unchanged at the minimum value; once the critical region is defined, the remaining space needs to be explored and a new critical region is generated; Step 3.2.5. According to the principle of receding horizon optimization, take the first term of the sequence of optimal variation U and apply it to the controlled object; judge whether is satisfied. If not, then let , return to Step 3.2.3, and obtain the corresponding set of active constraint subscripts corresponding state partition and the control law of the corresponding critical region in the same way; if so, end the iteration; After exploring the parameter space of the entire parameter polyhedron set X , obtain a polyhedron domain with the same form of optimal value function ; at this time, the receding horizon optimization process of model predictive control has been transformed into offline calculation. According to the current grid-connected industrial power grid system state variables x of the system, obtain the corresponding optimal value function ; Substitute the obtained optimal value function into , and obtain the explicit expression of the sequence of optimal variables U regarding x : , where H , F , G , W ,S is a constant parameter matrix; denote t the state of the grid-connected industrial power grid system at time the corresponding optimal solution is , and take the first term as the optimal output of the explicit model predictive controller to act on the controlled object.

[0012] According to an interconnected power grid regulation method based on explicit model predictive control provided by the present invention, the implementation manner of step 4 is as follows. Perform offline calculation according to the given constraint conditions, consider the multi-parameter theory to construct a multi-parameter quadratic programming problem, divide the state space of the grid-connected industrial power grid system into multiple state partitions, solve the multi-parameter quadratic programming problem to obtain the optimal control law corresponding to each state partition, and store it in the memory. Calculate the optimal output of the explicit model predictive controller for the grid-connected industrial power grid system under different states according to this optimal control law. According to the explicit model predictive control algorithm, obtain the explicit model predictive controller equation:

[0013] In the formula, m is the number of partitions of the explicit model predictive controller; solve the explicit model predictive controller equation to obtain the control quantity of the grid-connected industrial power grid system.

[0014] On the other hand, the present invention also provides an interconnected power grid regulation system based on explicit model predictive control, including the following modules. The first module is used to construct a control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the interconnected power grid tie line jointly participated by the electrolytic aluminum load and the secondary frequency regulation of the self-provided thermal power unit. The second module is used to establish a model predictive controller based on the control model, including determining the state quantity, disturbance quantity and control variable of the grid-connected industrial power grid system, constructing the corresponding optimization objective function of the model predictive control optimization problem of the grid-connected industrial power grid system, and converting it into a quadratic programming problem. The third module is used to convert the optimization objective function into a multi-parameter quadratic programming problem according to the multi-parameter theory, obtain the explicit relationship between the control variable and the state quantity of the grid-connected industrial power grid system by solving the multi-parameter quadratic programming problem, and convert the model predictive controller into an explicit model predictive controller. The fourth module is used to convert the online rolling optimization process of the model predictive controller into an offline calculation; obtain the optimal explicit control law of each grid-connected industrial power grid system through offline calculation; online monitor the current state of the grid-connected industrial power grid system through the explicit model predictive controller, call the corresponding optimal explicit control law, and calculate the corresponding control quantity.

[0015] In addition, the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the method for regulating an interconnected power grid based on explicit model predictive control is implemented.

[0016] In addition, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method for regulating an interconnected power grid based on explicit model predictive control is implemented.

[0017] In addition, the present invention further provides a computer program product, including a computer program. When the computer program is executed by a processor, the method for regulating an interconnected power grid based on explicit model predictive control is implemented.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) Based on the method of explicit model predictive control, through the multi-parameter theory, the complex rolling optimization process of ordinary model predictive control is transformed into two steps: offline dividing the parameter interval to find the optimal control law and online searching and calculating the control signal. This greatly accelerates the system response speed. When there are instantaneous fluctuations such as wind power fluctuations in the power grid, the controller can quickly respond, coordinate the secondary frequency modulation of thermal power units and the active power output of electrolytic aluminum loads, and quickly suppress the power fluctuations of the system tie line. Compared with traditional control methods, this method can better regulate the tie line power fluctuations caused by short-term disturbances in the interconnected power grid. Without a complex rolling optimization process, the optimal control signal can be directly calculated, and the response is faster and more flexible. This method also takes the electrolytic aluminum load as the adjustment object, adjusts the power fluctuations of the industrial power grid tie line from the load side, and solves the problems of slow response and limited adjustment capacity of the secondary frequency modulation of thermal power units.

[0019] (2) Based on high-precision ultra-short-term wind power prediction, the explicit model predictive controller can use the prediction model to predict the future dynamics of the system, and at the same time consider the system constraints during the control process to ensure the optimization feasibility. Therefore, compared with the traditional PID controller based only on current and historical information, it has a more obvious suppression effect. Under different wind power fluctuation scenarios, the tie line power fluctuations can be controlled within a reasonable range.

[0020] (3) The explicit model predictive controller makes full use of the advantages of sufficient secondary frequency modulation adjustment capacity of thermal power units and fast adjustment speed of electrolytic aluminum loads, reasonably distributes the adjustment amount, preferentially uses thermal power units for adjustment, and fully exerts the fast adjustment ability of electrolytic aluminum loads. Through the coordinated cooperation of the two, it can suppress large-amplitude and rapid fluctuations of wind power and reduce the impact on the normal production of electrolytic aluminum loads, which is more beneficial to the normal production of electrolytic aluminum loads.

[0021] (4) The explicit model predictive controller can more reasonably utilize the electrolytic aluminum load to participate in the control of the power fluctuation of the interconnection grid tie line. When a large number of wind farms are disconnected from the grid, causing instantaneous unbalanced power, the controller can quickly adjust the power of the electrolytic aluminum load to reduce the unbalanced amount of system power; at the same time, the controller reduces the adjustment amount of the electrolytic aluminum load power by adjusting the output of the thermal power unit, so as to restore the electrolytic aluminum load power to the rated value. Therefore, the explicit model predictive controller can more reasonably utilize the electrolytic aluminum load to participate in the suppression of tie line power fluctuations and reduce the impact on the production of the electrolytic aluminum load. Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the implementation examples or the description of the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0023] Figure 1 It is a flowchart of the method for controlling the power fluctuation of the interconnection grid tie line based on explicit model predictive control provided by the embodiment of the present invention; Figure 2 It is an equivalent model of the grid-connected industrial grid for the method of controlling the power fluctuation of the interconnection grid tie line based on explicit model predictive control provided by the embodiment of the present invention; Figure 3 It is a regulation model considering the electrolytic aluminum load control for the method of controlling the power fluctuation of the interconnection grid tie line based on explicit model predictive control provided by the embodiment of the present invention; Figure 4 It is a model predictive controller model for the method of controlling the power fluctuation of the interconnection grid tie line based on explicit model predictive control provided by the embodiment of the present invention; Figure 5 It is an explicit model predictive control structure diagram for the method of controlling the power fluctuation of the interconnection grid tie line based on explicit model predictive control provided by the embodiment of the present invention; Figure 6 It is a parameter polyhedron region division diagram for fixing other state variables and performing on state variables 1 and 2 provided by the embodiment of the present invention; Figure 7(a) is a schematic diagram of the difference in the closed-loop simulation results between the sub-optimal explicit MPC and the explicit MPC of the thermal power unit output provided by the embodiment of the present invention; Figure 7(b) is a schematic diagram of the difference in the closed-loop simulation results between the sub-optimal explicit MPC and the explicit MPC of the electrolytic aluminum load output provided by the embodiment of the present invention; Figure 7 (c) is a schematic diagram showing the difference in the closed-loop simulation results between the sub-optimal explicit MPC and the explicit MPC of the active power fluctuation of the tie line provided by the embodiment of the present invention; Figure 8 is a schematic diagram of the simulation control of applying the explicit model predictive controller to a simple power system provided by the embodiment of the present invention; Figure 9 is a schematic diagram of the structure of the electronic device provided by the embodiment of the present invention. Detailed implementation manners

[0024] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0025] This embodiment provides a method for controlling the power fluctuation of the tie line of an interconnected power grid based on explicit model predictive control. For the standard multi-parameter quadratic programming problem constructed for the regulation model of a grid-connected high-energy-consuming industrial power grid considering load control, the rolling optimization process is transformed into off-line calculation, the state partition is divided, and the optimal explicit control rate corresponding to each state partition is calculated. The optimal control signal is directly obtained according to the partition to which the current state of the system belongs, which greatly accelerates the method for suppressing the power fluctuation speed of thermal power units and electrolytic aluminum loads.

[0026] The technical solution adopted in this embodiment is as follows: construct the MPC optimization problem of the high-energy-consuming industrial power grid, and transform it into a multi-parameter quadratic programming problem considering the multi-parameter theory; determine the feasible parameter domain, and divide the critical domain according to the effective constraint conditions; assume that the effective constraint combination satisfies the linearly independent normal constraint conditions, and according to the optimal value and the related Lagrange multiplier vector uniquely define the corresponding state x of the affine function, obtain the optimal control law for this interval, and calculate the secondary frequency modulation output of the self-provided thermal power unit and the output value of the electrolytic aluminum load participating in power fluctuation suppression according to this control law. As Figure 1 shown, it specifically includes the following steps: Construct a control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the tie line of the interconnected power grid jointly participated by the electrolytic aluminum load and the secondary frequency modulation of the self-provided thermal power unit; the connection model between the grid-connected industrial power grid and the large power grid is as Figure 2 shown, Region 1 is the industrial power grid, including self-provided thermal power units, electrolytic aluminum loads , wind farm power and other loads , is the active power output of the self-provided thermal power unit, is the frequency fluctuation of the grid-connected industrial power grid system, T is the governor, G is the boiler controlled by the governor, R is the governor controller, is the active power controlled by the governor. Area 2 is the large power grid; the industrial power grid is connected to the large power grid in Area 2 through a tie line, is the tie line power, is the change in tie line power. The capacity of the large power grid is much larger than that of the industrial power grid. The self-provided thermal power units and wind turbines in Area 1 supply power to the electrolytic aluminum load and other loads. When the internal power supply capacity is insufficient, power is purchased from Area 2 through the tie line.

[0027] The regulation model of the grid-connected industrial power grid system considering the electrolytic aluminum load control is as Figure 3 shown. When the disturbance of other loads acts on the grid-connected industrial power grid system, it will cause the system frequency to change and the tie line power to fluctuate ; the regulation instructions of the thermal power unit and the electrolytic aluminum load and act on the governor-turbine dynamic model and the electrolytic aluminum load dynamic response model respectively, and the actual active power response of the turbine and the actual active power output regulation instruction of the electrolytic aluminum load are obtained; the differential controller and the proportional-integral regulator with a proportional coefficient of together constitute the feedback branch, and the feedback signal is obtained; and , , act on the grid-connected industrial power grid system together to regulate the frequency change of the grid-connected industrial power grid system. M is the system equivalent inertia coefficient, D is the system equivalent damping coefficient, r is the primary frequency regulation droop characteristic coefficient of the thermal power unit, is the electrolytic aluminum load dynamic response time constant, and are the turbine time constants, is the first-order inertia link gain of the electrolytic aluminum load dynamic response, s is the complex frequency.

[0028] The frequency change of the grid-connected industrial power grid system Expressed as:

[0029] Wherein M is the system equivalent inertia coefficient, D is the system equivalent damping coefficient, is the governor - turbine dynamic model, R is the primary frequency regulation droop characteristic coefficient, is the secondary frequency regulation power adjustment command of the thermal power unit, is the active power output adjustment command of the electrolytic aluminum load, is the disturbance quantity of other loads; represents the differential of the change in system frequency to the change in tie - line power, s is the complex frequency, is the dynamic response time constant of the electrolytic aluminum load; Change in tie - line power Expressed as:

[0030] According to the mean - value theorem, the change in steady - state system frequency is:

[0031] The change in steady - state tie - line power is:

[0032] It can be seen from this that when there is a power disturbance in the grid - connected industrial power grid, the change in tie - line power is related to the secondary frequency regulation adjustment of the thermal power unit and the adjustment of the electrolytic aluminum load. By coordinating the secondary frequency regulation of the thermal power unit and the adjustment of the electrolytic aluminum load, effective tie - line power regulation can be achieved and the tie - line power impact can be reduced.

[0033] It should be noted that since the reactive power of the grid - connected industrial power grid is sufficient, it is difficult to adjust the AC voltage on the high - voltage side of the electrolytic aluminum load, and the method of adjusting the load power based on the AC voltage on the high - voltage side cannot be realized. The grid - connection mode does not affect the control function of the current - stabilizing system. The adjustment method based on the saturable reactor can realize the adjustment of the active power of the electrolytic aluminum load by adjusting the control current of the saturable reactor of the electrolytic aluminum load. Therefore, this method is used to adjust the electrolytic aluminum load.

[0034] Based on the control model, a model predictive controller is established, including determining the state variables, disturbance variables and control variables of the grid - connected industrial power grid system, constructing the corresponding optimization objective function of the model predictive control optimization problem of the grid - connected industrial power grid system, and converting it into a quadratic programming problem; as Figure 4 shown, a secondary frequency regulation controller for the thermal power unit and a controller for the electrolytic aluminum load are established, and the wind power disturbance , other load disturbances and the change in tie-line power Input to the model predictive controller, and the regulation command of the thermal power unit is obtained through calculation and the regulation command of the electrolytic aluminum load ; r is the droop characteristic coefficient of the primary frequency regulation of the thermal power unit, which constitutes a feedback branch to the dynamic model of the steam turbine to obtain a feedback signal . Act together with on the speed governor of the thermal power unit to obtain the power regulation command of the boiler , Act on the boiler to obtain the active power controlled by the speed governor ; Act on the dynamic response model of the electrolytic aluminum load; is a differential controller, which together with the proportional-integral regulator with a proportional coefficient of constitutes a feedback branch to the grid-connected industrial power grid system to obtain a feedback signal ; Act together with , , on the grid-connected industrial power grid system to adjust the frequency change of the system ; is the actual active power response of the electrolytic aluminum load; T G is the speed governor time constant, T CH is the boiler time constant.

[0035] 2.1. Determine the state variables, disturbance variables and control variables of the grid-connected industrial power grid system; Establish a secondary frequency regulation controller for the thermal power unit and an electrolytic aluminum load controller, and upload the state variables of the grid-connected industrial power grid system , wind power disturbance and other load disturbance variables to the model predictive controller through the wide-area measurement system. The model predictive controller calculates the secondary frequency regulation power regulation command of the thermal power unit and the active power output regulation command of the electrolytic aluminum load , and send them to the electrolytic aluminum load controller and the secondary frequency regulation controller of the thermal power unit; where is the change in boiler active power, is the change in governor active power, is the frequency change of the grid-connected industrial power grid system, is the change in tie-line power.

[0036] 2.2. Construct the corresponding optimization objective function for the model predictive control optimization problem of the grid-connected industrial power grid system; The model predictive controller considering the feasible region of electrolytic aluminum load regulation adopts a discrete form, and the discrete state-space model is expressed as:

[0037] where is t the state variable of the grid-connected industrial power grid system at time , is the state variable of the grid-connected industrial power grid system at time, is t the control variable of the model predictive controller at time, is t the output of the grid-connected industrial power grid system at time, A is the state variable coefficient matrix of the state equation, B is the input variable coefficient matrix of the state equation, C is the state variable coefficient matrix of the output equation; Substitute the t control variables k predicted at time ,…, to get:

[0038] where represents t the state predicted at time ; is the dimensional coefficient matrix, is the th predicted control variable substituted into the state variable coefficient matrix corresponding to the state equation; Set the optimization objective function to minimize the system frequency deviation and the tie-line power fluctuation while reducing the electrolytic aluminum load response, expressed as:

[0039]

[0040] In the formula, J is the optimization objective function; is the state variable of the grid-connected industrial power grid system; is t the state variable predicted at time , is its transpose; represents t the tThe state variable at a certain moment, that is, the predicted initial state variable; is the secondary frequency regulation reference value of the self-provided thermal power unit and the active power regulation command of the electrolytic aluminum load, serving as the control variable of the model predictive controller; is the optimization objective function t The sequence of optimized variables predicted at a certain moment, represents the th control variable in this sequence of optimized variables, acting on at a certain moment, is the k th control variable, acting on at a certain moment, is its transpose, is its change amount; and are the upper and lower limits of the constraints of the control variable, and are the upper and lower limits of the constraints of the change amount of the control variable; is the change in the power of the tie line of the grid-connected industrial power grid system and the change amount of the system frequency, serving as the output variable of the system; is t the k th output variable predicted at a certain moment, acting on at a certain moment, and are its upper and lower limits of constraints; Q and R are the weight coefficient matrices corresponding to the state variable and the control variable in the optimization objective function respectively; and are the prediction step length and the control step length respectively; It should be noted that the change in wind power is input to the model predictive controller as an observable disturbance. For the control variable u , the constraints are specifically the upper and lower limits of the secondary regulation ability of the self-provided thermal power unit, the rate limit of the secondary frequency regulation of the thermal power unit, and the upper and lower limits of the regulation ability of the electrolytic aluminum load.

[0041] 2.3. Convert the optimization objective function of the model predictive control into a quadratic programming problem; If there exists a sequence of optimized variables U that satisfies , U contains variables, the optimization objective function can be converted into a constrained finite-time optimal control function:

[0042]

[0043] In the formula, represents the optimization function Take the minimum value under the optimized variable sequence which is the state variable predicted at t time ; denote t the state variable predicted at t time, that is, the predicted initial state variable which is t the state variable predicted at time denote t the k -th control variable in the optimized variable sequence predicted at time, acting on time; t is the output variable predicted at time; , is a positive semi - definite matrix; and respectively represent the matrix dimensions depending on the dimension of the state variable x and depending on the dimension of the optimal control variable u ; Adjust Q , R to change the weights of different state variables and control variables; K is the feedback gain; , , are the output time domain length, input time domain length and constraint time domain length respectively , that is, the control step; As described in step 2.2 t the state predicted at time is . Substitute this state equation into the above - mentioned constrained finite - time optimal control function for replacement, and obtain a quadratic programming problem of a discrete state - space model, whose form is:

[0044] In the formula, represents the cost function of this quadratic programming problem; , U is the optimized variable sequence denote its transpose; denote t the k -th control quantity in the optimized variable sequence at ; Represents the sequence of optimization variables U The dimensionality of the space where it is located; Is the t State variable of the grid-connected industrial power grid system at the moment, and H 、 F 、 Y 、 G 、 E Are constant parameter matrices and can be obtained through Q 、 R and the above formula.

[0045] Solving this quadratic programming problem can obtain the optimal active power outputs of the self-provided thermal power units and electrolytic aluminum loads, and can quickly suppress the power fluctuations of the interconnected power grid tie lines. However, this process involves a large amount of calculation and takes a lot of time. Therefore, an explicit model predictive controller is further used for control.

[0046] 3. According to the multi-parameter theory, transform the optimization objective function into a multi-parameter quadratic programming problem. By solving the multi-parameter quadratic programming problem, obtain the explicit relationship between the control variables and the state variables of the grid-connected industrial power grid system, and transform the model predictive controller into an explicit model predictive controller; The above quadratic programming problem must be solved through online calculation. Considering that the online solution process of model predictive control involves a large amount of calculation time, for this problem, the present invention provides a method of replacing the model predictive controller with an explicit model predictive controller, thereby improving the solution speed and enabling the controller to calculate the corresponding secondary frequency regulation output of the self-provided thermal power units and the input amount of electrolytic aluminum load power in real time according to the fluctuations of the interconnected power grid in a shorter time, specifically as follows: 3.1 Construct a multi-parameter quadratic programming problem (multiparametric quadratic program); Using variable substitution , and is the optimal value function, transform the quadratic programming problem of the discrete state space model into a standard multi-parameter quadratic programming problem, and its form is:

[0047] In the formula, Represents the cost function of this standard multi-parameter quadratic programming problem, and , where Represents the cost function of the quadratic programming problem of the above discrete state space model; S Is the coefficient matrix of the state variables in the constraint conditions, and , G, W Is a constant parameter matrix;x is the state variable of the grid-connected industrial power grid system, specifically , denotes its transpose; is a constant parameter F is the transpose matrix of denotes the constant parameter H is the inverse matrix of Given a closed set of parametric polyhedra: , and denotes the constraint boundary of the set, denotes the spatial dimension of the set; define denotes the feasible parameter domain of the multi-parametric quadratic programming problem, and the cost function is finite on this parameter domain; define any state of the grid-connected industrial power grid system, denotes the minimum value of the cost function of the multi-parametric quadratic programming problem in this state, and the optimal value corresponding to this state is .

[0048] Step 3.2. Solving the multi-parametric quadratic programming problem includes the following steps: Step 3.2.1. Define the constraint conditions: When the system is in the th state interval, it corresponds to the th set of constraint conditions of the multi-parametric quadratic programming problem, , , are constant matrices; if , then this constraint is called an effective constraint, denoted by , , to represent this effective constraint condition; if , then this constraint is called a non-effective constraint, denoted by , , to represent this non-effective constraint condition; let be the set of constraint subscripts, m denotes the number of groups of constraint conditions. For any set of effective constraint subscripts, the matrices and are and submatrices of; I The optimal partition of is , where

[0049] where denotes the set of effective constraints, Denote the set of non - active constraints; For the current state variables of a given grid - connected industrial power grid system , denote , the set is the critical region related to the set of active constraints L .

[0050] Define the linear independent constraint qualification (LICQ) condition: For a given set of active constraints L , if the rows of the coefficient matrix of the parameter variables in the corresponding constraint conditions are linearly independent, the set is called a linearly independent constraint qualification; Step 3.2.2. Solve for the initial state The corresponding set of active constraints ; 1) Obtain the minimum affine subspace containing the feasible parameter domain , whose dimension ; is the dimension of the parameter polyhedron set X ; If , an equation about the parameters defining can be found.

[0051] 2) Obtain the partition of the critical region of the feasible parameter domain , search for the cost function and an optimal value function ; As described in step 1), consider the multi - parameter quadratic programming problem, let , the feasible parameter domain is convex; the optimal value function is continuous and convex on the polyhedron, and is affine on each critical region; the cost function is continuous, convex and piece - wise quadratic on the polyhedron; As described in step 2), obtain an initial state , select as the Chebyshev ball center contained in the parameter polyhedron set X , this state is within the parameter polyhedron set X , and makes the multi - parameter quadratic programming problem feasible for ; Solve the multi - parameter quadratic programming problem to obtain the corresponding optimal value ; Since , the solution is unique, and an active constraint set outside the constraints of the multi - parameter quadratic programming problem can be uniquely determined, specifically as follows: Make For the center of the largest ball contained in the set X , solve the following linear programming problem (LP problem):

[0052]

[0053] where is the number of rows of the matrix T , is the decision variable, is the coefficient matrix of the parameter vector e in the x th constraint condition of this linear programming problem, and is its vector norm; if for such a , the multi-parameter quadratic programming problem is feasible, then is the Chebychev ball center of the set X ; if , for all X within the set x , the multi-parameter quadratic programming problem is infeasible; otherwise, determine the parameter .

[0054] Solve the multi-parameter quadratic programming problem to obtain the corresponding optimal solution ; since , the solution is unique and can uniquely determine a valid constraint set outside the constraints of the multi-parameter quadratic programming problem; Step 3.2.3. Determine the optimal value z and the affine relationship between x ; When this step is executed for the first time, let , let , consider an effective constraint combination , and assume that it satisfies the linear independent regularization constraint LICQ condition; on the corresponding critical domain , uniquely define the explicit affine function of z according to the optimal value and the related Lagrange multiplier vector x ; the process of determining this explicit relationship is as follows: The necessary condition for the first-order optimal solution (KKT condition) of the multi-parameter quadratic programming problem is expressed as:

[0055] where is the Lagrange multiplier,G , W , S is a constant matrix, , , is the parameter variable coefficient matrix, state variable coefficient matrix, and constant matrix corresponding to the i th set of constraint conditions; is the dimension of the Lagrange multiplier, is the G transpose matrix of.

[0056] Solving the above equation gives:

[0057] Complementary slackness condition:

[0058] In the formula, , are the inactive constraints and active constraints corresponding to the Lagrange multipliers respectively; for inactive constraints, ; for active constraints, , then: ; In the formula, , , are the active constraint combinations respectively, are linearly independent, exists; is an explicit affine function of x . Substituting into , the optimal value z is also an explicit affine function of x , and its form is: , denoted as ; Step 3.2.4. Determine the remaining critical region; Through , the critical region corresponding to the initial state is:

[0059] Eliminating the redundant inequalities in the above formula, the corresponding critical region is obtained as a compact expression; is a polyhedron in the X space, representing the largest set of that can keep the active constraint combination unchanged at the minimum value; once the critical region is defined, the remaining space needs to be explored and a new critical region is generated; The method for exploring the remaining space is as follows: Determine the state parameter space of the grid-connected industrial power grid system , For X the critical domain of, where , ; Let , then satisfies the following two conditions: ; ; In the formula, and are the coefficient matrix and the constant matrix of the constraint conditions corresponding to the subset , J is the parameter body included in the j th group of constraint conditions, is to determine the i th critical domain after the remaining critical domain, X is the complete set of the parameter space; Step 3.2.5. According to the principle of rolling optimization, take the first item of the optimal variable sequence U and act on the controlled object.

[0060] Judge whether it satisfies , if not, then let , return to Step 3.2.3, and use the same method to obtain the corresponding effective constraint subscript set the corresponding state partition and the control law of the corresponding critical domain ; if so, end the iteration; As Figure 6 shown, when the parameter space of the entire parameter polyhedron set X is explored, a polyhedron domain with the optimal value function of the same form is obtained. At this time, the rolling optimization process of the model predictive control has been transformed into an offline calculation. According to the current grid-connected industrial power grid system state variable x , the corresponding optimal value function can be obtained as ; Substitute the obtained optimal value function into to obtain the explicit expression of the optimal variable sequence U with respect to x :

[0061] Denote t the optimal solution corresponding to the state of the grid-connected industrial power grid system at time , take the first item and act on the controlled object, that is, the optimal output of the explicit model predictive controller is: .

[0062] It should be noted that if it is considered that there are too many divided regions, a sub-optimal explicit model controller can be adopted, that is, delete the regions where the Chebyshev radius is less than R r . This can reduce the number of divided regions to a certain extent, but may cause the controller performance to be poor in these regions. The differences are shown in Fig. 7(a), Fig. 7(b), and Fig. 7(c); Fig. 7(a) is a schematic diagram of the difference in the closed-loop simulation results between the sub-optimal explicit MPC and the explicit MPC of the thermal power unit output; Fig. 7(b) is a schematic diagram of the difference in the closed-loop simulation results between the sub-optimal explicit MPC and the explicit MPC of the electrolytic aluminum load output; Fig. 7(c) is a schematic diagram of the difference in the closed-loop simulation results between the sub-optimal explicit MPC and the explicit MPC of the active power fluctuation of the tie line.

[0063] 4. Convert the online rolling optimization process of the model predictive controller into an offline calculation; obtain the optimal explicit control law of each grid-connected industrial power grid system through offline calculation; monitor the current state of the grid-connected industrial power grid system online through the explicit model predictive controller, call the corresponding optimal explicit control law, and calculate the corresponding control quantity.

[0064] As Figure 5 shown, the explicit model predictive controller converts the online rolling optimization process of the model predictive controller into an offline calculation. According to the given constraint conditions, perform offline calculation, solve the multi-parameter quadratic programming problem described in step 3, divide the state space composed of the upper and lower limits of each state variable during the operation of the power grid system into multiple state partitions, calculate the optimal explicit control law corresponding to each state partition of the industrial power grid system, and store it in the memory. During operation, the explicit model predictive controller only needs to monitor the current state of the power grid system online x , determine the state partition to which the state belongs, call the corresponding optimal explicit control law from the memory, and then the corresponding control quantity can be calculated u ; the disturbances of the actual system w are input into the explicit model predictive controller as observable disturbances, specifically the wind power disturbance and other load disturbances, y and the output signal of the system, which is also introduced as feedback into the explicit model predictive controller, specifically the change in the frequency of the grid-connected industrial power grid system and the change in the tie line power.

[0065] During actual operation, the explicit model predictive controller only needs to execute the following online calculation steps: Step 4.1. Online measurement: Detect the current system state to obtain the current state variables of the grid-connected industrial power grid system ; Step 4.2. For the current state variable , the explicit model predictive controller performs the following steps: Step 4.2.1. Verify whether the constraint is satisfied; if the constraint is not satisfied, the explicit model predictive controller returns an error state and sets ; if the constraint is satisfied, perform Step 4.2.2; Step 4.2.2. Starting from the partition , sequentially detect whether belongs to the partition i . If , then belongs to the partition i , perform Step 4.2.3; otherwise perform Step 4.2.4; and represent the boundary condition coefficient matrix corresponding to the partition i ; Step 4.2.3. Calculate the control law and , return the status code and index indicating successful completion; do not test other partitions when returning; i and and are the optimal control law control matrices stored in the memory within the partition i ; Step 4.2.4. If does not belong to the partition i , calculate the violation term , which is the maximum value in ; if the violation term is 's minimum violation term, the explicit model predictive controller sets and ; perform Step 4.2.5; Step 4.2.5. The explicit model predictive controller increments and tests the next partition, returning to Step 4.2.2; after testing all partitions, enter Step 4.2.6; Step 4.2.6. If all partitions have passed the test and does not belong to any partition, the explicit model predictive controller obtains the coefficient matrices and from the memory, and calculates ; According to the explicit model predictive control algorithm, the explicit model predictive controller equation is obtained:

[0066] In the formula, m is the number of partitions of the explicit model predictive controller; the control quantity of the grid-connected industrial power grid system is obtained by solving the explicit model predictive controller equation.

[0067] F and G are stored in the memory in advance through offline calculation. The explicit model predictive controller only needs to detect the current state fluctuation of the interconnected power grid system in real time, index the partition to which the state belongs according to the above steps, and call the corresponding explicit control law, then the secondary frequency modulation output regulation instruction of the self-provided thermal power unit can be quickly calculated. and the active power output regulation instruction of the electrolytic aluminum load , and the two can quickly respond to suppress the power fluctuation of the tie line.

[0068] Applying the established explicit model predictive controller to the built grid-connected industrial power grid system model, the control result can be obtained, and the desired control curve can be obtained by adjusting the corresponding weight coefficient and state parameter range constraint. Figure 8 Illustrates the system change control in a simple state.

[0069] Figure 9 Illustrates a schematic physical structure diagram of an electronic device, as Figure 9 shown, the electronic device may include: a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus. The processor can call the logical instructions in the memory to execute the industrial power grid tie line power fluctuation control method based on explicit model prediction control.

[0070] In addition, when the logical instructions in the above-mentioned memory are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. And the aforementioned storage medium includes: USB flash drive, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other various media that can store program codes.

[0071] In another embodiment, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the interconnected power grid regulation method based on explicit model predictive control provided by the above-mentioned various methods.

[0072] In another embodiment, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the interconnected power grid regulation method based on explicit model predictive control provided by the above-mentioned various methods.

[0073] In another embodiment, the present invention also provides an interconnected power grid regulation system based on explicit model predictive control, including the following modules The first module is used to construct a control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the interconnected power grid tie line jointly participated by the electrolytic aluminum load and the secondary frequency modulation of the self-provided thermal power unit; The second module is used to establish a model predictive controller based on the control model, determine the state variables, disturbance variables and control variables of the grid-connected industrial power grid system, construct an optimization objective function corresponding to the model predictive control optimization problem of the grid-connected industrial power grid system, and convert it into a quadratic programming problem; The third module is used to convert the optimization objective function into a multi-parameter quadratic programming problem according to the multi-parameter theory. By solving the multi-parameter quadratic programming problem, an explicit relationship between the control variable and the state variable of the grid-connected industrial power grid system is obtained, and the model predictive controller is converted into an explicit model predictive controller; The fourth module is used to convert the online rolling optimization process of the model predictive controller into an offline calculation; obtain the optimal explicit control law of each grid-connected industrial power grid system through offline calculation; online monitor the current state of the grid-connected industrial power grid system through the explicit model predictive controller, call the corresponding optimal explicit control law, and calculate the corresponding control quantity.

[0074] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0075] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0076] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An interconnected power grid regulation method based on explicit model predictive control, characterized in that Including: Step 1. Construct a control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the interconnection grid tie line jointly participated by the electrolytic aluminum load and the secondary frequency regulation of the self-provided thermal power unit; Step 2. Based on the control model, establish a model predictive controller, including determining the state variables, disturbance variables, and control variables of the grid-connected industrial power grid system, constructing the corresponding optimization objective function of the model predictive control optimization problem of the grid-connected industrial power grid system, and converting it into a quadratic programming problem; Step 3. According to the multi-parameter theory, convert the optimization objective function into a multi-parameter quadratic programming problem. By solving the multi-parameter quadratic programming problem, obtain the explicit relationship between the control variables and the state variables of the grid-connected industrial power grid system, and convert the model predictive controller into an explicit model predictive controller; Step 4. Convert the online rolling optimization process of the model predictive controller into an offline calculation; obtain the optimal explicit control law of each grid-connected industrial power grid system through offline calculation; monitor the current state of the grid-connected industrial power grid system online through the explicit model predictive controller, call the corresponding optimal explicit control law, and calculate the corresponding control quantity.

2. The interconnected power grid regulation method based on explicit model predictive control according to claim 1, wherein The control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the interconnection grid tie line includes a first area control model and a second area control model; the first area control model includes self-provided thermal power unit control, electrolytic aluminum load control, self-provided wind power unit control, and other load disturbance control, and the second area control model is large power grid control. The first area control model and the second area control model are connected through a tie line; Frequency change amount of grid-connected industrial power grid system Expressed as where M is the system equivalent inertia coefficient, D is the system equivalent damping coefficient, is the governor - turbine dynamic model, R is the primary frequency regulation droop characteristic coefficient, is the secondary frequency regulation power adjustment command of the thermal power unit, is the active power output adjustment command of the electrolytic aluminum load, is the disturbance quantity of other loads; represents the differential of the system frequency change to the tie - line power change, s is the complex frequency, is the dynamic response time constant of the electrolytic aluminum load; Change in tie-line power Denoted as 。 3. The interconnected power grid regulation method based on explicit model predictive control according to claim 2, characterized in that The electrolytic aluminum load control realizes the regulation of the active power of the electrolytic aluminum load by adjusting the control current of the electrolytic aluminum load saturation reactor.

4. The interconnected power grid regulation method based on explicit model predictive control according to claim 2, wherein The implementation process of establishing the model predictive controller based on the control model includes the following sub-steps: Step 2.

1. The implementation methods for determining the state variables, disturbance variables, and control variables of the grid-connected industrial power grid system are as follows, Build a secondary frequency modulation controller for thermal power units and an electrolytic aluminum load controller, and use the state variables of the grid-connected industrial power grid system , wind power disturbances and other load disturbance quantities are uploaded to the model predictive controller through the wide-area measurement system. The model predictive controller calculates the secondary frequency modulation power adjustment command for thermal power units , the active power output adjustment command for electrolytic aluminum loads , and sends them to the electrolytic aluminum load controller and the secondary frequency modulation controller for thermal power units; where is the change in boiler active power, is the change in governor active power, is the change in the frequency of the grid-connected industrial power grid system, is the change in tie-line power; Step 2.

2. The implementation method for constructing the corresponding optimization objective function of the model predictive control optimization problem of the grid-connected industrial power grid system is as follows, The model predictive controller considering the feasible region of electrolytic aluminum load regulation adopts a discrete form, and its discrete state space equation is obtained. Substitute t the predicted k control variables , …, into it to obtain t the predicted state of the grid-connected industrial power grid system at . Among them, is the secondary frequency modulation power regulation command of the thermal power unit and the active power output regulation command of the electrolytic aluminum load as the control variables of the model predictive controller; aiming to minimize the system frequency deviation and the tie-line power change while reducing the electrolytic aluminum load response, an optimization objective function of the model predictive control is constructed. Step 2.

3. The implementation method for converting the optimization objective function of the model predictive control into a quadratic programming problem is as follows, If there exists an optimal variable sequence , U including variables, satisfying , then the optimization objective function of the model predictive control constructed in step 2.2 can be transformed into a constrained finite-time optimal control function; where represents the th control variable in the optimal variable sequence, acting at time, is t the state variable of the grid-connected industrial power grid system at ; Substitute t the predicted state variables of the grid-connected industrial power grid system at the moment into the above-mentioned constrained finite-time optimal control function for substitution, and a quadratic programming problem of a discrete state space model is obtained.

5. The interconnected power grid regulation method based on explicit model predictive control according to claim 4, characterized in that The conversion of the model predictive controller into an explicit model predictive controller includes the following steps, Step 3.

1. The implementation method for constructing the multi-parameter quadratic programming problem is as follows, Use variable substitution to transform the quadratic programming problem of the discrete state-space model into a standard multi-parameter quadratic programming problem; given a closed parametric polyhedron set X , define to represent the feasible parameter domain of the multi-parameter quadratic programming problem, and the cost function on this parameter domain is finite; define any grid-connected industrial power grid system state , to represent the minimum value of the cost function of the multi-parameter quadratic programming problem in this state, and the optimal value corresponding to this state is ; Step 3.

2. The steps for solving the multi-parameter quadratic programming problem are as follows, Step 3.2.

1. The implementation method for defining the constraint conditions is as follows, When the system is in the i th state interval, it corresponds to the i th set of constraint conditions of the multi-parameter quadratic programming problem , , are constant matrices. If , then this constraint is an effective constraint, denoted by , , ; if , then this constraint is a non-effective constraint, denoted by , , ; let be the set of constraint subscripts, m represents the number of sets of constraint conditions. For any set of effective constraint subscripts , the matrices and represent the submatrices of and respectively; for a given current state quantity of a grid-connected industrial power grid system, the set represents the critical domain related to the set of effective constraint subscripts L . Step 3.2.

2. Solve the initial state The corresponding set of valid constraint subscripts is implemented as follows: Consider the multi-parameter quadratic programming problem, and let the constant parameter matrix , and the feasible parameter domain is convex; the optimal value function is continuous and convex on the polyhedron, and is affine on each critical domain; the cost function is continuous, convex and piecewise quadratic on the polyhedron; Obtain the initial state of a grid-connected industrial power grid system , select as the Chebyshev ball center contained in the parametric polyhedron set X , the initial state is located within the parametric polyhedron set X , and make the multi-parameter quadratic programming problem feasible; Solve the multi-parameter quadratic programming problem to obtain the corresponding optimal value , since , the solution is unique and can uniquely determine a valid constraint subscript set outside the constraints of the multi-parameter quadratic programming problem ; Step 3.2.

3. Determine the optimal value and x has an affine relationship. The implementation method is as follows When this step is executed for the first time, let , , consider a set of effective constraint subscripts , and assume that the linear independent return method specification constraint LICQ condition is satisfied; on the corresponding critical domain of the set of effective constraint subscripts , solve the optimal value according to the necessary condition KKT condition of the first-order optimal solution of the multi-parameter quadratic programming problem, and find the explicit affine function of x , denoted as the optimal value function ; Step 3.2.

4. The implementation method for determining the remaining critical domain is as follows, Through the optimal value function Obtain the critical region expression corresponding to the initial state, eliminate the redundant inequalities therein, and obtain the effective constraint subscript set The corresponding critical region A compact expression; the critical region Is a polyhedron in the parameter polyhedron set X Space, representing The largest set that can keep the effective constraint subscript set Invariant at the minimum value; once the critical region Is defined, the remaining space Needs to be explored and a new critical region is generated; Step 3.2.

5. According to the rolling optimization principle, take the first item of the optimized change amount sequence U and act on the controlled object; judge whether it satisfies i = m , if not, then let i = i +1, return to Step 3.2.3, and obtain the corresponding effective constraint subscript set the corresponding state partition and the corresponding critical region of the control law; if so, end the iteration; When exploring the entire parametric polyhedron set X of the parameter space, a polyhedron domain with the optimal value function of the same form is obtained; at this time, the rolling optimization process of model predictive control has been transformed into offline calculation, and according to the current grid-connected industrial power grid system state variables x of the system, the corresponding optimal value function is obtained; Substitute the obtained optimal value function into to obtain the explicit expression of the optimized variable sequence U with respect to x : , where H , F , G , W , S are constant parameter matrices; denote the optimal solution corresponding to the state t of the grid-connected industrial power grid system at time as , and take the first term as the optimal output of the explicit model predictive controller to act on the controlled object.

6. The interconnected power grid regulation method based on explicit model predictive control according to claim 5, characterized in that The implementation method of Step 4 is as follows, Perform offline calculation according to the given constraint conditions, consider the multi-parameter theory to construct a multi-parameter quadratic programming problem, divide the state space of the grid-connected industrial power grid system into multiple state partitions, solve the multi-parameter quadratic programming problem to obtain the optimal control law corresponding to each state partition, and store it in the memory. Calculate the optimal output of the explicit model predictive controller of the grid-connected industrial power grid system under different states according to this optimal control law; According to the explicit model predictive control algorithm, obtain the explicit model predictive controller equation In the formula, m is the number of partitions of the explicit model predictive controller; the control quantity of the grid-connected industrial power grid system is obtained by solving the explicit model predictive controller equation.

7. A system for implementing the interconnected power grid regulation method based on explicit model predictive control according to any one of claims 1-6, characterized in that, Including the following modules, The first module is used to construct a control model for the frequency fluctuation of the grid-connected industrial power grid system and the power fluctuation of the interconnection grid tie line, in which the electrolytic aluminum load and the secondary frequency regulation of the self-provided thermal power unit participate together; The second module is used to establish a model predictive controller based on the control model, determine the state variables, disturbance variables and control variables of the grid-connected industrial power grid system, construct an optimization objective function corresponding to the model predictive control optimization problem of the grid-connected industrial power grid system, and convert it into a quadratic programming problem; The third module is used to convert the optimization objective function into a multi-parameter quadratic programming problem according to the multi-parameter theory. By solving the multi-parameter quadratic programming problem, an explicit relationship between the control variable and the state variable of the grid-connected industrial power grid system is obtained, and the model predictive controller is converted into an explicit model predictive controller; The fourth module is used to convert the online rolling optimization process of the model predictive controller into an offline calculation; the optimal explicit control law of each grid-connected industrial power grid system is obtained through the offline calculation; the current state of the grid-connected industrial power grid system is monitored online through the explicit model predictive controller, and the corresponding optimal explicit control law is called to calculate the corresponding control quantity.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the interconnected grid regulation method based on explicit model predictive control according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the interconnected grid regulation method based on explicit model predictive control according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the interconnected grid regulation method based on explicit model predictive control according to any one of claims 1 to 6.

Images