A method and system for dispatching industrial microgrids containing flexible electrolytic aluminum loads.

Abstract

This invention belongs to the technical field of electrical engineering and discloses a scheduling method and system for an industrial microgrid containing flexible electrolytic aluminum loads. The scheduling method uses minimizing the operating cost of the industrial microgrid system as the objective function, establishing a scheduling model for the industrial microgrid containing flexible electrolytic aluminum loads. This model is reconstructed as a Markov decision for single-time-period solution. By employing piecewise linear convex functions, the three state variable sequences—liquid layer temperature, solid layer temperature, and cumulative production—in the Markov decision process are approximated by value functions, thus obtaining a three-dimensional approximate value function. The slope of this value function can be iteratively updated under randomly generated training scenarios. The Bellman equation is solved to obtain an approximate globally optimal decision for industrial microgrid scheduling. This invention effectively addresses the uncertainty of new energy output while satisfying the operational constraints of the industrial microgrid and flexible electrolytic aluminum loads, providing an optimal scheduling scheme for industrial microgrids under stochastic scenarios.

Description

Technical Field

[0001] This invention belongs to the technical field of electrical engineering, and more specifically, relates to a method and system for dispatching an industrial microgrid containing a flexible electrolytic aluminum load. Background Technology

[0002] With the rapid development of renewable energy and the increase in electricity demand, the importance of industrial microgrids in energy management is becoming increasingly prominent. Especially in energy-intensive industries such as electrolytic aluminum production, optimizing load management and dispatch to improve energy efficiency, reduce costs, and decrease carbon emissions has become an urgent problem to be solved. During the electrolytic aluminum production process, the electricity demand of the electrolytic cells is enormous and continuous; therefore, rational dispatching of power load is crucial for reducing production costs and improving economic efficiency.

[0003] Since electrolytic aluminum is a flexible load, the optimization scheduling model that takes into account the randomness of the load and renewable energy is very complex, takes a long time to solve, and sometimes it is difficult to achieve the optimal scheduling result.

[0004] Therefore, how to make scheduling decisions for industrial microgrids containing flexible electrolytic aluminum loads and obtain optimal scheduling results is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] In view of the above-mentioned defects or improvement needs of the existing technology, the present invention provides an industrial microgrid scheduling method and system with flexible electrolytic aluminum load, the purpose of which is to make scheduling decisions for industrial microgrids with flexible electrolytic aluminum load and obtain the optimal scheduling result.

[0006] To achieve the above objectives, the present invention provides a method for dispatching an industrial microgrid containing a flexible electrolytic aluminum load, comprising:

[0007] S1. An industrial microgrid scheduling model with flexible electrolytic aluminum load is established with the objective function of minimizing the operating cost of the industrial microgrid. The cost includes the operating cost of conventional units, the cost of purchasing electricity, and the cost of curtailing renewable energy. The model satisfies the grid operation constraints and the flexible electrolytic aluminum load regulation constraints.

[0008] S2, reconstruct the model into a single-time Markov decision; where the Markov decision at time t has the set of state variables S at time t. t Decision variable set x t Uncertain information set W t And the Bellman equation:

[0009] S tIncludes the liquid layer temperature, solid layer temperature, and cumulative output of the flexible electrolytic aluminum load at time t, as well as the maximum output of new energy sources and the predicted output of fixed loads at time t, the active power output of conventional units at the previous time, and the electricity purchase price of the microgrid from the external grid at time t; x t Includes the active power output of conventional generating units at time t, the electricity purchased by the microgrid from the external grid, the operating current of the flexible electrolytic aluminum load, and the power curtailment of renewable energy; W t This includes the prediction errors of renewable energy output at time t, fixed load, and electricity purchase price; the Bellman equation is V t =min{C t (S t ,x t )+V t x}, where C t (S t ,x t Let V be the operating cost of the industrial microgrid system at time t. t Let V be the value function before the decision at time t. t x Let be the value function after the decision at time t, and be equal to the sum of piecewise linear convex functions of the value functions of liquid layer temperature, solid layer temperature, and cumulative production after the decision at time t;

[0010] S3 obtains the optimal scheduling value of the decision variable set by iteratively updating the slope of each piecewise linear convex function under randomly generated training scenarios and solving the Bellman equation.

[0011] In one embodiment, the model is:

[0012]

[0013] In the formula, F is the system operating cost of the industrial microgrid, T is the total scheduling time, and C is the total system operating cost of the industrial microgrid. G For conventional unit operating costs, C grid For the electricity purchase cost of microgrids, C cur t represents the cost of curtailing wind and solar power, and t represents the time index.

[0014] In one embodiment, the grid operation constraints include power balance constraints, upper and lower limits of conventional unit output, conventional unit ramp rate constraints, upper limit of renewable energy curtailment power, and upper and lower limits of power exchange between microgrids and external grids; the flexible electrolytic aluminum load regulation constraints include initial temperature constraints of liquid and solid layers, upper and lower limits of operating current, upper and lower limits of operating current adjustment, upper and lower limits of liquid layer temperature, upper and lower limits of solid layer temperature, and electrolytic aluminum production constraints.

[0015] The initial temperature constraints of the liquid and solid layers of the flexible electrolytic aluminum load include:

[0016]

[0017]

[0018] In the formula, a1 to a6 are the calculated coefficients, and H e For the thermal conductivity between the solid layer and the liquid layer, H s For the thermal conductivity between the chamber air and the solid layer, C s For the thermal mass of the solid layer, C e k is the thermal mass of the liquid layer. e The current coefficient;

[0019] The initial temperature of the liquid layer under the flexible electrolytic aluminum load. The initial temperature of the solid layer under the flexible electrolytic aluminum load. The rated operating current for flexible electrolytic aluminum loads.

[0020] In one embodiment, the upper and lower limits of the current adjustment for the flexible electrolytic aluminum load are:

[0021]

[0022] In the formula, These represent the upward and downward adjustment currents of the flexible electrolytic aluminum load at time t, respectively. Let I represent the upward adjustment state of the flexible electrolytic aluminum load at time t, a 0-1 variable where 0 indicates a downward adjustment of the current and 1 indicates an upward adjustment of the current. EA,min I EA,max It is divided into the minimum operating current and the maximum operating current of flexible electrolytic aluminum load.

[0023] In one embodiment, the electrolytic aluminum production constraint of the flexible electrolytic aluminum load is:

[0024]

[0025] In the formula, Y EA,t Y represents the cumulative output of the flexible electrolytic aluminum load at time t. EA,t-Δt The cumulative output of the flexible electrolytic aluminum load at time t-Δt, where Δt is the length of a single time period. Let K be the operating current of the flexible electrolytic aluminum load at time t. I2P Y is the conversion factor from operating current to power for flexible electrolytic aluminum loads. EA,end To schedule the cumulative output of flexible electrolytic aluminum load at the end of the time period. The rated operating current of the flexible electrolytic aluminum load is T, and the total number of scheduling periods is T.

[0026] In one embodiment, it is characterized in that,

[0027] The power balance constraint is:

[0028]

[0029] In the formula, P G,t Let P be the active power output of the conventional unit at time t. grid,t Let P be the amount of electricity purchased by the microgrid from the external grid at time t, where purchasing is positive and selling is negative. W,t Contribute to the prediction of new energy sources at time t. Let P be the power of renewable energy curtailment at time t. EA,t Let P be the active power of the flexible electrolytic aluminum load at time t. D,t The predicted output of the fixed load at time t;

[0030] The upper and lower limits of the output of the conventional generating units are constrained as follows:

[0031] P G,min ≤P G,t ≤P G,max

[0032] In the formula, P G,min P G,max These are the minimum and maximum active power outputs of conventional generating units, respectively.

[0033] The conventional unit ramp rate constraint is:

[0034]

[0035] In the formula, This is the downhill / climbing rate for a conventional unit. For conventional units, the ramp rate is P. G,t-Δt The active power output of the conventional unit at time t-Δt is the length of a single time period;

[0036] The upper limit constraint on the amount of abandoned renewable energy is:

[0037]

[0038] The upper and lower limits of power exchange between the microgrid and the external power grid are:

[0039] P grid,min ≤P grid,t ≤P grid,max

[0040] In the formula, P grid,min P represents the minimum power exchanged between the microgrid and the external power grid. grid,max This represents the maximum power exchanged between the microgrid and the external power grid.

[0041] The upper and lower limits of the operating current of the flexible electrolytic aluminum load are constrained as follows:

[0042]

[0043] In the formula, Let I be the operating current of the flexible electrolytic aluminum load at time t. EA,min For the minimum operating current of flexible electrolytic aluminum load, I EA,max This is the maximum operating current for flexible electrolytic aluminum loads;

[0044] The upper and lower limits of the liquid layer temperature of the flexible electrolytic aluminum load are constrained as follows:

[0045]

[0046] In the formula, T t e Let t be the temperature of the liquid layer under the flexible electrolytic aluminum load. This is the lowest temperature of the liquid layer in the flexible electrolytic aluminum load. This is the highest temperature of the liquid layer under the flexible electrolytic aluminum load.

[0047] The upper and lower limits of the solid layer temperature of the flexible electrolytic aluminum load are constrained as follows:

[0048]

[0049] In the formula, T t s Let t be the temperature of the solid layer under the flexible electrolytic aluminum load. This is the lowest temperature of the solid layer supporting the flexible electrolytic aluminum load. This represents the highest temperature of the solid layer under the load of flexible electrolytic aluminum.

[0050] In one embodiment, in step S3, the method for generating training scenarios includes: considering that the prediction error of new energy output follows a normal distribution, forming a series of random training scenarios of new energy day-ahead prediction values;

[0051] Iteratively update the slope of each piecewise linear convex function, including:

[0052] First, calculate the slope of each segment in the piecewise linear convex function using the following formula:

[0053]

[0054] In the formula, Let i be the i-th state variable after the decision at time t in the nth iteration, where i = 1, 2, 3. The three state variables after the decision are the liquid layer temperature, etc. solid layer temperature and cumulative production In the nth iteration The slope sample value of the k-th segment in the corresponding piecewise linear convex function. In the nth iteration Compared to the increase in the previous moment, In the nth iteration and An approximate function of the sum. In the nth iteration Approximate function, Let i be the i-th state variable after the decision at time t-Δt in the nth iteration. In the nth iteration The slope value of the k-th segment of the corresponding piecewise linear convex function, θ n-1 To update the step size for the slope, In the (n-1)th iteration The slope value of the k-th segment of the corresponding piecewise linear convex function;

[0055] Then, select any calculated slope of the k-th segment as a reference, and adjust the slope according to the following formula:

[0056]

[0057] In the formula, the subscript y is the index of the segment;

[0058] Substitute the adjusted slope into the piecewise linear convex function and solve the Bellman equation.

[0059] This invention also provides an industrial microgrid dispatching system containing flexible electrolytic aluminum loads, comprising:

[0060] The model building module is used to establish an industrial microgrid dispatch model containing flexible electrolytic aluminum load with the objective function of minimizing the operating cost of the industrial microgrid. The cost includes the operating cost of conventional units, the cost of purchasing electricity, and the cost of curtailing renewable energy. The model satisfies the grid operation constraints and the flexible electrolytic aluminum load regulation constraints.

[0061] The reconstruction module is used to reconstruct the model into a single-time Markov decision; wherein the Markov decision at time t has a set S of state variables at time t. t Decision variable set x t Uncertain information set W t And Bellman equation: S t Includes the liquid layer temperature, solid layer temperature, and cumulative output of the flexible electrolytic aluminum load at time t, as well as the maximum output of new energy sources and the predicted output of fixed loads at time t, the active power output of conventional units at the previous time, and the electricity purchase price of the microgrid from the external grid at time t; x tIncludes the active power output of conventional generating units at time t, the electricity purchased by the microgrid from the external grid, the operating current of the flexible electrolytic aluminum load, and the power curtailment of renewable energy; W t This includes the prediction errors for new energy output at time t, the prediction errors for fixed load, and the prediction errors for electricity purchase price; the Bellman equation is... In the formula, C t (S t ,x t Let V be the operating cost of the industrial microgrid system at time t. t Let V be the value function before the decision at time t. t x Let be the value function after the decision at time t, and be equal to the sum of piecewise linear convex functions of the value functions of liquid layer temperature, solid layer temperature, and cumulative production after the decision at time t;

[0062] The solution module is used to obtain the optimal scheduling value of the decision variable set by iteratively updating the slope of each piecewise linear convex function and solving the Bellman equation under randomly generated training scenarios.

[0063] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any of the preceding claims.

[0064] The present invention also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the method described in any of the preceding claims.

[0065] In summary, compared with the prior art, the technical solutions conceived in this invention have the following main advantages:

[0066] 1. This invention establishes a real-time scheduling model for an industrial microgrid containing a flexible electrolytic aluminum load, with the objective function of minimizing the operating cost of the industrial microgrid system. The model is reconstructed into a Markov decision-making process for single-time-period solution, thereby reducing the complexity of the solution. By employing piecewise linear convex functions, during the Markov decision-making process, the three state variable sequences of the flexible electrolytic aluminum load—liquid layer temperature, solid layer temperature, and cumulative production—are approximated by value functions. The piecewise linear convex functions of the three state variables are superimposed to obtain a three-dimensional approximate value function. This three-dimensional approximate value function is used as the value function in the Bellman equation of the Markov decision-making process. The slope of this value function can be iteratively updated under randomly generated training scenarios, thereby reducing the curse of dimensionality caused by the Markov decision-making process traversing the state variable space and decision variable space, and decreasing the computational load. In this invention, by analyzing each state variable, the coupling relationship between state variables in the state variable sequence is fully considered. Since the flexible electrolytic aluminum model is represented by differential equations, the liquid layer temperature and solid layer temperature, two state variables, have a time-dependent coupling relationship. The cumulative output must reach its rated value at the end of the scheduling process, which also has a time-dependent coupling relationship. Therefore, selecting the liquid layer temperature, solid layer temperature, and cumulative output, three state variable sequences, for value function approximation respectively ensures the stability and effectiveness of the iterative training of solving the time-valued function, guaranteeing the final iterative convergence. Finally, solving the Bellman equation yields an approximate globally optimal decision for industrial microgrid scheduling. This invention not only reduces the operating cost of the industrial microgrid system but also ensures the optimization accuracy under stochastic scenarios. Under the premise of satisfying the operating constraints of the industrial microgrid and flexible electrolytic aluminum load, it effectively addresses the uncertainty of new energy output and provides an optimal scheduling scheme for industrial microgrids under stochastic scenarios.

[0067] 2. In a further embodiment, specific grid operation constraints and flexible electrolytic aluminum load regulation constraints are given. This series of constraints can ensure the safe operation of industrial microgrids containing flexible electrolytic aluminum loads.

[0068] 3. In a further embodiment, upper and lower limits of current adjustment for electrolytic aluminum load are given. By decomposing it into upward and downward current and introducing 0-1 variables, the upward and downward adjustment of current can be accurately limited.

[0069] 4. In a further embodiment, an iterative update method for the slope of a piecewise linear convex function is given. First, the slope of each segment is calculated using an iterative formula, and then the slope is adjusted. This ensures that the piecewise linear function remains a convex function after each iterative update. Attached Figure Description

[0070] Figure 1 This is a flowchart of the steps of an industrial microgrid scheduling method with flexible electrolytic aluminum load according to an embodiment of the present invention;

[0071] Figure 2 This is a random training scenario for the day-ahead wind power prediction value in one embodiment of the present invention;

[0072] Figure 3 This is an example of an optimal flexible electrolytic aluminum load scheduling plan obtained using the scheduling method proposed in this invention;

[0073] Figure 4 This is the optimal flexible electrolytic aluminum load liquid layer temperature obtained by using the scheduling method proposed in this invention in one embodiment;

[0074] Figure 5 This is a schematic diagram of the structure of an industrial microgrid dispatching system according to an embodiment of the present invention. Detailed Implementation

[0075] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0076] Example 1

[0077] like Figure 1 The diagram shows a flowchart of the industrial microgrid scheduling method with flexible electrolytic aluminum load according to an embodiment of the present invention, which mainly includes steps S1 to S3. Each step is described below.

[0078] S1 establishes an industrial microgrid dispatch model with flexible electrolytic aluminum load as the objective function, with the minimum operating cost of the industrial microgrid as the objective function. The cost includes the operating cost of conventional units, the cost of purchasing electricity and the cost of curtailing renewable energy. The model satisfies the grid operation constraints and the flexible electrolytic aluminum load regulation constraints.

[0079] Understandably, before modeling, it is necessary to collect equipment parameters, load data, new energy data, and flexible electrolytic aluminum load parameters of the industrial microgrid to be scheduled.

[0080] Equipment parameters include: the number of conventional units. Rated capacity P of conventional units base,G The maximum output P of conventional units max,G Minimum output P of conventional units min,G Conventional unit downhill ramp rate Conventional unit ramp rate

[0081] Load data includes: the day-ahead forecast load power curve P D .

[0082] New energy data includes: the previously predicted new energy output curve P. W .

[0083] The load parameters for flexible electrolytic aluminum include: the highest temperature of the liquid layer. Lowest temperature of the liquid layer The highest temperature of the solid layer Minimum temperature of solid layer Maximum operating current I EA,max Minimum operating current I EA,min The working voltage V of the electrolytic aluminum load base Number of electrolytic cells M EA Electrolytic aluminum load operating current to power conversion coefficient K I2P Rated operating current of electrolytic aluminum load

[0084] In one embodiment, the real-time dispatch model for an industrial microgrid considering a flexible electrolytic aluminum load is as follows:

[0085]

[0086] Where F is the system operating cost of the industrial microgrid, T is the total scheduling time, and C is the total system operating cost of the industrial microgrid. G For conventional unit operating costs, C grid For the electricity purchase cost of microgrids, C cur t represents the cost of curtailing wind and solar power, and t represents the time index.

[0087] Meanwhile, the model also needs to meet the constraints of power grid operation and the load regulation constraints of flexible electrolytic aluminum.

[0088] Grid operation constraints can be related to conventional generating units, purchased electricity, and curtailment of renewable energy. For example, they can include power balance constraints, upper and lower limits of conventional generating unit output, conventional generating unit ramp rate constraints, upper limit of renewable energy curtailment power, and upper and lower limits of power exchange between microgrids and the external grid. Examples of each constraint are given below.

[0089] The power balance constraint is:

[0090]

[0091] In the formula, P G,t Let P be the active power output of the conventional unit at time t. grid,t Let P be the amount of electricity purchased by the microgrid from the external grid at time t, where purchasing is positive and selling is negative. W,t Contribute to the prediction of new energy sources at time t. Let P be the power of renewable energy curtailment at time t. EA,tLet P be the active power of the flexible electrolytic aluminum load at time t. D,t This represents the predicted output of the fixed load at time t.

[0092] The upper and lower limits of the output of conventional generating units are constrained as follows:

[0093] P G,min ≤P G,t ≤P G,max

[0094] In the formula, P G,min P G,max These are the minimum and maximum active power outputs of conventional generating units, respectively.

[0095] The conventional unit ramp rate constraint is:

[0096]

[0097] In the formula, This is the downhill / climbing rate for a conventional unit. For conventional units, the ramp rate is P. G,t-Δt The active power output of the conventional unit at time t-Δt is the length of a single time period.

[0098] The upper limit constraint on the curtailment power of renewable energy is:

[0099]

[0100] In the formula, P W,t Contribute to the prediction of new energy sources at time t. Let t be the power of renewable energy curtailment.

[0101] The upper and lower limits of power exchange between the microgrid and the external power grid are:

[0102] P grid,min ≤P grid,t ≤P grid,max

[0103] In the formula, P grid,min P represents the minimum power exchanged between the microgrid and the external power grid. grid,max This represents the maximum power exchanged between the microgrid and the external power grid.

[0104] The load regulation constraint for flexible electrolytic aluminum is a constraint on the operation of the electrolytic aluminum load. The integrated thermal parameter model of the flexible electrolytic aluminum load can be represented by the following set of differential equations:

[0105]

[0106]

[0107] Where τ is the time index, a is the coefficient, and Ta T represents the temperature of the air in the chamber. s T is the temperature of the solid layer. e H represents the temperature of the liquid layer. s For the thermal conductivity between the chamber air and the solid layer, H e For the thermal conductivity between the solid layer and the liquid layer, C s For the thermal mass of the solid layer, C e k is the thermal mass of the liquid layer. e I is the current coefficient. e The current intensity that generates heat.

[0108] In one embodiment, the load regulation constraints for flexible electrolytic aluminum include initial temperature constraints of the liquid layer and solid layer, upper and lower limits constraints of the operating current, upper and lower limits constraints of the operating current adjustment, upper and lower limits constraints of the liquid layer temperature, upper and lower limits constraints of the solid layer temperature, and electrolytic aluminum production constraints. Examples of each constraint are given below.

[0109] First, initial temperature constraints for the liquid and solid layers are constructed based on an integrated thermal parameter model. The initial temperatures of the liquid and solid layers can be directly calculated based on these constraints.

[0110] The initial temperature constraints for the liquid and solid layers are as follows:

[0111]

[0112]

[0113] In the formula, a1 to a6 are the calculated coefficients, and H e For the thermal conductivity between the solid layer and the liquid layer, H s For the thermal conductivity between the chamber air and the solid layer, C s For the thermal mass of the solid layer, C e k is the thermal mass of the liquid layer. e For current coefficient, The initial temperature of the liquid layer under the flexible electrolytic aluminum load. The initial temperature of the solid layer under the flexible electrolytic aluminum load. The rated operating current for flexible electrolytic aluminum loads.

[0114] The upper and lower limits of current adjustment for flexible electrolytic aluminum loads are as follows:

[0115]

[0116] In the formula, These represent the upward and downward adjustment currents of the flexible electrolytic aluminum load at time t, respectively. Let I represent the upward adjustment state of the flexible electrolytic aluminum load at time t, a 0-1 variable where 0 indicates a downward adjustment of the current and 1 indicates an upward adjustment of the current.EA,min I EA,max It is divided into the minimum operating current and the maximum operating current of flexible electrolytic aluminum load.

[0117] The upper and lower limits of the operating current of the flexible electrolytic aluminum load are constrained as follows:

[0118]

[0119] In the formula, Let I be the operating current of the flexible electrolytic aluminum load at time t. EA,min For the minimum operating current of flexible electrolytic aluminum load, I EA,max This is the maximum operating current for flexible electrolytic aluminum loads.

[0120] The upper and lower limits of the liquid layer temperature for flexible electrolytic aluminum loads are:

[0121]

[0122] In the formula, T t e Let t be the temperature of the liquid layer under the flexible electrolytic aluminum load. This is the lowest temperature of the liquid layer in the flexible electrolytic aluminum load. This represents the highest temperature of the liquid layer in the flexible electrolytic aluminum load.

[0123] The upper and lower limits of the solid layer temperature for flexible electrolytic aluminum loads are:

[0124]

[0125] In the formula, T t s Let t be the temperature of the solid layer under the flexible electrolytic aluminum load. This is the lowest temperature of the solid layer supporting the flexible electrolytic aluminum load. This represents the highest temperature of the solid layer under the load of flexible electrolytic aluminum.

[0126] The electrolytic aluminum production constraint for flexible electrolytic aluminum loads is:

[0127]

[0128] In the formula, Y EA,t Y represents the cumulative output of the flexible electrolytic aluminum load at time t. EA,t-Δt The cumulative output of the flexible electrolytic aluminum load at time t-Δt, where Δt is the length of a single time period. Let K be the operating current of the flexible electrolytic aluminum load at time t. I2P Y is the conversion factor from operating current to power for flexible electrolytic aluminum loads. EA,end To schedule the cumulative output of flexible electrolytic aluminum load at the end of the time period. The rated operating current of the flexible electrolytic aluminum load is T, and the total number of scheduling periods is T.

[0129] S2 reconstructs the model into a single-time Markov decision; where the Markov decision at time t has the set of state variables S at time t. t Decision variable set x t Uncertain information set W t And the Bellman equation, the Bellman equation is In the formula, C t (S t ,x t Let V be the operating cost of the industrial microgrid system at time t. t Let be the value function before the decision at time t. Let be the value function after the decision at time t, and be equal to the sum of piecewise linear convex functions of the value functions of the liquid layer temperature, solid layer temperature, and cumulative production after the decision at time t.

[0130] Making Markov decisions requires determining state variables, decision variables (i.e., action variables), and uncertain variables.

[0131] In this invention, the set of state variables S t It includes the liquid layer temperature, solid layer temperature and cumulative output of the flexible electrolytic aluminum load at time t, as well as the maximum output of new energy and the predicted output of fixed load at time t, the active power output of conventional units at the previous time, and the electricity purchase price of the microgrid from the external grid at time t.

[0132] In this embodiment, it is necessary to know the maximum power output P that the wind power can obtain at time t. W,max,t The temperature T of the liquid layer in the flexible electrolytic aluminum load at time t. t e The temperature T of the solid layer of the flexible electrolytic aluminum load at time t. t s The cumulative output Y of flexible electrolytic aluminum load at time t EA,t The magnitude of the active power P of the system's fixed load at time t D,t The active power output P of the conventional unit at the previous time t-Δt G,t-Δt At time t, the electricity price p purchased by the microgrid from the external grid t Therefore, the set of state variables S of the microgrid at time t is... t for:

[0133] S t ={P W,max,t ,T t e ,T t s ,Y EA,t ,P D,t ,P G,t-Δt,p t}

[0134] In this invention, the set of decision variables x t This includes the active power output of conventional generating units at time t, the electricity purchased by the microgrid from the external grid, the operating current of the flexible electrolytic aluminum load, and the power of abandoned renewable energy.

[0135] In this embodiment, solving the industrial microgrid scheduling problem at time t requires determining the active power output P of the conventional generating units at time t. G,t At time t, the amount of electricity P purchased by the microgrid from the external grid grid,t (Purchasing electricity is positive, selling electricity is negative), the magnitude of the operating current of the flexible electrolytic aluminum load at time t. The magnitude of the upward adjustment working current of the flexible electrolytic aluminum load at time t The magnitude of the reduced operating current of the flexible electrolytic aluminum load at time t The upward adjustment state of the flexible electrolytic aluminum load at time t System renewable energy curtailment power at time t Therefore, the set of decision variables x of the microgrid at time t t for:

[0136]

[0137] In this invention, the set of uncertain information W t Prediction error including renewable energy output at time t Prediction error of fixed load And the prediction error of electricity purchase price Uncertain information set W t for:

[0138]

[0139] In this invention, by reconstructing the Markov decision process, the Bellman equation is specifically as follows:

[0140]

[0141] Among them, C t (S t ,x t ) represents the cost function of a real-time scheduling model for an industrial microgrid containing flexible electrolytic aluminum loads. V is the set of state variables after the decision. t and These are the value functions before and after the decision, respectively.

[0142] However, in dynamic programming, obtaining the optimal decision at each time step requires traversing both the state variable space and the decision variable space, leading to a massive computational burden and the "curse of dimensionality." Therefore, this invention employs three piecewise linear convex functions to approximate the liquid layer temperature, solid layer temperature, and cumulative yield in the state variable sequence during the Markov decision process. The three piecewise linear convex functions are superimposed to obtain a three-dimensional approximate value function, which is then directly used as the value function after the decision.

[0143] In this embodiment, to address the "curse of dimensionality" problem in solving the real-time scheduling model of an industrial microgrid with flexible electrolytic aluminum load, a real-time scheduling method for an industrial microgrid with flexible electrolytic aluminum load based on approximate dynamic programming is adopted. The liquid layer temperature, solid layer temperature, and cumulative production in the state variable sequence are piecewise linearly represented as three-dimensional piecewise linear convex functions to approximate the state value function, thereby solving the Bellman equation and obtaining an approximately optimal decision set.

[0144] Specifically, the three-dimensional piecewise linear convex function is:

[0145]

[0146]

[0147] In the formula, T t e,x Let T be the temperature of the liquid layer after the decision at time t. t s,x Let t be the temperature of the solid layer after the decision at time t. Let be the cumulative output after the decision at time t. Let T be the temperature of the liquid layer after the decision at time t. t e,x Approximate function, Let T be the solid layer temperature after the decision at time t. t s,x Approximate function, The cumulative output after the decision at time t Approximate function, Let i be the i-th state variable for value function approximation, where i = 1, 2, 3. for for for k is the segment index of the piecewise linear convex function, K i for The number of segments in the corresponding piecewise linear convex function, v i,k,t Let r be the slope of the k-th piecewise linear convex function at time t. i,k Let x be the length of the k-th segment of the piecewise linear convex function mapped to the x-axis.

[0148] S3 obtains the optimal scheduling value of the decision variable set by iteratively updating the slope of each piecewise linear convex function under randomly generated training scenarios and solving the Bellman equation.

[0149] Solving the Bellman equation in the Markov decision process actually involves solving for the following decision variables:

[0150]

[0151] This invention solves the Bellman equation using an iterative method, updating the slope of each piecewise linear convex function in each iteration. After multiple iterations, V t It will gradually decrease and converge, yielding the optimal scheduling value.

[0152] Methods for updating the slope of each piecewise linear convex function include:

[0153] First, considering that the prediction error of new energy output follows a normal distribution, a series of random training scenarios for the day-ahead prediction values ​​of new energy are formed.

[0154] Specifically, considering that the wind power output prediction error follows a normal distribution, for example, satisfying W t ~N(0,0.2) 2 ), generating 600 random training scenarios using the Monte Carlo method, such as the random training scenarios for the day-ahead wind power forecast. Figure 2 As shown, different curves represent different training scenarios.

[0155] Then, the slope of each segment in the piecewise linear convex function is calculated according to the iterative formula.

[0156] Specifically, the slope is updated iteratively using 600 randomly generated training scenarios, with each iteration corresponding to one scenario. The formula used is as follows:

[0157]

[0158] In the formula, Let i be the i-th state variable after the decision at time t in the nth iteration, where i = 1, 2, 3. The three state variables after the decision are the liquid layer temperature, etc. solid layer temperature and cumulative production In the nth iteration The slope sample value of the k-th segment in the corresponding piecewise linear convex function. In the nth iteration Compared to the increase in the previous moment, In the nth iteration and An approximate function of the sum. In the nth iteration Approximate function, Let i be the i-th state variable after the decision at time t-Δt in the nth iteration. In the nth iteration The slope value of the k-th segment of the corresponding piecewise linear convex function, θ n-1 To update the step size for the slope, In the (n-1)th iteration The slope value of the k-th segment of the corresponding piecewise linear convex function.

[0159] Finally, using the slope of any calculated k-th segment as a reference, adjust the slope according to the following formula:

[0160]

[0161] In the formula, the subscript y is the index of the segment;

[0162] By substituting the three-dimensional approximation function obtained through training on random scenarios into the Bellman function in the Markov decision process and solving it time-by-time, the current running decision can be obtained, which has the characteristic of being approximately globally optimal.

[0163] like Figure 3 The figure shows an optimal flexible electrolytic aluminum load scheduling plan obtained using the scheduling method proposed in this invention in one embodiment. Figure 4 The figure shows the optimal liquid layer temperature of the flexible electrolytic aluminum load obtained by the scheduling method proposed in this invention in one embodiment. Multiple test scenarios were designed, and the optimization error of this invention was evaluated based on the solution results of MILP. The mean optimization error of this invention is on the order of 5.028e-5, which shows that the optimization effect is excellent and can obtain an approximate global optimal solution.

[0164] It should be noted that for a more detailed explanation of the working principle and procedures of this embodiment, please refer to the relevant description in Embodiment 1, but not limited to this one.

[0165] Example 2

[0166] This invention also relates to an industrial microgrid dispatching system containing flexible electrolytic aluminum loads, such as... Figure 5 The diagram shown is a structural schematic of an industrial microgrid dispatching system according to an embodiment of the present invention, which includes:

[0167] The model building module is used to establish an industrial microgrid dispatch model with flexible electrolytic aluminum load as the objective function, with the goal of minimizing the operating cost of the industrial microgrid. The cost includes the operating cost of conventional units, the cost of purchasing electricity and the cost of curtailing renewable energy. The model satisfies the grid operation constraints and the flexible electrolytic aluminum load regulation constraints.

[0168] The reconstruction module is used to reconstruct the model into a single-time Markov decision; where the Markov decision at time t has the set of state variables S at time t. t Decision variable set x t Uncertain information set W t And Bellman equation: S t Includes the liquid layer temperature, solid layer temperature, and cumulative output of the flexible electrolytic aluminum load at time t, as well as the maximum output of new energy sources and the predicted output of fixed loads at time t, the active power output of conventional units at the previous time, and the electricity purchase price of the microgrid from the external grid at time t; x t Includes the active power output of conventional generating units at time t, the electricity purchased by the microgrid from the external grid, the operating current of the flexible electrolytic aluminum load, and the power curtailment of renewable energy; W t This includes the prediction errors of renewable energy output at time t, fixed load, and electricity purchase price; the Bellman equation is V t =min{C t (S t ,x t )+V t x}, where C t (S t ,x t Let V be the operating cost of the industrial microgrid system at time t. t Let V be the value function before the decision at time t. t x Let be the value function after the decision at time t, and be equal to the sum of piecewise linear convex functions of the value functions of liquid layer temperature, solid layer temperature, and cumulative production after the decision at time t;

[0169] The solution module is used to obtain the optimal scheduling value of the decision variable set by iteratively updating the slope of each piecewise linear convex function and solving the Bellman equation under randomly generated training scenarios.

[0170] Example 3

[0171] The present invention also relates to a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0172] Specifically, the memory may include high-speed random access memory, as well as non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart media cards (SMC), secure digital cards (SD), flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0173] Example 4

[0174] This invention provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps of the method described in the above embodiments of this invention.

[0175] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification. It should be noted that the terms "in one embodiment," "for example," and "again" are intended to illustrate the present invention and are not intended to limit the present invention.

[0176] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for dispatching an industrial microgrid containing flexible electrolytic aluminum loads, characterized in that, include: S1. An industrial microgrid scheduling model with flexible electrolytic aluminum load is established with the objective function of minimizing the operating cost of the industrial microgrid. The cost includes the operating cost of conventional units, the cost of purchasing electricity, and the cost of curtailing renewable energy. The model satisfies the grid operation constraints and the flexible electrolytic aluminum load regulation constraints. S2, the model is reconstructed into a single-time Markov decision; wherein, Markov decision-making at any given moment has Set of state variables at time 1 Set of decision variables Uncertain information set And the Bellman equation: Include The liquid layer temperature, solid layer temperature, and cumulative output of flexible electrolytic aluminum at any given time, and The maximum output of new energy sources and the predicted output of fixed load at any given time, and the active power output of conventional units at the previous time. The electricity price that a microgrid purchases from the external power grid at any given time; Include The active power output of conventional generating units at any given time, the amount of electricity purchased by the microgrid from the external grid, the operating current of the flexible electrolytic aluminum load, and the amount of renewable energy curtailed. Include The prediction errors for renewable energy output, fixed load, and electricity purchase price at any given time; the Bellman equation is... In the formula, for The operating cost of industrial microgrid systems at all times for The value function before the decision time. Let be the value function after the decision at time step and equal to The sum of piecewise linear convex functions of the values ​​of liquid layer temperature, solid layer temperature, and cumulative production after the time-decision decision; S3, by iteratively updating the slope of each piecewise linear convex function under randomly generated training scenarios and solving the Bellman equation, the optimal scheduling value of the decision variable set is obtained; The grid operation constraints include power balance constraints, upper and lower limits of conventional unit output, conventional unit ramp rate constraints, upper limit of renewable energy curtailment power, and upper and lower limits of power exchange between microgrids and external grids; the flexible electrolytic aluminum load regulation constraints include initial temperature constraints of liquid and solid layers, upper and lower limits of operating current, upper and lower limits of operating current adjustment, upper and lower limits of liquid layer temperature, upper and lower limits of solid layer temperature, and electrolytic aluminum production constraints. The initial temperature constraints of the liquid and solid layers of the flexible electrolytic aluminum load include: In the formula, For the calculated coefficients, For the thermal conductivity between the solid layer and the liquid layer, For the thermal conductivity between the chamber air and the solid layer, The temperature of the air in the chamber. For the thermal mass of the solid layer, For the thermal mass of the liquid layer, The current coefficient; The initial temperature of the liquid layer under the flexible electrolytic aluminum load. The initial temperature of the solid layer under the flexible electrolytic aluminum load. The rated operating current for flexible electrolytic aluminum loads; The upper and lower limits of the current adjustment for the flexible electrolytic aluminum load are as follows: In the formula, , They are respectively Adjusting the upward and downward current of the flexible electrolytic aluminum load at all times. for The current adjustment state of the flexible electrolytic aluminum load at any time is a 0-1 variable, where 0 indicates a decrease in current and 1 indicates an increase in current. , It is divided into the minimum operating current and the maximum operating current of flexible electrolytic aluminum load.

2. The industrial microgrid dispatching method as described in claim 1, characterized in that, The model is as follows: In the formula, For the system operating cost of industrial microgrids, For the total scheduling duration, For the operating costs of conventional generating units, For the electricity purchase cost of microgrids, To account for the cost of wind and solar power curtailment For time indexing.

3. The industrial microgrid dispatching method as described in claim 1, characterized in that, The electrolytic aluminum production constraint of the flexible electrolytic aluminum load is: In the formula, for The cumulative output of flexible electrolytic aluminum under constant load. for The cumulative output of flexible electrolytic aluminum under constant load. For the length of a single time period, for The operating current of the flexible electrolytic aluminum load at all times. The conversion factor from operating current to power for flexible electrolytic aluminum loads. To schedule the cumulative output of flexible electrolytic aluminum load at the end of the time period. This is the rated operating current for the flexible electrolytic aluminum load. This represents the total number of scheduling periods.

4. The industrial microgrid dispatching method as described in claim 1, characterized in that, The power balance constraint is: In the formula, for The active power output of conventional generating units at all times. for The microgrid purchases electricity from the external grid at any given time; purchasing electricity is positive, and selling electricity is negative. for Real-time new energy power output forecast for The amount of renewable energy curtailed at all times for The active power of the flexible electrolytic aluminum load at all times. for Predicted output of fixed load at all times; The upper and lower limits of the output of the conventional generating units are constrained as follows: In the formula, , These are the minimum and maximum active power outputs of conventional generating units, respectively. The conventional unit ramp rate constraint is: In the formula, This is the downhill / climbing rate for a conventional unit. This is the ramp rate for conventional generator units. for The active power output of conventional generating units at all times. The length of a single time period; The upper limit constraint on the amount of abandoned renewable energy is: The upper and lower limits of power exchange between the microgrid and the external power grid are: In the formula, This represents the minimum power exchanged between the microgrid and the external power grid. This represents the maximum power exchanged between the microgrid and the external power grid. The upper and lower limits of the operating current of the flexible electrolytic aluminum load are constrained as follows: In the formula, for The operating current of the flexible electrolytic aluminum load at all times. The minimum operating current for flexible electrolytic aluminum load. This is the maximum operating current for flexible electrolytic aluminum loads; The upper and lower limits of the liquid layer temperature of the flexible electrolytic aluminum load are constrained as follows: In the formula, for The temperature of the liquid layer under the flexible electrolytic aluminum load at all times. This is the lowest temperature of the liquid layer in the flexible electrolytic aluminum load. This is the highest temperature of the liquid layer under the flexible electrolytic aluminum load. The upper and lower limits of the solid layer temperature of the flexible electrolytic aluminum load are constrained as follows: In the formula, for The temperature of the solid layer under the load of flexible electrolytic aluminum at all times. This is the lowest temperature of the solid layer supporting the flexible electrolytic aluminum load. This represents the highest temperature of the solid layer under the load of flexible electrolytic aluminum.

5. The industrial microgrid dispatching method as described in claim 1, characterized in that, In step S3, the method for generating training scenarios includes: considering that the prediction error of new energy output follows a normal distribution, forming a series of random training scenarios of new energy day-ahead prediction values; Iteratively update the slope of each piecewise linear convex function, including: First, calculate the slope of each segment in the piecewise linear convex function using the following formula: In the formula, For the first In the next iteration The i-th state variable after the decision at time step i, i=1,2,3, and the three state variables after the decision are the liquid layer temperature, etc. Solid layer temperature and cumulative production , For the first In the next iteration In the corresponding piecewise linear convex function, the first The slope sampling value of the segment. For the first In the next iteration Compared to the increase in the previous moment, For the first In the next iteration and An approximate function of the sum. For the first In the next iteration Approximate function, For the first In the next iteration The i-th state variable after the decision at time step For the first In the next iteration The corresponding piecewise linear convex function The slope value of the segment, Update the step size for slope. For the first In the next iteration The corresponding piecewise linear convex function The slope value of the segment; Then, select any calculated slope of the k-th segment as a reference, and adjust the slope according to the following formula: In the formula, the subscript y is the index of the segment; Substitute the adjusted slope into the piecewise linear convex function and solve the Bellman equation.

6. An industrial microgrid dispatching system containing flexible electrolytic aluminum loads, characterized in that, include: The model building module is used to establish an industrial microgrid dispatch model containing flexible electrolytic aluminum load with the objective function of minimizing the operating cost of the industrial microgrid. The cost includes the operating cost of conventional units, the cost of purchasing electricity, and the cost of curtailing renewable energy. The model satisfies the grid operation constraints and the flexible electrolytic aluminum load regulation constraints. The reconstruction module is used to reconstruct the model into a single-time Markov decision; wherein, Markov decision-making at any given moment has Set of state variables at time 1 Set of decision variables Uncertain information set And the Bellman equation: Include The liquid layer temperature, solid layer temperature, and cumulative output of flexible electrolytic aluminum at any given time, and The maximum output of new energy sources and the predicted output of fixed load at any given time, and the active power output of conventional units at the previous time. The electricity price that a microgrid purchases from the external power grid at any given time; Include The active power output of conventional generating units at any given time, the amount of electricity purchased by the microgrid from the external grid, the operating current of the flexible electrolytic aluminum load, and the amount of renewable energy curtailed. Include The prediction errors for renewable energy output, fixed load, and electricity purchase price at any given time; the Bellman equation is... In the formula, for The operating cost of industrial microgrid systems at all times for The value function before the decision time. Let be the value function after the decision at time step and equal to The sum of piecewise linear convex functions of the values ​​of liquid layer temperature, solid layer temperature, and cumulative production after the time-decision decision; The solution module is used to obtain the optimal scheduling value of the decision variable set by iteratively updating the slope of each piecewise linear convex function and solving the Bellman equation under randomly generated training scenarios. The grid operation constraints include power balance constraints, upper and lower limits of conventional unit output, conventional unit ramp rate constraints, upper limit of renewable energy curtailment power, and upper and lower limits of power exchange between microgrids and external grids; the flexible electrolytic aluminum load regulation constraints include initial temperature constraints of liquid and solid layers, upper and lower limits of operating current, upper and lower limits of operating current adjustment, upper and lower limits of liquid layer temperature, upper and lower limits of solid layer temperature, and electrolytic aluminum production constraints. The initial temperature constraints of the liquid and solid layers of the flexible electrolytic aluminum load include: In the formula, For the calculated coefficients, For the thermal conductivity between the solid layer and the liquid layer, For the thermal conductivity between the chamber air and the solid layer, The temperature of the air in the chamber. For the thermal mass of the solid layer, For the thermal mass of the liquid layer, The current coefficient; The initial temperature of the liquid layer under the flexible electrolytic aluminum load. The initial temperature of the solid layer under the flexible electrolytic aluminum load. The rated operating current for flexible electrolytic aluminum loads; The upper and lower limits of the current adjustment for the flexible electrolytic aluminum load are as follows: In the formula, , They are respectively Adjusting the upward and downward current of the flexible electrolytic aluminum load at all times. for The current adjustment state of the flexible electrolytic aluminum load at any time is a 0-1 variable, where 0 indicates a decrease in current and 1 indicates an increase in current. , It is divided into the minimum operating current and the maximum operating current of flexible electrolytic aluminum load.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

8. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method as described in any one of claims 1 to 5.

Images