Dynamic optimization method for identifying key planning decision point of integrated energy system

Abstract

The invention discloses a dynamic optimization method for identifying a key planning decision point of an integrated energy system, and relates to the field of power system scheduling and control, and the method comprises the steps: firstly, building a planning and operation combined optimization model considering the equipment operation constraint based on the physical composition of the integrated energy system; secondly, calculating and outputting the value of the current planning scheme by taking a determined equipment capacity scheme as an input parameter and taking the operation condition in a relatively long operation period as evaluation, and forming a value function model of a planning decision; then, a rolling time domain optimization model and a system rolling call value function model of the integrated energy system are constructed, the value in the current state is calculated, and a continuous value curve is obtained through linear interpolation and reverse recursion; according to the multi-stage dynamic planning method, a one-time complex investment problem is converted into a multi-stage dynamic planning scheme of the comprehensive energy system through a non-decreasing increment planning mode.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power system dispatching and control, and particularly relates to a dynamic optimization method for identifying key planning decision points of a comprehensive energy system. BACKGROUND

[0002] As an important part of energy transformation and smart grid development, the development status and future trend of park micro-energy network comprehensive energy system are of great concern. At present, electricity, heat, cold, gas and other heterogeneous energy flows in the park are coupled and converted with each other, forming a multi-energy network in the region. The comprehensive energy system can realize optimal allocation of energy and improve energy efficiency due to its characteristics of multi-energy complementation and supply-demand interaction, and can meet the energy demand of various users within the scope of the comprehensive energy system, and gradually become an important carrier to promote low-carbon operation of industrial parks and improve wind and light consumption rate. The current research and development trend of park micro-energy network comprehensive energy system mainly includes two aspects of optimal dispatching operation and park planning design method. The optimal operation of park micro-energy network comprehensive energy system includes fine modeling of park micro-energy network, economic operation and dispatching, main-distribution-micro collaborative operation, and participation of park micro-energy network in power market. The planning and design of park micro-energy network comprehensive energy system includes configuration of multi-energy devices in the park, selection and sizing of multi-energy devices, production line process design and planning, and economic investment planning.

[0003] The development stage and industrial structure of industrial park comprehensive energy system are different, and the load demand fluctuation and source side uncertainty will also be amplified over time in the long time scale of planning. The diversity and spatio-temporal heterogeneity of load demand put forward higher requirements for the reliability and quality of energy supply, which requires the park micro-energy network to fully consider the influence of uncertain events in the planning stage and to be able to adaptively adjust the planning scheme. In view of this, how to identify the key planning decision points, decompose the planning scheme of a comprehensive energy system into multi-stage dynamic planning scheme in a long period, and avoid the risk of one-time investment planning is particularly important.

[0004] Prior art 1 (CN120410105A) A successive approximation random dynamic programming method of a watershed water and scenery integrated system. The invention discloses a successive approximation random dynamic programming method for a watershed water and scenery integrated system, aiming to solve the uncertainty optimization problem in multi-energy coupling complementary scheduling. The specific steps include: first, the theoretical probability distribution function is used to sample discrete values to describe the uncertainty of runoff, photovoltaic and wind power output; then a transition probability matrix is constructed using Markov chain to describe the time series continuity of input variables; then a complementary optimization scheduling model is established combining the system objective function and constraint conditions; then the successive approximation-random dynamic programming (SDP-SA) method is used, taking the reservoir period initial reservoir capacity, runoff, wind and light output as multi-dimensional state variables, and the period end reservoir capacity as the decision variable, to derive the scheduling decision table; finally, in actual operation, according to the real-time reservoir capacity, inflow and wind and light forecast data, the optimal scheduling decision is determined by table lookup interpolation. But its method is based on the short-term scheduling strategy of approximation-random dynamic programming, and does not consider the long-term planning decision problem.

[0005] Prior art 2 (CN120389434A) Electric vehicle V2G microgrid energy storage capacity optimization method based on dynamic programming. The invention discloses an electric vehicle V2G microgrid energy storage capacity optimization method based on dynamic programming, belonging to the technical field of microgrid energy storage optimization, aiming to solve the energy storage capacity optimization defects caused by spatial dynamic characteristics caused by electric vehicle mobility. The method first collects power grid topology parameters and electric vehicle mobile trajectory data, divides node positions by geographic grid and generates a space-time matrix combined with time stamps to identify high-frequency access nodes; then based on the impedance matrix between nodes, a state transition equation is constructed to set the node-level safety boundary with the line capacity change rate as the dynamic constraint; then a phased reverse dynamic programming algorithm is used to preferentially optimize the spatial dimension of high-frequency access nodes, generating a charge and discharge constraint that integrates voltage deviation and capacity overrun penalty term; finally, the constraints are integrated to minimize the total operating cost as the target for global optimization, outputting the energy storage capacity configuration strategy for each period and generating an energy storage deployment scheme. But it is to solve the spatial optimization and operation problem of deploying electric vehicles, and it focuses on short-term deployment and operation decision, and does not involve long-term investment planning. The method does not use time sequence rolling optimization, and the calculation cost of this method increases significantly in long time scale.

[0006] The prior art generally focuses on using a dynamic programming method for short-term scheduling decisions, and less considers investment planning problems in a long period scale; meanwhile, current stochastic dynamic programming relies on expected values of predictions, and prediction accuracy will affect planning schemes; although the dynamic programming method is applied in many fields, it is not common to use it in-depth in the joint optimization operation of comprehensive energy system and industrial park planning, and the double-layer architecture formed by the time sequence rolling optimization method and the value function model proposed in the application realizes dimension reduction of the dynamic programming problem in the time dimension, taking into account the robustness of the planning scheme while avoiding the dimension disaster of the traditional dynamic programming. SUMMARY

[0007] The application aims to solve the problems in the prior art, and provides a dynamic optimization method for identifying key planning decision points of a comprehensive energy system, establishes a planning and operation joint optimization model considering equipment operation constraints based on the physical composition of the comprehensive energy system, evaluates the value of the current planning scheme by constructing a value function model of the planning decision, constructs a rolling time domain optimization model of the comprehensive energy system, calls the value function model rolling, and obtains a continuous value curve through linear interpolation and reverse recursion. The key planning decision points are obtained based on a peak value identification algorithm. Finally, the one-time complex investment problem is converted into a multi-stage dynamic planning scheme of the comprehensive energy system, reduces the planning risk of the park comprehensive energy system, and improves the system operation income.

[0008] This section aims to summarize some aspects of the embodiments of the application and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of the specification to avoid obscuring the purpose of this section, the abstract and the title, and such simplifications or omissions cannot be used to limit the scope of the application.

[0009] In order to achieve the above-mentioned purpose, the application has the following technical solutions, comprising: Step S1: based on the energy supply structure and supply and demand equipment of the comprehensive energy system, constructing an optimization operation model of the comprehensive energy system; Step S2: considering the operation condition of a specific planning decision in a long time in the future to evaluate the value of the decision, constructing a planning value function model of the comprehensive energy system; Step S3: constructing a rolling time domain optimization model of the comprehensive energy system, and proposing a method for identifying key decision points in a planning period; Step S4: based on a small number of high planning value decision points in a long planning period, proposing a multi-stage dynamic planning scheme of the comprehensive energy system based on the key decision points.

[0010] As a preferred scheme of the dynamic optimization method for identifying key planning decision points of the comprehensive energy system, the step S1 is analyzed, wherein: Based on the energy supply structure and supply and demand equipment of the integrated energy system, an optimized operation model for the integrated energy system is constructed. 1) Objective function The optimization objective is to minimize the total cost of the system. Its annualized investment cost and cycle operating costs constitute.

[0011] Annualized investment cost Determined by the capital recovery factor, it is the sum of the annualized investments in all equipment, satisfying the following formula: in, This indicates that the equipment belongs to the photovoltaic category. Wind power Energy storage ,gas turbine A set; For equipment The capacity; The unit investment cost of the equipment; The expected lifespan of the equipment; is the discount rate.

[0012] Cycle operating costs The total cost of electricity purchase, fuel, and operation and maintenance during the dispatch cycle is expressed by the following formula: in, , , These represent the electricity purchase and sales cost, gas purchase cost, and energy storage operation cost coefficients, respectively. , , , , These represent the integrated energy system's power purchase from the grid, power sales from the grid, power generation from the gas-fired power generation unit, charging power from the energy storage unit, and discharging power from the energy storage unit, respectively. T Indicates the total runtime period.

[0013] 2) Real-time power balance constraints At every moment t The total power generation of the system must equal the total power consumption, and the system should satisfy the following constraints: in, Indicates time t User load.

[0014] 3) Renewable energy output constraints where, , denote the uncertainty coefficients of photovoltaic and wind power output, respectively; , denote the expected values of photovoltaic and wind power generation at time t, respectively; t , denote the upper limits of photovoltaic and wind power installed capacity, respectively; , denote the photovoltaic and wind power generation at time t, respectively.

[0015] 4) Load response constraints where, denotes the uncertainty coefficient of load; denotes the expected value of load at time t; denotes the load at time t.

[0016] 5) Micro gas turbine output constraints where, denotes the output power of the gas turbine at time t; t denotes the upper limit of the gas turbine output.

[0017] 6) Energy storage system constraints where, , denote the state of charge capacity of the energy storage at the current time t and the previous time t-1, respectively; denotes the charging power of the energy storage at time t; denotes the discharging power of the energy storage at time t; , denote the charging and discharging efficiencies, respectively.

[0018] where, , denote the minimum and maximum state of charge coefficients of the energy storage, respectively; denotes the upper limit of the state of charge capacity of the energy storage.

[0019] where,​​ This represents the charge / discharge identifier at time t, where 0 indicates no charge / discharge and 1 indicates charge / discharge. This indicates the upper limit of the maximum charging and discharging power.

[0020] Energy storage systems are expected to meet periodic power balance constraints during long-term operation.

[0021] 7) Interaction constraints between integrated energy system and power grid in, , These represent the power purchased and sold at time t, respectively. This represents the electricity sales identifier at time t, where 0 indicates no electricity sales and 1 indicates electricity sales. This indicates the upper limit of the power that can be interacted with the power grid.

[0022] in, , These represent the planned capacity and minimum planned capacity of device i, respectively.

[0023] As a preferred embodiment of the dynamic optimization method for identifying key planning decision points in an integrated energy system according to the present invention, step S2 is specifically analyzed, wherein: Step S2 considers the operational status of a specific planning decision over a longer period to evaluate the value of the decision and constructs a planning value function model for the integrated energy system, specifically including: 1) Functional model of the value assessment module in, The value function represents the planning scheme, and the relative magnitude of its value indicates the applicability of the planning scheme. It is a functional relationship that takes the planning scheme value as input and the value function value as output; This represents the input set consisting of the planning scheme and external operating parameters; These represent the input sets containing the upper limits of planned energy storage capacity, planned photovoltaic power, planned wind power, and planned gas turbine output, respectively.

[0024] function The calculation results depend on a short-running optimization problem solved internally. We define the decision variables and objective function of this internal problem using the following formulas.

[0025] wherein, denotes the time window for evaluation; denotes the time window for each time instant; denotes the time instant corresponding electricity purchase price; , denotes the electricity purchase / sale power at time instant ; denotes the gas purchase cost price; denotes the power of the gas turbine at time instant ; denotes the penalty coefficient for system power imbalance within the evaluation time window ; denotes the system imbalance power within the evaluation time window ; denotes the algebraic sum of the absolute values of the system imbalance power within the evaluation time window ; denotes the comprehensive cost of operation and imbalance power within the time window .

[0026] 2) Value function calculation wherein, denotes the value of that has been minimized; scaling factor for adjusting the result to a suitable order of magnitude.

[0027] As a preferred scheme of the dynamic optimization method for identifying key planning decision points of an integrated energy system, the step S3 is specifically analyzed, wherein: The step S3 constructs a rolling time domain optimization model of the integrated energy system, and proposes a method for identifying key decision points in a planning period, specifically including: S3.1, through rolling optimization iteration, simulating the annual operation of the system, and calculating the current decision value function for each rolling step, forming a sequence of value functions; S3.2, based on the sequence of value functions obtained in S3.1, performing reverse iteration on the value curve from back to front, and constructing a value curve that is continuous and reflects long-term cumulative value in a long planning period; S3.3, automatically finding the peak value on the final value curve, thereby determining the key decision point, and realizing the identification and output of the key decision point.

[0028] The step S3.1 forms a sequence of value functions through rolling optimization iteration, specifically including: 1) Initialization of the rolling optimization iteration The process starts at the first time point At this time, the system solves the integrated energy system planning and operation joint optimization model based on step S1 to determine the initial baseline configuration of the system.

[0029] wherein, represents the starting time point of the rolling optimization; represents the optimal decision variable set at which makes the objective function reach the minimum value.

[0030] Based on the value function calculation, the decision value at the first time point is calculated.

[0031] wherein, represents the decision value at the first time point .

[0032] 2) Rolling iteration For each subsequent rolling time point , the following sub-models are repeatedly solved. The state update equation needs to be set for the time sequence variable during the rolling iteration process, and the model specifies the initial capacity of the energy storage .

[0033] wherein, represents the time interval between the starting time of the subsequent rolling optimization and the starting time , which is called the sliding time window length of the rolling optimization; the initial capacity of the energy storage k represents the initial capacity of the energy storage at the time point S ; represents the optimal decision variable set at S which makes the objective function reach the minimum value; represents the initial capacity of the energy storage at the time point ; represents the optimal decision variable set at which makes the objective function reach the minimum value; represents the initial capacity of the energy storage at the time point ; represents the initial capacity of the energy storage at the time point which is equal to the value of the optimized energy storage capacity at in ; which is called the time window length of a single rolling optimization; then represents the initial capacity of the energy storage at the time point ; represents the value of the optimized energy storage capacity at in W , which is called the time window length of a single rolling optimization; then ​The decision quantity set after time optimization Calculation The decision value at the time .

[0034] 3) Value function sequence output When the rolling cycle traverses the time points of the whole cycle planning, the final output is a sparse state value function point sequence: Step S3.2 builds a value curve in a long planning period that is continuous and reflects long-term cumulative value, specifically including: 1) Interpolation construction of continuous value curve Connect discrete value points to form a continuous value curve defined at each hour in the whole planning period.

[0035] Among them, represents the continuous value function curve formed after linear interpolation processing.

[0036] 2) Reverse order iteration of value curve Based on the Bellman equation idea in dynamic programming, the total value at a time depends not only on the present, but also on all future possible values. The value function should satisfy the following formula after reverse order iteration.

[0037] Among them, represents the value function at the critical point T of the planning period, which should be equal; represents the immediate planning decision value at t; is the future value discount factor, which represents the coefficient of discounting future value to the present time; represents the discounted value of the sum of all values from t+1 to the future; represents the cumulative value function of the potential value of the system planning scheme from now until the future.

[0038] Step S3.3 identification and output of key decision points, specifically including: The goal is to filter out a set of time points from the value function curve through strict mathematical conditions Each element in the set is defined as a key decision point.

[0039] 1) Local optimality condition This condition ensures that the selected The value is higher than its adjacent left and right points, and is the optimal value in the local range.

[0040] wherein, represents the key decision point The value at the moment; represents the logical "and".

[0041] 2) Significance condition This condition ensures that the value function of the selected key decision point must be large enough to filter out insignificant fluctuations and ensure that the identified decision point has significant value.

[0042] wherein, represents the minimum peak height threshold; is the percentile threshold, i.e. the minimum peak threshold should not be higher than the percentile of the value curve %; is an indicator function that takes 1 when the condition is true, otherwise 0; represents the lower bound, i.e. the minimum value that satisfies the condition; represents the minimum value filtered out after satisfying the condition.

[0043] 3) Sparsity condition This condition ensures that any two selected key decision points are sufficiently spaced in time, avoiding excessive concentration of decision points and making the multi-stage dynamic programming more realistic.

[0044] wherein, , represent the selected i th, the j th decision point; represents the minimum time interval; represents the set of key decision points.

[0045] As a preferred scheme of the dynamic optimization method for identifying key planning decision points of a comprehensive energy system according to the application, the step S4 is specifically analyzed, wherein: Step S4 proposes a multi-stage dynamic programming scheme for the comprehensive energy system based on key decision points based on a small number of high planning value decision points in a long planning period, specifically including: 1) Data input Based on the key decision point set obtained in the above step S3.3, initialize the minimum planning capacity of each device in the initial stage wherein, denotes initialization of the minimum planning capacity set; denote the minimum planning capacity of energy storage, photovoltaic, wind power and gas turbine respectively.

[0046] 2) Objective function wherein, denote the annualized investment cost and operation cost calculated within the planning window of the current stage k respectively. 3) Incremental planning constraints

[0047] The current stage The planned capacity of each device must be greater than or equal to the optimal capacity determined in the last stage k -1. k

[0048] wherein, denotes the planning capacity value within the k th planning stage, denotes the optimal planning value within the k -1 stage.

[0049] Further, the present application also provides a computer device comprising a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0050] Further, the present application also provides a computer readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the method.

[0051] The present application has the following beneficial effects: The present application is based on the physical structure of the integrated energy system, establishes a planning and operation joint optimization model considering the operation constraints of the device; takes a determined device capacity scheme as the input parameter, takes the operation condition within a long operation period as the evaluation, calculates the value of the current planning scheme, forms a value function model of the planning decision, constructs a rolling time domain optimization model of the integrated energy system, the system calls the value function model in a rolling manner, calculates the value under the current state, and obtains a continuous value curve through linear interpolation and reverse recursion. Based on the peak value identification algorithm, the key planning decision points are obtained; based on a small number of key decision points within a long planning period, the identified key decision points are taken as the decision stage, and through the non-decreasing incremental planning mode, the one-time complex investment problem is converted into a multi-stage dynamic planning scheme of the integrated energy system. BRIEF DESCRIPTION OF DRAWINGS​

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0053] Figure 1 A flowchart of a dynamic optimization method for identifying key planning decision points of a comprehensive energy system according to an embodiment of the present application; Figure 2 A value function curve and a key decision point diagram according to an embodiment of the present application; Figure 3 A planning scheme evolution process diagram according to an embodiment of the present application; Figure 4 A change diagram of energy storage capacity in the last scheduling period in time sequence rolling according to an embodiment of the present application; Figure 5 A power balance analysis diagram in the last scheduling period in time sequence rolling according to an embodiment of the present application; Figure 6 A distribution diagram of annualized investment cost according to an embodiment of the present application; DETAILED DESCRIPTION In the following description, specific details are set forth in order to provide a thorough understanding of embodiments of the present application. However, persons of ordinary skill in the art will readily appreciate that embodiments of the present application can be practiced without these specific details. In other instances, well-known structures, devices, circuits, and methods have not been described in detail in order to avoid obscuring the present application.

[0054] Embodiment 1 Reference Figure 1 According to a first embodiment of the present application, the embodiment provides a dynamic optimization method for identifying key planning decision points of a comprehensive energy system, comprising: Step S1: constructing an optimization operation model of the comprehensive energy system based on an energy supply structure and supply and demand equipment of the comprehensive energy system; Step S2: constructing a planning value function model of the comprehensive energy system by considering the value of a specific planning decision in the future long-time operation state to evaluate the value of the decision; Step S3: constructing a rolling time domain optimization model of the comprehensive energy system, and proposing a method for identifying key decision points in a planning period; Step S4: proposing a multi-stage dynamic planning scheme of the comprehensive energy system based on key decision points based on a small number of high planning value decision points in a long planning period.

[0055] As a preferred scheme of the dynamic optimization method for identifying key planning decision points of an integrated energy system, the step S1 is analyzed, wherein: Based on the energy supply structure and supply and demand equipment of the integrated energy system, an optimized operation model of the integrated energy system is constructed, characterized in that: 1) Objective function The optimization objective is to minimize the total cost of the system , which is composed of the annualized investment cost and the periodic operation cost .

[0056] The annualized investment cost is the sum of the annualized investment of all equipment, which satisfies the following formula: Wherein, represents the set of equipment types belonging to photovoltaic , wind power , energy storage , gas turbine ; is the capacity of the equipment ; is the unit investment cost of the equipment; is the expected life of the equipment; is the discount rate.

[0057] The periodic operation cost is the sum of all costs such as electricity purchase, fuel and operation and maintenance within the dispatching period, which satisfies the following formula: Wherein, , , respectively represent the electricity purchase and sale price cost, gas purchase cost, and energy storage operation cost coefficient; , , , , respectively represent the purchased power, sold power, gas power generation unit power, energy storage unit charging power, and energy storage unit discharging power of the integrated energy system to the power grid; T represents the total operation period.

[0058] 2) Real-time power balance constraint At each time t , the total power generation of the system must be equal to the total power consumption, and the system should satisfy the following constraint: where, represents the user load at time t .

[0059] 3) Renewable energy generation constraints where, , represent the uncertainty coefficients of photovoltaic and wind power generation, respectively; , represent the expected values of photovoltaic and wind power generation at time t . , represent the upper limits of installed capacity of photovoltaic and wind power, respectively; , represent the power generation of photovoltaic and wind power at time t, respectively.

[0060] 4) Load response constraints where, represents the uncertainty coefficient of load; represents the expected value of load at time t; represents the load at time t.

[0061] 5) Micro gas turbine generation constraints where, represents the t power output of the gas turbine at time t; represents the upper limit of the gas turbine output.

[0062] 6) Energy storage system constraints where, , represent the state-of-charge capacity of the energy storage at the current time t and the previous time t-1; represents the charging power of the energy storage at time t; represents the discharging power of the energy storage at time t; , represent the charging and discharging efficiencies, respectively.

[0063] where, , represent the minimum and maximum state-of-charge coefficients of the energy storage, respectively; represents the upper limit of the state-of-charge capacity of the energy storage.​​​​​​​​​​​​​​​​​​​​​​​​​

[0064] wherein, represents the charging and discharging identifier at time t, 0 represents no charging and discharging, and 1 represents charging and discharging; represents the maximum charging and discharging power upper limit.

[0065] The energy storage system is considered to satisfy the periodic power balance constraint during long-period operation.

[0066] 7) Comprehensive energy system and grid interaction constraint wherein, , respectively represent the power purchase and power sale at time t; represents the power sale identifier at time t, 0 represents no power sale, and 1 represents power sale; represents the upper limit of the power interaction with the grid.

[0067] wherein, , respectively represent the device i planning capacity and the minimum planning capacity.

[0068] As a preferred scheme of the dynamic optimization method for identifying key planning decision points of a comprehensive energy system, the step S2 is specifically analyzed, wherein: The step S2 considers the running condition of a specific planning decision in a future long time to evaluate the value of the decision, and constructs a planning value function model of the comprehensive energy system, specifically including: 1) Function model of the value evaluation module wherein, represents the value function of the planning scheme, and the relative size of the value represents the applicability of the planning scheme; is a function relationship with the planning scheme value as input and the value function value as output; represents an input set composed of the planning scheme and external running parameters; respectively represent the input set with the energy storage planning capacity upper limit, the photovoltaic planning power upper limit, the wind power planning power upper limit, and the gas turbine output planning upper limit as elements.

[0069] the function The result is dependent on a short-term operation optimization problem solved internally. We define the decision variables and the objective function of this internal problem as follows.

[0070] where, denotes the time window for evaluation; denotes each time instant of the time window denotes the purchase price of electricity at time instant , denotes the purchase / sale power at time instant denotes the cost price of gas; denotes the power of the gas turbine at time instant denotes the penalty coefficient for system power imbalance within the evaluation time window denotes the system imbalance power within the evaluation time window denotes the algebraic sum of the absolute values of the system imbalance power within the evaluation time window denotes the total cost of operation and imbalance power within the evaluation time window

[0071] 2) Value function calculation where, denotes the minimum value of scaling factor to adjust the result to a suitable order of magnitude.

[0072] As a preferred solution of the dynamic optimization method for identifying key planning decision points of the integrated energy system, the step S3 is specifically analyzed, wherein: The step S3 constructs a rolling time domain optimization model of the integrated energy system, and proposes a method for identifying key decision points in the planning period, specifically including: S3.1, through rolling optimization iteration, simulating the annual operation of the system, and calculating the current decision value function for each rolling step, forming a sequence of value functions; S3.2, based on the sequence of value functions obtained in S3.1, iteratively from back to front, constructing a value curve that is continuous and reflects long-term cumulative value in a long planning period; ​​​​​​​​​S3.3 Automatically find the peak value on the final value curve to determine the key decision point and realize the identification and output of the key decision point.

[0073] Step S3.1 involves forming a value function sequence through rolling optimization iterations, specifically including: 1) Initialization of rolling optimization iteration The process begins at the first point in time. At this point, the system solves the joint optimization model for integrated energy system planning and operation based on step S1 to determine the initial baseline configuration of the system.

[0074] in, This indicates the starting point of the rolling optimization; Indicates in Time optimization achieved The set of optimal decision quantities with minimum value.

[0075] Based on the value function calculation, the first time point is calculated. The decision-making value.

[0076] in, Indicates the first time point The decision-making value.

[0077] 2) Rolling Iteration For each subsequent rolling time point Repeat the solution process for the following sub-models. During the rolling iteration, it is necessary to set state update equations for the time-series variables; in this model, this specifically refers to the initial energy storage capacity. .

[0078] in, This indicates the starting point and start time of subsequent rolling optimizations. difference k indivual S The time interval, S The sliding time window length, referred to as rolling optimization; initial energy storage capacity. express The initial energy storage capacity at any given time; Indicates in Time optimization achieved The set of optimal decision quantities for minimizing the value; express Initial energy storage capacity at any given time equal The optimized energy storage capacity at all times value , W The time window length is called the single-roll optimization; then it is determined by... The set of decision quantities optimized at each time step calculate The value of decision-making at any moment .

[0079] 3) Output of the value function sequence After the rolling loop iterates through all the time points of the full cycle planning, the final output is a sparse sequence of state-value function points: Step S3.2 Constructs a value curve that is continuous over a long planning period and reflects long-term cumulative value, specifically including: 1) Construction of continuous value curve interpolation By connecting discrete value points, a continuous value curve is formed, defined for each hour within the entire planning cycle.

[0080] in, This indicates that a continuous value function curve is formed after linear interpolation.

[0081] 2) Reverse iteration of the value curve Based on the Bellman equation in dynamic programming, the total value at a given moment depends not only on the present but also on all possible future values. The value function, after reverse iteration, should satisfy the following formula.

[0082] in, The value functions at the critical point T of the planning cycle should be equal; This represents the immediate planning and decision-making value at time t. It is the future value discount factor, which represents the coefficient by which future value is discounted to the present moment; It represents the present value of the sum of all values ​​from time t+1 up to the future; It represents the cumulative value function of the potential of the system planning scheme from now until the future.

[0083] Step S3.3 Identification and output of key decision points, specifically including: The goal is to select a set of time points from the value function curve using rigorous mathematical conditions. each element in the set is defined as a key decision point.

[0084] 1) Local optimality condition This condition ensures that the selected value is higher than its immediate left and right points, which is the optimal value in the local range.

[0085] wherein, represents the value of the key decision point moment; represents the logical "and".

[0086] 2) Significance condition This condition ensures that the value function of the selected key decision point must be large enough to filter out insignificant fluctuations, ensuring that the identified decision points have significant value.

[0087] wherein, represents the minimum peak height threshold; is the percentile threshold, i.e. the minimum peak threshold should not be higher than the percentile of the value curve %; is an indicator function that takes 1 when the condition is true, otherwise 0; represents the lower bound, i.e. the minimum value that satisfies the condition; represents the minimum value filtered out after satisfying the condition.

[0088] 3) Sparsity condition This condition ensures that any two selected key decision points have sufficient time interval, avoiding excessive concentration of decision points, making the multi-stage dynamic programming more practical.

[0089] wherein, , represent the i th, the j th decision point selected; represents the minimum time interval; represents the set of key decision points.

[0090] As a preferred scheme of the dynamic optimization method for identifying key planning decision points of a comprehensive energy system according to the present application, the step S4 is specifically analyzed, wherein: Step S4 proposes a multi-stage dynamic planning scheme for the integrated energy system based on a small number of high-planning-value decision points within a long planning period. Specifically, this includes: 1) Data Input Based on the above step S3.3, obtain the set of key decision points. Minimum planned capacity of each device in the initial stage of initialization in, This represents the initial minimum planning capacity set; These represent the minimum planned capacity for energy storage, photovoltaic power, wind power, and gas turbines, respectively.

[0091] 2) Objective function in, , These represent the annualized investment cost and operating cost calculated within the planning window of the current stage k, respectively.

[0092] 3) Incremental programming constraints Current stage k The planned capacity of each piece of equipment must be greater than or equal to that of the previous phase. k -1 represents the already determined optimal capacity.

[0093] in, Indicates the first k Planning capacity values ​​within each planning phase express k The optimal planning value within the -1 stage.

[0094] Example 2 Reference Figures 2-6 As an embodiment of the present invention, a dynamic optimization method for identifying key planning decision points of an integrated energy system is provided. To verify the beneficial effects of the present invention, a comparative experiment is conducted for scientific demonstration.

[0095] The example demonstrates the planning of an integrated energy system for an industrial park, with the window width used for rolling optimization. W Set to 120 hours, rolling step S Set to 96 hours, the value function evaluation window is 2. W That is, 240 hours. Future value discount factor. gamma Set to 0.1. Parameters such as equipment investment and operation and maintenance costs are all set in the model, as shown in Table 1.

[0096] Table 1. Park Equipment and Planned Operation Parameters Table 1. Park equipment parameters First, the system annual operation is simulated by the flow of S1-S3, and an enhanced value curve reflecting the long-term cumulative value is generated V ( t ). Then, the key decision points (KDPs) are automatically screened out by the peak identification algorithm. The value function curve and the key decision points are shown in Figure 2 .

[0097] Table 2. Key decision points and value functions Table 2. Key decision points and value functions KDPs are the peaks of the system "value function". The higher the value of a point, the more acute the contradiction between supply and demand, the maximum operating cost pressure, or the most significant potential economic opportunity at that moment, which is the time to make planning. According to the results in Table 2, the following analysis can be made: 865 hours in the winter, which is one of the coldest periods of the year, is in the winter. The short day and weak solar intensity result in the lowest photovoltaic output in the year, while the winter heating demand leads to high total load. In this extreme case, the system must rely heavily on expensive gas turbine power generation or purchase peak / flat power from the grid, resulting in a sharp rise in operating costs.

[0098] 1537 and 1921 hours are in the typical seasonal transition period, when the supply and demand relationship fluctuates dramatically. When the load is not high but the wind and light are abundant, the system has a high risk of abandoning wind and light; while the load surges but the wind and light drop sharply, the system faces high standby costs. The value function forms a peak at this point, indicating that the system has the highest potential benefit of energy time shift by configuring energy storage to smooth fluctuations. This is the key period for determining the power and capacity ratio of energy storage system to respond to high-frequency fluctuations.

[0099] 3649 hours are in the early summer, when the cooling load begins to appear and gradually becomes dominant. The light resource is good, and the photovoltaic output enters the high-output period. The supply and demand structure changes with the season, and the value function peaks at this point, indicating whether the existing energy storage capacity is sufficient to transfer the surplus photovoltaic power in the daytime to the new peak in the evening.

[0100] 4897 hours in the summer, the hottest period of the year, the cooling load reaches its peak, and the peak-valley electricity price difference is also the most significant. During the day, photovoltaic output is the strongest, and the valley electricity price is the lowest, which is the golden period for energy storage charging; in the evening, the electricity load is the highest, and the peak electricity price is the most expensive, which is the largest window for energy storage discharging to make a profit. In the evening, the system faces the largest power gap in the year, and the dependence on gas turbines, energy storage, and grid purchases reaches its limit. This time point is the extreme embodiment of the economic benefits and power supply costs throughout the year. The highest peak of the value function usually appears here, as it represents the maximum potential benefit that can be brought by investing in an energy storage system.

[0101] 6433 hours in the summer, the hottest period of the year, the cooling load reaches its peak, and the peak-valley electricity price difference is also the most significant. During the day, photovoltaic output is the strongest, and the valley electricity price is the lowest, which is the golden period for energy storage charging; in the evening, the electricity load is the highest, and the peak electricity price is the most expensive, which is the largest window for energy storage discharging to make a profit. In the evening, the system faces the largest power gap in the year, and the dependence on gas turbines, energy storage, and grid purchases reaches its limit. This time point is the extreme embodiment of the economic benefits and power supply costs throughout the year. The highest peak of the value function usually appears here, as it represents the maximum potential benefit that can be brought by investing in an energy storage system.

[0102] Figure 3 presents the incremental changes in the configuration capacity after planning at the above KDPs. From the evolution of the capacity of each device in this example, it can be seen that a certain scale of wind power and photovoltaic power can be configured at the beginning of the planning to make up for the energy gap within the industrial park and reduce the purchase cost. As the configuration capacity of new energy continues to increase, the energy supply side of the system gradually faces the pressure of consumption due to the uncertainty of new energy output. Therefore, subsequent planning needs to gradually increase the configuration of energy storage to enhance the system's consumption capacity, while using the operation and scheduling of energy storage to smooth the fluctuations in output and load, and further reduce the system's operating cost by using electricity peak-valley arbitrage.

[0103] Figure 4 、 Figure 5 The energy storage capacity changes and system power supply and demand balance states in the last planning window are shown respectively, which intuitively presents the system's running characteristics after the final landing of this planning scheme. The system chooses to purchase electricity from the grid during the low electricity consumption period, and sells electricity to the grid during the high electricity consumption period to achieve peak-valley arbitrage; at the same time, the energy storage charges in coordination with the peak photovoltaic output to consume excess electricity, and discharges off-peak during the subsequent load peak period to supply load electricity, and its capacity fluctuates periodically with the changes in electricity price and new energy output. Figure 6The annual investment cost composition of the final overall scheme is shown: the investment proportions of photovoltaic and wind power are 2.7% and 3.5% respectively, the investment proportion of gas turbine is 13.1%, and the investment proportion of energy storage is as high as 80.7%, and the cumulative investment amount is 6721730 yuan. The cumulative final planning scheme is: energy storage capacity 4631kWh, energy storage power 972kW, photovoltaic capacity 2540kWh, wind power capacity 942kWh, and gas turbine 529kWh.

[0104] In view of the small energy consumption scale of the industrial park comprehensive energy system set in the example, the investment scale of wind power and photovoltaic is correspondingly small, and the system mainly reduces the subsequent operation cost by increasing the energy storage investment.

[0105] The application is beneficial to improving the scientificity of investment decision of the comprehensive energy system, enhancing the economy and engineering feasibility of long-period planning scheme, promoting efficient consumption and utilization of renewable energy, and has significance in promoting scientific planning and sustainable development of complex energy system. The dynamic optimization method for identifying key planning decision points of the comprehensive energy system proposed in the application identifies the key planning decision points with the most significant influence on the system full planning period cost by constructing the economic value function throughout the year, and generates a staged investment path with economic optimization and engineering practicability.

[0106] The present disclosure can be a system, a method, and / or a computer program product. The computer program product can include a computer readable storage medium having computer readable program instructions embodied therewith, wherein the computer readable program instructions are used to cause a processor to implement various aspects of the present disclosure.

[0107] The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium can be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium include the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or punched tape, and any suitable combination of the foregoing. A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0108] Computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and / or a wireless network. The network can comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing / processing device.

[0109] Computer readable program instructions for carrying out operations of the present disclosure can be assembly instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions can execute entirely on the user's computing device, partly on the user's computing device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user's computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computing device, for example, through the Internet using an Internet Service Provider. In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0110] Finally, it should be noted that the above-mentioned embodiments are merely intended for describing and illustrating, but not limiting the technical solutions of the present application. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application. Any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered within the protection scope of the claims of the present application.

Claims

1. A dynamic optimization method for identifying key planning decision points in an integrated energy system, characterized in that, Includes the following steps: Step S1: Based on the energy supply structure and supply and demand equipment of the integrated energy system, construct an optimized operation model for the integrated energy system; Step S2: Consider the operational status of a specific planning decision over a longer period of time to evaluate the value of the decision and construct a planning value function model for the integrated energy system; Step S3: Construct a rolling time-domain optimization model for the integrated energy system and propose a method for identifying key decision points within the planning period; Step S4: Based on a small number of high-planning-value decision points within a long planning period, propose a multi-stage dynamic planning scheme for the integrated energy system based on key decision points.

2. The dynamic optimization method for identifying key planning decision points in an integrated energy system according to claim 1, characterized in that, In step S1, based on the energy supply structure and supply and demand equipment of the integrated energy system, the optimized operation model of the integrated energy system is constructed as follows: 1) Objective function The optimization objective is to minimize the total cost C of the system. total The annualized investment cost C inv and cycle operating costs C om constitute: minC total =C inv +C om Annualized investment cost C inv Determined by the capital recovery factor, it is the sum of the annualized investments of all equipment, satisfying the following formula: Where i∈{pv,wt,bat,g} represents the set of equipment types belonging to photovoltaic pv, wind power wt, energy storage bat, and gas turbine g; Let i be the capacity of device i; N represents the unit investment cost of the equipment. i The expected lifespan of the equipment; r p The discount rate; Cycle operating cost C om The total cost of electricity purchase, fuel, and operation and maintenance during the dispatch cycle is expressed by the following formula: in, C fuel , These represent the electricity purchase and sales cost, gas purchase cost, and energy storage operation cost coefficients, respectively; P t buy P t sell P t g P t ch P t dis These represent the integrated energy system's power purchase from the grid, power sales, power generation from the gas-fired power generation unit, charging power from the energy storage unit, and discharging power from the energy storage unit, respectively; T represents the total operating time. 2) Real-time power balance constraints At each time t, the total power generation of the system must equal the total power consumption, and the system should satisfy the following constraint: P t buy +P t g +P t dis +P t wt +P t pv =P t load +P t ch +P t sell Among them, P t load This represents the user load at time t; 3) Renewable energy output constraints Where, ω pv ω w These represent the uncertainty coefficients for photovoltaic power output and wind power output, respectively. Let represent the expected power generation values ​​of photovoltaic and wind power at time t, respectively; These represent the upper limits of installed capacity for photovoltaic and wind power, respectively; P t pv P t wt Let represent the power generation of photovoltaic and wind power at time t, respectively; 4) Load response constraints Where, ω l The uncertainty factor representing the load; P represents the expected load at time t; t load This represents the load at time t; 5) Output constraints of micro gas turbines Among them, P t g This represents the output power of the gas turbine at time t; Indicates the upper limit of gas turbine output; 6) Constraints of energy storage systems in, P represents the state-of-charge capacity of the stored energy at time t and time t-1 before; t ch P represents the energy storage charging power at time t; t dis η represents the energy storage and discharge power at time t; ch η dis These represent the charge and discharge efficiencies, respectively. Where, α min α max These represent the minimum and maximum state-of-charge coefficients for energy storage, respectively. Indicates the upper limit of the energy storage state-of-charge capacity; in, This represents the charge / discharge identifier at time t, where 0 indicates no charge / discharge and 1 indicates charge / discharge. Indicates the upper limit of maximum charging and discharging power; Energy storage systems are expected to meet periodic power balance constraints during long-term operation. 7) Interaction constraints between integrated energy system and power grid in, P t sell These represent the power purchased and sold at time t, respectively. This represents the electricity sales identifier at time t, where 0 indicates no electricity sales and 1 indicates electricity sales. Indicates the upper limit of power interaction with the power grid; in, These represent the planned capacity and minimum planned capacity of device i, respectively.

3. The dynamic optimization method for identifying key planning decision points in an integrated energy system according to claim 2, characterized in that, In step S2, the value of a specific planning decision is evaluated by considering its operational status over a longer period. The planning value function model for the integrated energy system is constructed as follows: 1) Functional model of the value assessment module Q=F value (θ) Where Q represents the value function of the planning scheme, and its relative value represents the applicability of the planning scheme; F value It is a functional relationship that takes the planning scheme value as input and the value function value as output; θ represents the input set consisting of the planning scheme and external operating parameters; These represent the input sets containing the upper limits of planned energy storage capacity, planned photovoltaic power, planned wind power, and planned gas turbine output, respectively. The function F value The calculation results depend on a short-running optimization problem solved internally, whose decision variables and objective function are defined by the following formulas: Among them, T eval t' represents the time window used for evaluation; t' represents the time window T. eval Every moment; This represents the electricity price at time t′; C represents the power purchased and sold at time t′; fuel This indicates the cost price of purchasing gas; λ represents the power of the gas turbine at time t′; λ represents the power output within the evaluation time window T. eval Penalty coefficient for internal system electrical imbalance; ΔP t′ Indicates the evaluation time window T eval Unbalanced power within the system; Indicates the evaluation time window T eval The algebraic sum of the absolute values ​​of the unbalanced electrical quantities within the system; Indicates in T eval The combined cost of operation and unbalanced power generation is considered within the time window; 2) Value function calculation in, This indicates that the minimum has been obtained. The value of M; scale The scaling factor is used to adjust the result to a suitable order of magnitude.

4. The dynamic optimization method for identifying key planning decision points in an integrated energy system according to claim 3, characterized in that, The steps in step S3 for constructing a rolling time-domain optimization model for the integrated energy system and proposing a method for identifying key decision points within the planning period are as follows: S3.

1. Through rolling optimization iteration, the system's operation throughout the year is simulated, and the current decision value function is calculated for each rolling step, forming a sequence of value functions; S3.2 Based on the sequence of value functions obtained in S3.1, the value curves are iterated in reverse order from back to front to construct a continuous value curve that reflects long-term accumulated value within the long planning period. S3.3 Automatically find the peak value on the final value curve to determine the key decision point and realize the identification and output of the key decision point.

5. The dynamic optimization method for identifying key planning decision points in an integrated energy system according to claim 4, characterized in that, Step S3.1, which generates a value function sequence through rolling optimization iteration, specifically includes: 1) Initialization of rolling optimization iteration The process begins at the first time point t0, at which point the system solves the joint optimization model for integrated energy system planning and operation based on step S1 to determine the initial baseline configuration of the system: Where t0 represents the starting time point of the rolling optimization; This indicates that the optimization reaches C at time t0. inv +C om The set of optimal decision quantities for minimizing the value; Based on the value function calculation, the decision value at the first time point t0 is calculated: Where V(t0) represents the decision value at the first time point t0; 2) Rolling Iteration For each subsequent rolling time point t k =t0+k·S, and the following sub-model is solved repeatedly. During the rolling iteration, it is necessary to set the state update equation for the time series variables. In this model, it refers to the initial energy storage capacity E. start,k : Among them, t k =t0+k·S represents the time interval kS between the starting point of subsequent rolling optimization and the initial time t0, where S is called the sliding time window length of rolling optimization; initial energy storage capacity E start,k Indicates t k The initial energy storage capacity at any given time; Indicates at t k Time optimization to reach C inv +C om The set of optimal decision quantities for minimizing the value; Indicates t k+1 Energy storage initial capacity E at time start,k+1 equal to t k The optimized energy storage capacity at time t k The value of +W-1 W is called the time window length for single-cycle rolling optimization; then t k The set of decision quantities optimized at each time step Calculate t k The value of decision-making at any moment 3) Output of the value function sequence After the rolling loop iterates through all the time points of the full cycle planning, the final output is a sparse sequence of state-value function points: {V(t0),V(t1),V(t2),…}.

6. The dynamic optimization method for identifying key planning decision points in an integrated energy system according to claim 5, characterized in that, Step S3.2, which involves constructing a continuous value curve that reflects long-term cumulative value over a long planning period, specifically includes: 1) Construction of continuous value curve interpolation Connecting the discrete value points forms a continuous value curve defined for each hour throughout the entire planning period: Among them, V interp (t) represents the continuous value function curve formed after linear interpolation. 2) Reverse iteration of the value curve The value function, after reverse iteration, satisfies the following formula: V final (T)=V interp (T) V final (t)=V interp (t)+γ·V final (t+1) Among them, V final (T)=V interp (T) indicates that the value functions at the critical point T of the planning cycle should be equal; V interp (t) represents the immediate planning and decision value at time t; γ is the future value discount factor, representing the coefficient by which future value is discounted to the present time; γ·V final (t+1) represents the present value of the sum of all values ​​from time t+1 until the future; V final (t) represents the cumulative value function of the potential of the system planning scheme from now until the future.

7. The dynamic optimization method for identifying key planning decision points in an integrated energy system according to claim 6, characterized in that, The identification and output of key decision points in step S3.3 specifically includes: The goal is to select a set of time points from the value function curve using rigorous mathematical conditions. Each element t in the set * Both are defined as a key decision point; 1) Local optimality condition V(t * )>V(t * -1)∧V(t * )>V(t * +1) Wherein, V(t) * ) represents the critical decision point t * The value of a moment; ∧ represents the logical AND; 2) Significance condition V(t * )≥V threshold Among them, V threshold P represents the minimum peak height threshold; th It refers to the percentile threshold, meaning the minimum and peak thresholds should not exceed the 100th percentile of the value curve. th %; It is an indicator function that takes the value 1 when the condition is true and 0 otherwise; inf{·} represents taking the infimum, that is, the minimum value that satisfies the condition; v represents the minimum value selected after satisfying the condition. 3) Sparsity condition in, D represents the i-th and j-th selected decision points, respectively; min Indicates the minimum time interval; This represents the set of key decision points.

8. The dynamic optimization method for identifying key planning decision points in an integrated energy system according to claim 7, characterized in that, Step S4 proposes a multi-stage dynamic planning scheme for the integrated energy system based on a small number of high-planning-value decision points within a long planning period, specifically including: 1) Data Input Based on the above step S3.3, obtain the set of key decision points. Minimum planned capacity of each device in the initial initial stage in, This represents the initial minimum planning capacity set; These represent the minimum planned capacities for energy storage, photovoltaic power, wind power, and gas turbines, respectively. 2) Objective function Among them, C inv,k C om,k These represent the annualized investment cost and operating cost calculated within the planning window of the current stage k, respectively. 3) Incremental planning constraints The capacity of each device in the current stage k is greater than or equal to the optimal capacity determined in the previous stage k-1: Where, θ k This represents the planned capacity value in the k-th planning stage. This represents the optimal planning value within stage k-1.

Images