A medium and long term random production simulation method, system, medium and device

Abstract

The application discloses a kind of medium and long term random production simulation method, system, medium and equipment, obtain the basic technical data of power system containing renewable energy and the predicted distribution data of renewable energy;Based on the basic technical data of power system and predicted distribution data, to realize system power supply safety, renewable energy consumption, reduce the target of system total cost, in the case where complex constraints are considered Markov decision process is constructed;Approximate dynamic programming method is used to iteratively train Markov decision process, and an optimal set of decisions is searched as a medium and long term operation mode.The application searches the optimal operation mode through the distribution of random factors, avoids the interference of subjective factors, reduces the total cost of the system under the condition of ensuring power supply safety and renewable energy consumption in extreme scenarios, and finally obtains a medium and long term operation mode and evaluation results that perform better than traditional solutions, providing a reference for the safe operation of power systems.

Description

Technical Field

[0001] This invention belongs to the field of power system optimization and dispatching technology, specifically relating to a medium- and long-term stochastic production simulation method, system, medium, and equipment. Background Technology

[0002] As the penetration rate of renewable energy in the power system gradually increases, the power system exhibits highly probabilistic characteristics on a medium- to long-term research scale. The problems of renewable energy curtailment and power supply shortages are intertwined, posing new and significant challenges to grid security.

[0003] In the past, to address the aforementioned medium- to long-term balance issues, medium- to long-term production simulation methods were typically used. These methods usually target monthly, annual, or longer study periods to determine the operation and maintenance schedules of various power generation resources, focusing on the system's power balance and the medium- to long-term regularities of each power generation resource. They are mainly divided into two categories: stochastic production simulation and time-series production simulation. Traditional stochastic production simulation, represented by the equivalent power function method, requires numerous heuristic mechanisms to cope with the large-scale integration of various renewable energy sources into the power system, making the formulation of medium- to long-term operation modes increasingly cumbersome and reducing the scalability of these methods. Traditional time-series production simulation methods, represented by robust methods or scenario-based methods, while achieving optimal decisions under the worst-case disturbance scenario, yield overly conservative results in other scenarios. Scenario-based methods transform various uncertainties in the system into deterministic scenarios for balance analysis; however, with the further increase in the proportion of renewable energy, the number of scenarios to be considered increases dramatically, and the selection of scenarios is often subjective, making it difficult to accommodate diverse scenarios and thus exacerbating the optimization difficulty. Furthermore, the surge in uncertainties poses a challenge to the solvability of large-scale time-series optimization models. Therefore, how to address the over-reliance on heuristic principles and scenario selection in traditional medium- and long-term production simulations, and improve the system's performance in other scenarios while ensuring power supply security in extreme scenarios and renewable energy consumption, has become a key issue that needs to be addressed in this field. Summary of the Invention

[0004] The technical problem to be solved by this invention is to address the shortcomings of the prior art by providing a medium- and long-term stochastic production simulation method, system, medium, and equipment. This invention aims to solve the problem of traditional medium- and long-term production simulation relying excessively on heuristic principles and scenario selection. It significantly reduces the total cost of the system while ensuring power supply security in extreme scenarios and renewable energy consumption, and guarantees the solvability of the model.

[0005] The present invention adopts the following technical solution:

[0006] A medium- to long-term stochastic production simulation method includes the following steps:

[0007] S1. Obtain basic technical data of power systems containing renewable energy and predicted distribution data of renewable energy;

[0008] S2. Based on the basic technical data and predicted distribution data of the power system obtained in step S1, and with the goal of achieving system power supply security, renewable energy consumption and reducing the total system cost, a Markov decision process is constructed under the consideration of complex constraints.

[0009] S3. The Markov decision process obtained in step S2 is iteratively trained using an approximate dynamic programming method to search for the optimal set of decisions as the medium- to long-term operating mode.

[0010] Preferably, the basic technical data of the power system includes unit parameters, and the predicted distribution of renewable energy and load.

[0011] Preferably, the construction of the Markov decision process is as follows:

[0012] Under the premise of adhering to the unit's operating constraints, the dispatch center uses operating information such as new energy output forecasts, water inflow forecasts, and system load forecasts to determine the appropriate parameters. t Simultaneously, combining system status information such as unit maintenance status and water storage status, S t Provide the operation mode decision A t Based on S t and A t The result, through the PS operation process Q t The system state S for the new time period is derived. t+1 The cost of this round of decision-making is quantitatively analyzed by calculating system balance indicators and power generation costs; the above calculations are repeated until the system state transitions from S0 to S... T The ADP search is used to find the decision sequence that minimizes the total cost of the power system and to form the operating mode; the dispatch center determines the operating mode based on the state variable S during time period t. t and operating condition variable G t Make decision A t Then, the power system balance calculation is performed using probabilistic sequence arithmetic rules to derive the system state S for time period t+1. t+1 .

[0013] More preferably, the power system balance calculation specifically includes:

[0014] By sorting and renumbering the generating units according to their power generation economics, an ordered matching set Q can be obtained. t If the economic order is n t The generating capacity of the unit in time period t is and and These represent the distribution of local residual load electricity and the distribution of electricity consumption after the unit participates in balancing, respectively. Convolution operation Θ and crossover operation are introduced. The order within time period t is n t Unit load balancing calculation

[0015] More preferably, for new energy units, the operating condition variable G is directly applied. t WD in w,t and PV v,t The process is transformed into a discrete sequence. For thermal power resources, the impact of random unit failures and maintenance decisions on available capacity is fully considered to obtain the sequence. For hydropower resources, considering the state variable S during time period t... t Operating condition variable G t Decision variable A t Determine the water storage and electricity sequence C of the reservoir in time period t+1. r,t+1 If reservoir r corresponds to N r There are n hydroelectric generating units, which are simultaneously ordered as n within the entire system. t The hydroelectric power units in this N r The economic order among the individual units is m r,t The discrete sequence of the power generation of the hydropower unit in time period t is obtained. The hydropower unit participates in system balancing, and the discrete sequence of the unit's power generation is calculated. Then, the total generating capacity of other units not involved in balancing related to reservoir r is calculated. The dispatch center ensures that the total capacity of units not undergoing maintenance during time period t is greater than the maximum load power value during that time period and has a preset margin. For the economic order n... t The unit's power generation cost during time period t The calculations are performed, and then the results of all units are summed to obtain the system's power generation cost in time period t.

[0016] More preferably, the evaluation metrics for system balance include the expected energy deficit (EUE) and the load shedding ratio (LOLR) for time period t, as follows:

[0017]

[0018] Among them, i t and I t These are discrete sequences The state sequence number and the number of states in L; t,i and It is a discrete sequence The state result and probability of the i-th state; p(·) evaluates the input value, returning 0 if it is positive, otherwise returning the input value; j t and Jt These are the scene number and the total number of scenes; and The jth t Power loss and total load power in each scenario.

[0019] More preferably, the state variable S t :

[0020] S t ={U g,t U h,t C r,t}

[0021] Among them, U g,t and U h,t C represents the remaining downtime days for thermal power unit g and hydropower unit h during time period t. r,t This represents the distribution of water volume and electricity stored in reservoir r during time period t; t is the corresponding time period number.

[0022] Operating condition variable G t express:

[0023] G t ={WD w,t ,PV v,t H r,t ,L t}

[0024] Among them, WD w,t It is the predicted power output distribution of wind turbine unit w in time period t; PV v,t It is the predicted output distribution of the photovoltaic unit v in time period t; H r,t It is the predicted distribution of water and electricity generated by reservoir r during time period t; L t This is the predicted distribution of load over time period t; t is the time period number.

[0025] Decision variable A t :

[0026] A t ={u g,t ,u h,t ,c r,t}

[0027] Among them, u g,t and u h,t This represents the maintenance decision result for thermal power unit g and hydropower unit h during time period t. A value of 0 indicates a maintenance decision, otherwise 1. Here, g and h are the original unit numbers; c r,t It is the decision result of the water storage capacity of reservoir r at time t;

[0028] State S t The optimal total cost function V of the system under the following conditions t (St ):

[0029]

[0030] Where E(·) is the expected function; γ is the return factor ranging from 0 to 1;

[0031] State S t Execution Decision A t And transferred to S t+1 The instantaneous cost C generated t (S t A t ):

[0032] C t (S t A t ) = GC(S t A t )+α·HC(S t A t )

[0033] +β·NC(S t A t )+χ·LC(S t A t )

[0034] Where GC(·), HC(·), NC(·), and LC(·) are the power generation cost function, water curtailment cost function, renewable energy curtailment cost function, and load shedding cost function, respectively, and α, β, and χ are the corresponding cost coefficients.

[0035] Preferably, in step S3, for time period t of the nth iteration of training, the scheduling center bases the state variable S... t Make the optimal decision to transition the power system state to S. t+1 ;

[0036] The scheduling center first determines an initial solution in the MDP framework and obtains the corresponding initial state sequence through simple methods or historical experience. Then, it changes a small increment above and below the state sequence to form a corridor. Then, it searches within the corridor using conventional dynamic programming and uses the result of this optimization as the initial solution for the next search. The process is repeated until the convergence condition is met.

[0037] ADP based on a table model can solve the original problem in polynomial time complexity. By recording the solutions to subproblems, it avoids repeated calculations of subproblems. If the dispatch center is in state S of the power system... t Make a decision A t After actual execution, the new state of the system in the next time period and the related discrete sequence are uniquely determined.

[0038] More preferably, the optimal operating mode of the system is as follows:

[0039]

[0040] Among them, V t (n) (S t ) and V t (n-1) (S t The states S obtained in the nth and (n-1)th iterations are respectively. t The corresponding value function; σ is a constant ranging from 0 to 1.

[0041] Secondly, embodiments of the present invention provide a medium- to long-term stochastic production simulation system, comprising:

[0042] The data module acquires basic technical data of power systems containing renewable energy and predicted distribution data of renewable energy.

[0043] The module is built based on basic power system technical data and predicted distribution data, with the goals of achieving system power supply security, renewable energy consumption, and reducing total system cost. It constructs a Markov decision process under complex constraints.

[0044] The simulation module uses an approximate dynamic programming method to iteratively train the Markov decision process and search for the optimal set of decisions as the medium- to long-term operating mode.

[0045] Thirdly, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described medium- and long-term stochastic production simulation method.

[0046] Fourthly, embodiments of the present invention provide a computer-readable storage medium including a computer program, which, when executed by a processor, implements the steps of the above-described medium- and long-term randomized production simulation method.

[0047] Compared with the prior art, the present invention has at least the following beneficial effects:

[0048] This invention presents a medium- to long-term stochastic production simulation method aimed at overcoming the limitations of current production simulation technologies in handling diverse scenarios. In terms of methodology, this invention cleverly uses stochastic production simulation results based on probabilistic sequences (PS) and a balance index as the cost function of approximate dynamic programming (ADP). Guided by PS, ADP can autonomously search for the optimal operating mode. Its unique computational mode allows ADP to output the current optimal result before convergence, avoiding the solvability problem inherent in traditional optimization models. Furthermore, this invention introduces relevant acceleration techniques to improve computational efficiency. In terms of innovative effects, ADP and PS complement each other. ADP endows PS-based stochastic production simulation with the ability to optimize the search, reducing the reliance on heuristic mechanisms while considering numerous random factors and improving the scalability of its models. PS and its computational theory enable ADP to be free from the negative impact of scenario selection or sampling, and instead determine the optimal operating mode directly based on the distribution of random factors, thereby avoiding interference from human subjective factors. The results can significantly reduce the total system cost while ensuring power supply security in extreme scenarios and renewable energy consumption, providing a reference for the safe operation of power systems.

[0049] Furthermore, basic power system data and predicted distribution data of renewable energy are prerequisites for constructing Markov decision processes for the power system.

[0050] Furthermore, the Markov decision process is based on finding the optimal operating mode through approximate dynamic programming.

[0051] Furthermore, the medium- and long-term stochastic production simulation calculation system proposed in this invention is used to obtain operating modes and evaluation indicators. The former includes unit maintenance plans, power generation plans, and water use plans for each period, while the latter includes total system cost, renewable energy curtailment rate, and system load shedding rate. These results are the basis for the safe operation of the power system.

[0052] It is understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0053] In summary, this invention obtains the optimal operating mode through the distribution search of random factors, avoiding interference from subjective factors. It reduces the total system cost while ensuring power supply security in extreme scenarios and renewable energy consumption. Ultimately, it yields a medium- to long-term operating mode and evaluation results that outperform traditional solutions, providing a reference for the safe operation of power systems.

[0054] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the process of the present invention;

[0056] Figure 2 A diagram of the Markov decision process in a power system;

[0057] Figure 3 Flowchart for medium- and long-term random production simulation calculation;

[0058] Figure 4 The diagram shows the production simulation results of the method under the average scenario, where (a) is the maintenance decision result of the method of the present invention under the average scenario, (b) is the water storage decision result of the method of the present invention under the average scenario, and (c) is the power generation result of the method of the present invention under the average scenario.

[0059] Figure 5 The figures show the production simulation results of the method under extreme scenarios, where (a) is the production simulation result of the method of the present invention under a severe power shortage scenario, and (b) is the production simulation result of the method of the present invention under a scenario of abundant renewable energy generation.

[0060] Figure 6 This is a diagram illustrating the iterative process of the algorithm.

[0061] Figure 7 A schematic diagram of a computer device provided in an embodiment of the present invention;

[0062] Figure 8 This is a block diagram of a chip provided according to an embodiment of the present invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0065] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0066] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this invention generally indicates that the preceding and following objects have an "or" relationship.

[0067] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0068] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0069] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0070] This invention provides a medium- to long-term stochastic production simulation method. It constructs a Markov decision process for the power system based on comprehensive consideration of various complex constraints, where the system state for each time period corresponds to multiple probabilistic sequences. Before the start of a new time period, a decision on the system's operating mode must be made, and the power system is then transitioned to the next state based on the probabilistic sequences and their computational theory. Next, approximate dynamic programming is introduced to train the Markov decision process, and related techniques are used to improve the model's training efficiency. This invention avoids the negative impact of subjective factors in traditional medium- to long-term production simulations, significantly reducing the total system cost while ensuring power supply security in extreme scenarios and the absorption of renewable energy, providing a reference for the safe operation of power systems.

[0071] For power systems with large-scale clean energy access, this invention combines probabilistic sequence and approximate dynamic programming to propose a new medium- and long-term production simulation method.

[0072] In terms of implementation, this invention uses the results of random production simulation based on sequence operation theory as the cost function of approximate dynamic programming, enabling approximate dynamic programming to automatically search for the optimal operating mode. This invention can output the current optimal result before the approximate dynamic programming converges, thereby avoiding the solvability problem of traditional optimization models. At the same time, this invention introduces relevant acceleration technologies to improve computational efficiency.

[0073] In terms of innovative effects, approximate dynamic programming enables stochastic production simulation based on sequence operation theory to have stronger optimization search capabilities. This invention can reduce the dependence on heuristic mechanisms while considering numerous random factors, thus improving the scalability of its model. Probabilistic sequence theory makes approximate dynamic programming no longer subject to the difficulties of scenario selection or sampling, but can directly determine the optimal operating mode based on the predicted distribution of random factors, thereby avoiding interference from human subjective factors. The result is that while ensuring power supply security in extreme scenarios and renewable energy consumption, it can significantly reduce the total system cost.

[0074] Please see Figure 1 This invention discloses a medium- to long-term stochastic production simulation method that comprehensively considers the changes in stochastic factors such as load, water inflow, wind power, and photovoltaic power at different time periods, as well as system safety constraints. With the goals of ensuring power supply security, renewable energy consumption, and reducing total system costs, it constructs a medium- to long-term stochastic production simulation process. This method can directly search for the optimal operating mode through the distribution of stochastic factors, avoiding the negative impact of subjective factors. While ensuring power supply security and renewable energy consumption in extreme scenarios, it can significantly reduce total system costs, providing a reference for the safe operation of power systems. The specific steps are as follows:

[0075] S1. Obtain basic technical data of power systems containing renewable energy and predicted distribution data of renewable energy;

[0076] Basic technical data includes: operating parameters of various generating units, and predicted distribution of renewable energy and load.

[0077] S2. Based on the basic technical data and predicted distribution obtained in step S1, a Markov decision process is constructed for the power system under full consideration of complex constraints, which serves as the overall framework for subsequent calculations.

[0078] Please see Figure 2 The state change process of a power system consists of T time periods, each corresponding to a system state, S0 and S2. T These represent the initial and final states of the system; here, we take the system state transition from time t to t+1 as an example. Figure 2 To illustrate, at this point, t = 0, 1, ..., T-1.

[0079] First, the dispatch center must, under the premise of complying with the unit's operating constraints, base its decisions on operating condition information such as the predicted output of new energy sources, the predicted water inflow, and the predicted system load. t Simultaneously, combining system status information such as unit maintenance status and water storage status, S t Provide the operation mode decision A t ;

[0080] Next, based on S t and A t The result, through the PS operation process Q t The system state S for the new time period is derived. t+1 Furthermore, the costs of this round of decision-making are quantitatively analyzed by calculating system balance indicators and power generation costs.

[0081] Then, repeat the above calculations until the system state transitions from S0 to S... T ;

[0082] Finally, ADP is used to search for the decision sequence that minimizes the total cost of the system and to determine the operating mode.

[0083] for Figure 2 In the constructed MDP, the system's uncertainties and their distributions will be used as state variables S. t or operating environment variable G t The method of this invention directly participates in production simulation calculations, no longer relying on scenario selection and sampling. This is the main difference between the method of this invention and traditional production simulation strategies.

[0084] To construct the above MDP framework, basic elements need to be defined, including the state variable S. t Operating condition variable G t Decision variable A t Value function V t The specific content is as follows:

[0085] 1. Definition of state variables

[0086] State variable S t This reflects the state of the system at time t:

[0087] S t ={U g,t U h,t C r,t} (1)

[0088] Among them, U g,t and U h,t This represents the remaining downtime days for thermal power unit g and hydropower unit h during time period t; if they can be started, the value is 0. r,t It represents the distribution of water volume and electricity stored in reservoir r during time period t; t is the corresponding time period number.

[0089] 2. Definition of operating condition variables

[0090] Operating condition variable G t This represents the system's operating environment during time period t, primarily composed of predicted results for renewable resources and system load.

[0091] G t ={WD w,t ,PV v,t H r,t ,L t} (2)

[0092] Among them, WD w,t It is the predicted power output distribution of wind turbine unit w in time period t; PV v,t It is the predicted output distribution of the photovoltaic unit v in time period t; H r,t It is the predicted distribution of water and electricity generated by reservoir r during time period t; L t It is the predicted distribution of load in time period t; t is the time period number.

[0093] 3. Definition of decision variables

[0094] Decision variable A t This reflects the operating mode selected by the dispatch center during time period t:

[0095] A t ={u g,t ,u h,t ,c r,t} (3)

[0096] Among them, u g,t and u h,tThis represents the maintenance decision results for thermal power unit g and hydropower unit h during time period t. A value of 0 indicates a maintenance decision, otherwise 1. Here, g and h are the original unit numbers, which will be reordered based on unit economics in Chapter 2; c r,t It is the decision result of the reservoir r storing water and electricity at time t.

[0097] 4. Definition of value function

[0098] Within the framework of dynamic programming, the optimal operating cost can be obtained by recursively solving the Bellman equation:

[0099]

[0100] Among them, V t (S t ) is in state S t The system's optimal total cost function is defined as follows: E(·) is the expected value function; γ is a return factor ranging from 0 to 1. If γ = 0, the value function is only related to the current time cost; the closer γ is to 1, the more the value function considers future returns and the more forward-looking the optimization becomes. t (S t A t ) is from state S t Execution Decision A t And transferred to S t+1 The instantaneous cost incurred is expressed as follows:

[0101]

[0102] Wherein, GC(·), HC(·), NC(·), and LC(·) are the power generation cost function, the water curtailment cost function, the renewable energy curtailment cost function, and the load shedding cost function, respectively. The values ​​of each of these can be determined through the state transition operation described later. α, β, and χ are the corresponding cost coefficients.

[0103] The dispatch center determines the time period t based on the state variable S. t and operating condition variable G t Make decision A t Then, system equilibrium calculations need to be performed using probabilistic sequence arithmetic rules to derive the system state S for time period t+1. t+1 .

[0104] Here, we choose time period t as an example for the balancing operation. First, the generating units are sorted and renumbered according to their power generation economics, resulting in an ordered matching set Q. t It is expressed as follows:

[0105]

[0106] Where, N tThis represents the number of available power generation devices during time period t. The order within time period t is n t The balancing calculation of the generating units and the local load.

[0107] If the economic order is n t The generating capacity of a unit in time period t is defined as and and These are defined as the distribution of local residual load electricity after the unit participates in balancing and the distribution of electricity consumed after the unit participates in balancing, respectively. Then, by introducing the convolution operation "Θ" and the product operation... back It is represented as:

[0108]

[0109] for and The calculation method first takes the maximum common factor of the power generation of each unit in time period t as the common off-track length ΔE. t Then for the new order n t The unit, if its rated power is Then its power generation distribution will be in the range Discretized into Each state is used to obtain the sequence. Here T t It is the length of time interval t. Obtained in the same way. For calculation The value of is introduced here by a function. <x>Let x represent integers not exceeding x, then we have:

[0110]

[0111] The number of states of the discrete sequence is derived based on formula (8). The generateable power or predicted load power in the i-th state is defined as:

[0112]

[0113] in accordance with The probability of their occurrence constitutes a discrete sequence of the unit's power generation and the predicted load power, which is used to approximate the actual distribution.

[0114] For new energy generating units, the operating condition variable G in equation (2) can be directly applied. t WD in w,t and PV v,t Transform it into a discrete sequence using the method described above, for ease of discussion. In the following text, it also refers to its corresponding discrete sequence.

[0115] For thermal power resources, taking into full account the impact of random unit failures and maintenance decisions on available capacity, therefore:

[0116]

[0117] in, Represents the probability of the i-th state occurring in a discrete distribution sequence, when hour The values ​​are all 0; and The economic order is n r,t The probability of normal operation and maintenance status of the unit during time period t; The value of 0 or 1 indicates whether the unit is under maintenance or not, and the decision is made by the dispatch center at time t. t It is estimated that each unit will only need to be inspected once.

[0118] For hydropower resources, considering the state variable S during time period t... t Operating condition variable G t Decision variable A t Given that this has been determined, the water storage and electricity sequence C of the reservoir at time t+1 is... r,t+1 Represented as:

[0119]

[0120] Among them, symbols "Θ" These represent the convolution operation, convolution difference operation, and crossover operation, respectively. It is a discrete distribution sequence of the actual water volume stored in reservoir r during time period t; It is the discrete distribution sequence of the total power generation of the corresponding hydropower units in reservoir r during time period t; H r,t It is the discrete distribution sequence of the predicted water and electricity flow from reservoir r in time period t; c r,t It is the discrete decision sequence of the dispatch center for the water and electricity stored in reservoir r during time period t, which can be expressed by c in equation (3). r,t The calculation shows that the specific method involves first determining the distance step length Δc based on actual needs. t and decision ceiling Then, the number of states of the discrete decision sequence in time period t can be obtained. Then initialize a sequence of all-zero elements with the same number of states, and place c in this sequence. r,t The probability of the nearest discrete state being found is set to 1, thus obtaining c. r,t To avoid the above transformation process, decision variable A can be pre-defined. t c in r,t Replace with c r,t .

[0121] To ensure that the results meet the actual constraints, the discrete state sequence C of reservoir r in time period t+1 is obtained through equation (12). r,t+1 The following calculations are then required:

[0122] First, remove C. r,t+1 The state that exceeds the actual limit is identified, and the probability of the remaining state is reset according to the ratio of the original probability. After processing, the sum of the probabilities of each state in the sequence is 1.

[0123] Then, in order to ensure that the final state of reservoir r returns to its initial value as much as possible under complex and diverse environments, the dispatch center automatically selects a feasible decision range within the final time interval t=T to make C r,t+1 The decision action whose expected value is closest to the initial value.

[0124] In addition, those that make C r,t+1 Decisions where the expected value significantly exceeds the actual limit will not be considered by the scheduling center. This approach ensures that the results reflect reality while significantly reducing the number of system states and decisions the scheduling center needs to consider, thereby improving the overall system computational efficiency.

[0125] Based on the above results, if reservoir r corresponds to N r There are n hydroelectric generating units, which are simultaneously ordered as n within the entire system. t The hydroelectric power units in this N r The economic order among the individual units is m r,t Then, the discrete sequence of the power generation of the hydropower unit in time period t. Represented as:

[0126]

[0127] Among them, symbols It is a product operation, indicating At the same time, it is subject to multiple restrictions such as hydrological conditions and operating methods; The first m corresponding to reservoir r r,t -The total power generation of the remaining units after one hydropower unit participates in system balancing; It is the power generation of the hydropower unit when only maintenance and capacity are considered. Its calculation method and required parameters are the same as those in equation (10).

[0128] exist Once determined, the hydropower unit participates in system balancing, and the discrete sequence of the unit's power generation is calculated using equation (8). Then, the total generating capacity of other units not involved in the balancing process related to reservoir r is calculated using the following method:

[0129]

[0130] To ensure power balance, the dispatch center must ensure that the total capacity of units not undergoing maintenance during time period t is greater than the maximum load power value during that time period, with a preset margin. Based on power constraints and the results of equations (6) to (14), the economic order n is determined by equation (18). t The unit's power generation cost during time period t The calculation is performed, and then the results of all units are summed to obtain the power generation cost of the system in time period t in equation (5).

[0131]

[0132] in, is n t The power generation cost function of the unit in time period t; and It is a discrete distribution sequence The state sequence number and the number of states; It is a discrete sequence Lidi The amount of electricity consumed in each state.

[0133] Regarding the evaluation of the calculation results, here we complete Q in equation (6) through equations (7) to (18). t To evaluate the system's balancing effect, we first present two evaluation indicators for time period t: Expected Unserved Energy (EUE) and Loss of Load Ratio (LOLR). The calculation methods are as follows:

[0134]

[0135] Among them, i t and I t These are discrete sequences The state sequence number and the number of states in L; t,i and It is a discrete sequence The state result and probability of the i-th state; p(·) evaluates the input value, returning 0 if it is positive, otherwise returning the input value; j t and J t These are the scene number and the total number of scenes, the latter being a relatively large value; and The jth t The power loss and total load power of the method proposed in this invention in various scenarios.

[0136] Considering that security is the primary requirement of the system, EUE t It must be lower than the preset threshold value δ t Only the corresponding decision will be considered; if none of these conditions are met, the lowest EUE will be retained first. t The decisions corresponding to the indicators. Then, the retained EUE. t Multiplying it by the relevant cost coefficient χ can directly serve as the loss of load cost in equation (5). The cost of abandoning water and the cost of abandoning new energy can be obtained in a similar way.

[0137] The above results are added to the calculated power generation cost to obtain the instantaneous revenue C for time period t in equation (5). t (S t A t Ultimately, a complete MDP framework was formed for the power system.

[0138] Finally, in order to facilitate the comparison of the strategy proposed in this invention with other methods, a more general evaluation index LOLR is given in equation (22), and the curtailment rate of renewable energy is obtained in a similar manner. After obtaining the operation mode through a certain method, a large number of specific scenarios will be extracted based on the distribution of each random factor and the average value of the relevant indexes will be calculated to evaluate the performance of the operation mode.

[0139] S3. The Markov decision process obtained in step S2 is iteratively trained using an approximate dynamic programming method to search for the optimal set of decisions as the medium- to long-term operating mode, and the results are evaluated.

[0140] Regarding the compression of state and action spaces, this invention utilizes the operational constraints of each power generation resource to eliminate invalid states and decision spaces, and introduces discrete differential dynamic programming based on ADP. Through these methods, the compression of state and action spaces is achieved, thereby significantly improving computational efficiency.

[0141] First, for time period t during the nth iteration of training, the scheduling center needs to base its decision on the state variable S. t Make the optimal decision to transition the power system state to S. t+1 To ensure that the system state transitions closely resemble reality, relevant constraints were added to each power generation resource as described above. States and actions that do not meet the requirements will be directly eliminated; therefore, S is used as the basis for this. t The new state S that can be transitioned to from the starting point t+1 It is finite, and at the same time in S t Available decision variables A under the given state t It is also limited, thereby enabling a significant compression of the system state and action space, ultimately improving computational efficiency.

[0142] Then, a discrete differential dynamic programming (MDP) method is introduced to reduce computational storage and processing time. The scheduling center can first determine an initial solution in the MDP framework using a simple method or historical experience, and obtain the corresponding initial state sequence. Then, a small increment is made above and below this state sequence to form a corridor. Then, conventional dynamic programming is used to search within this corridor. The result of this optimization can also be used as the initial solution for the next search. This process is repeated iteratively until the convergence condition is met.

[0143] In terms of constructing and updating value functions, ADP improves training efficiency by constructing specific approximation models of value functions. ADP based on table models can solve the original problem in polynomial time complexity. By recording the solutions to subproblems, it avoids repeated calculations of subproblems, which is a convenient and effective solution method. Therefore, this paper chooses this model to construct and train value functions.

[0144] Before formally introducing the table model, based on the previous derivation, if the dispatch center is in state S of the power system... t Make a decision A t After actual execution, the new state of the system in the next time period and the related discrete sequence can be uniquely determined, that is, the state transition of the MDP in this paper is deterministic, so formula (4) is simplified to:

[0145] V t (S t ) = C t (S t A t )+γV t+1 (S t+1 |S t A t (20)

[0146] Figure 3 The training iteration process of the table model is given, and it will be explained below:

[0147] First, the value function table is initialized, at which point the elements in the table are defined as V. t (0) (S t ), representing the state variable S during time period t. t The corresponding initial value of the value function;

[0148] Then, for the nth iteration training time period t, the relevant continuous variables need to be transformed into discrete sequences according to the method described above, and then based on formula (20), the state S of the power system in time period t is determined. t The optimal decision is determined by the value function table constructed in the previous training.

[0149]

[0150] Where n is the number of iterations; argmin{·} is the function used by the scheduling center to make the optimal decision selection according to the table model training method; It will return the value function result of the corresponding state in time period t+1 based on the value function table obtained from the previous training.

[0151] Determine the optimal decision according to formula (21) After that, state S t The corresponding approximate function estimate The result is calculated using formula (20), and then updated according to formula (22) and recorded in the value function table. This process is repeated until the value function table converges, and the value function table corresponds to the optimal operating mode of the system.

[0152]

[0153] Among them, V t (n) (S t ) and V t (n-1) (S t The states S obtained in the nth and (n-1)th iterations are respectively. t The corresponding value function; σ is a constant ranging from 0 to 1. Adjusting σ changes the proportion of the original value function in this update.

[0154] In another embodiment of the present invention, a medium- to long-term stochastic production simulation system is provided. This system can be used to implement the above-mentioned medium- to long-term stochastic production simulation method. Specifically, the medium- to long-term stochastic production simulation system includes a data module, a construction module, and a simulation module.

[0155] The data module acquires basic technical data of the power system containing renewable energy and predicted distribution data of renewable energy.

[0156] The module is built based on basic power system technical data and predicted distribution data, with the goals of achieving system power supply security, renewable energy consumption, and reducing total system cost. It constructs a Markov decision process under complex constraints.

[0157] The simulation module uses an approximate dynamic programming method to iteratively train the Markov decision process and search for the optimal set of decisions as the medium- to long-term operating mode.

[0158] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used for the operation of a medium- to long-term random production simulation method, including:

[0159] Acquire basic technical data of power systems containing renewable energy and predicted distribution data of renewable energy; based on the basic technical data and predicted distribution data of power systems, construct a Markov decision process under complex constraints with the goals of achieving system power supply security, renewable energy consumption and reducing total system cost; use an approximate dynamic programming method to iteratively train the Markov decision process and search for the optimal set of decisions as the medium- and long-term operation mode.

[0160] In another embodiment of the present invention, a storage medium is also provided, specifically a computer-readable storage medium (memory). This computer-readable storage medium is a memory device in a terminal device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device.

[0161] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the medium- and long-term random production simulation method in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps:

[0162] Acquire basic technical data of power systems containing renewable energy and predicted distribution data of renewable energy; based on the basic technical data and predicted distribution data of power systems, construct a Markov decision process under complex constraints with the goals of achieving system power supply security, renewable energy consumption and reducing total system cost; use an approximate dynamic programming method to iteratively train the Markov decision process and search for the optimal set of decisions as the medium- and long-term operation mode.

[0163] Please see Figure 7 The terminal device is a computer device. In this embodiment, the computer device 60 includes a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When executed by the processor 61, the computer program 63 implements the fluid composition calculation method in the reservoir stimulation wellbore of this embodiment. To avoid repetition, these details are not elaborated here. Alternatively, when executed by the processor 61, the computer program 63 implements the functions of each model / unit in the fluid composition calculation system in the reservoir stimulation wellbore of this embodiment. To avoid repetition, these details are not elaborated here.

[0164] Computer device 60 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will understand that... Figure 7 This is merely an example of computer device 60 and does not constitute a limitation on computer device 60. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device may also include input / output devices, network access devices, buses, etc.

[0165] The processor 61 may be a central processing unit (CPU), or other general-purpose processors, CPUs, graphics processing units (GPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, quantum computing-based data processing logic units, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0166] The memory 62 can be an internal storage unit of the computer device 60, such as a hard disk or RAM of the computer device 60. The memory 62 can also be an external storage device of the computer device 60, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on the computer device 60.

[0167] Furthermore, the memory 62 may include both internal storage units of the computer device 60 and external storage devices. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 can also be used to temporarily store data that has been output or will be output.

[0168] Any references to memory, databases, or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0169] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0170] Please see Figure 8 The terminal device is a chip. In this embodiment, the chip 600 includes a processor 622, which may be one or more, and a memory 632 for storing computer programs executable by the processor 622. The computer program stored in the memory 632 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processor 622 may be configured to execute the computer program to perform the generalizable monocular absolute depth map estimation method described above.

[0171] Additionally, chip 600 may also include a power supply component 626 and a communication component 650. The power supply component 626 can be configured to perform power management of chip 600, and the communication component 650 can be configured to enable communication of chip 600, such as wired or wireless communication. Furthermore, chip 600 may also include an input / output interface 658. Chip 600 can operate on an operating system stored in memory 632.

[0172] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0173] Case Analysis

[0174] The effectiveness and advancement of the proposed method were verified using the XJTU-ROTS testing system. This system includes 32 thermal power units, 10 hydropower units, 5 wind power systems, and 4 photovoltaic power systems. To better reflect the impact of the randomness of renewable energy on the power system, the unit capacities were appropriately adjusted. After adjustment, the installed capacity of thermal power accounts for approximately 45%, hydropower approximately 28%, and renewable energy approximately 27%, with a total installed capacity of 14100MW and a maximum annual load of 10421MW. This paper uses Matlab R2022a software to call CPLEX for simulation analysis, running on a 64-bit Windows system and an Intel Core i7-7700@3.6GHz environment. The following sections will verify the method of this invention through production simulation results and system production indicators in some scenarios.

[0175] 1) Analysis of production simulation results for average scenarios

[0176] To demonstrate the effectiveness and superiority of the method of this invention, this section describes the operation modes of the test system at the year-to-month time scale using the method of this invention, the scenario method, and the robust method. The results of these operation modes on the system under certain specific scenarios are then studied and compared. To realistically reflect the impact of various random factors on the system, this paper uses real data from a southern province to set the monthly changes in load and power generation resources in the test system.

[0177] Please see Figure 4 The figure shows the production simulation results of the method of the present invention under a certain average scenario, in which the load and each power generation resource are at the average value and have the highest probability of occurrence.

[0178] exist Figure 4 In (a), the maintenance of thermal power plants is concentrated in the low load period in February and the flood season in June. No thermal power units are maintained during the high load period at the end of the year, which ensures power supply safety and avoids the negative impact of unit failures. The maintenance of hydropower plants is concentrated in the low load period in February and the dry season at the end of the year. No maintenance is arranged during the flood season from May to October, which ensures the full utilization of hydropower resources.

[0179] exist Figure 4 In (b), the total available electricity of the reservoirs showed a downward trend during the period of low water inflow in the first half of the year, which indicates that the reservoirs were using the pre-stored electricity to support the system balance. As the system entered the flood season in June, the power generation and total available electricity of the reservoirs increased significantly. For different reservoirs, the No. 1 and No. 2 reservoirs, which are more economical, had larger adjustment ranges, while the No. 3 reservoir, which is less economical and subject to minimum water storage restrictions, changed more gradually. After adjustment, each reservoir basically returned to its initial state.

[0180] exist Figure 4 In section (c), hydropower, wind power, and photovoltaic resources complement each other, resulting in a more even distribution of power generation from thermal power units across months. This avoids the unfavorable situation where the available power generation capacity is low during a particular period, but the power supply workload is excessive. Regarding system costs, the variation range of each random factor is limited to near its mean. The average cost of each method is then calculated for each month. Finally, all results are converted to per-unit values ​​based on the results of the method described in this invention. The final results are as follows: Figure 4 (c) shows the three broken lines at the bottom. The above results demonstrate that the method of the present invention can fully consider the changes in various system constraints and random factors, ensure the safety of system power supply and the absorption of renewable energy, and at the same time reduce the total cost of the system to a low level.

[0181] 2) Analysis of Production Simulation Results in Extreme Scenarios

[0182] Please see Figure 5 The figure shows the production simulation results of the method of the present invention under two extreme scenarios: severe power shortage and large-scale generation of new energy sources. The power generation of various power generation resources is described by columns of different shades. Figure 5 The darkest bars at the bottom represent actual power generation, indicating the amount of electricity actually generated by the corresponding unit during the corresponding time period; the lighter bars represent reserve power, which is obtained by multiplying the capacity of the units not under maintenance during the corresponding time period by the length of that time period; the lightest bars represent maintenance power, which can be obtained by multiplying the capacity of the units under maintenance during the corresponding time period by the length of that time period.

[0183] exist Figure 5 In scenario (a), the system's load and power generation resources for each month are located at the upper and lower boundaries of their respective distributions, thus the system faces significant power supply pressure. Under these circumstances, the solution proposed in this invention enables the full absorption of renewable energy, although... Figure 5 (a) The hydropower units show a relatively large reserve of electricity, but due to constraints on inflow and reservoir capacity, the monthly power generation of the hydropower units is very limited. It is worth noting that the method of this invention, through the reasonable coordination of water storage decisions and maintenance decisions, enables… Figure 5 The system in (a) has adequate generating capacity in each month, avoiding power supply security issues caused by various random factors, including unit failures. To compare with existing solutions, in Figure 5 (a) provides additional load loss results for the scenario method and the robust method under this severe power shortage scenario. It can be seen that the method of the present invention and the robust method can ensure system balance at all time periods, but the scenario method has a large number of load loss cases.

[0184] exist Figure 5 In scenario (b), the system's load and power generation resources for each month are located at the lower and upper boundaries of the corresponding distribution areas, respectively. Therefore, in this scenario, renewable energy generation is abundant, and the system needs to absorb renewable energy as much as possible while reducing power generation costs. In this case, the method of this invention keeps the amount of renewable energy curtailment at a relatively low level through reasonable reservoir regulation decisions. The scenario-based method results in more renewable energy curtailment, while the robust method exhibits a large amount of renewable energy curtailment during the flood season.

[0185] Scenario-based methods typically require manual selection of scenarios and the assignment of scenario weights based on their probability of occurrence. This makes it prone to overlooking certain extreme scenarios with low occurrence probabilities, leading to... Figure 5 In such cases, it is difficult to guarantee the system's power supply security or the absorption of renewable energy. The robustness principle focuses on extreme scenarios of severe power shortages, therefore it... Figure 4 Average scenarios and Figure 5 (b) In scenarios involving large-scale renewable energy generation, it is difficult to achieve good performance. The traditional solutions described above all formulate annual operating modes for specific scenarios and do not consider the impact of unit failures, which is clearly unsuitable for systems with numerous and highly variable random factors. The method of this invention is no longer limited to specific scenarios but directly searches for the optimal operating mode based on the distribution of various random factors. This allows for a more objective and comprehensive analysis of the problem and considers the operating results under various scenarios as much as possible. Therefore, the method of this invention not only... Figure 4 (c) has a lower total system cost in the average scenario, and in Figure 5 It also performs best in extreme scenarios.

[0186] 3) Production simulation index analysis

[0187] To further compare the overall performance of the various methods, this section sets up three complete production simulation schemes. First, the operating mode is determined using the method of this invention, the scenario method, and the robust method at the year-to-month time level. To accurately evaluate the effect of a certain operating mode on system balance, a large number of specific scenarios are extracted based on the distribution of random factors in the system. Then, within the month, the unit combination is optimized hourly based on the operating mode and the extracted scenarios. Finally, the production simulation indicators are determined based on the average value of the results, as shown in Table 1.

[0188] Table 1 Comparison of system production simulation indicators. Fig. 3 Comparison of results of production simulation.

[0189]

[0190] In Table 1, the total system cost of the method of this invention is set as the baseline value, and the per-unit value of the total cost of each method can be obtained, which can more intuitively reflect the cost differences between the methods. As can be seen from the above results, the three production simulation indicators of the method of this invention all reach or approach the minimum level, exhibiting the best overall performance. The scenario method achieves lower cost results, but the neglect of certain extreme scenarios leads to a significantly higher probability of renewable energy curtailment and load shedding. The robust method can fully guarantee power supply security, but its overly conservative strategy results in excessively high total system cost and renewable energy curtailment probability. Table 1, from an overall perspective, confirms the analysis above. As the proportion of renewable energy increases, the randomness of the system and the number of scenarios that need to be considered both increase significantly. If the operating mode is determined based on only a few scenarios, it will inevitably lead to a situation where the scenario method and the robust method in Table 1 are inadequate in their consideration of one aspect while neglecting another. The method of this invention, which is directly based on the distribution of random factors, can better take into account the operating results under various scenarios.

[0191] To demonstrate the computational performance of the method of this invention, a comparison of the iterative processes of the method of this invention and conventional methods is provided here. Figure 6 As shown by the red and blue lines, their total costs have been standardized based on the final convergence result. The method of this invention first quickly obtains an initial operating mode using traditional methods, and then searches for the optimal operating mode around the initial result using ADP. Therefore, the method of this invention has a lower total cost from the beginning of the iteration, and converges to the optimal solution in only about 400 iterations. Traditional methods directly use ADP to search for the optimal operating mode, but because the search range is too large and lacks a clear search direction, the convergence speed is significantly slower. It is worth noting that regardless of whether the method of this invention or the traditional method is used for iteration, they can stop training at any time and provide the current optimal operating mode; after appropriate training, even if the optimal solution is not converged, the total system cost of the method of this invention or the traditional method is close to the minimum level, proving that the output operating mode is completely effective and feasible. These characteristics greatly avoid the solvability problem encountered by traditional optimization models in medium- and long-term production simulations, demonstrating the computational advantages of the method proposed in this paper.

[0192] In summary, this invention provides a medium- to long-term stochastic production simulation method, system, medium, and equipment. Verification using the ROTS system and real renewable energy data from a southern province demonstrates the following advantages: First, the new method reduces reliance on scenario selection and heuristic principles, avoiding interference from subjective factors. It can directly search for the optimal operating mode based on the distribution of stochastic factors, proving that it can reduce the total system cost while ensuring power supply security in extreme scenarios and renewable energy consumption, showing significantly better overall performance than traditional methods. Second, the new method exhibits good computational performance. Results show that it converges to the optimal solution with a lower number of iterations, and the results can be derived at any time with a low total system cost before convergence, avoiding the solvability problem of traditional optimization models. Furthermore, the new method has good scalability. It can fully consider numerous stochastic factors and various power generation resources, and can freely adjust modules such as stochastic factor distribution prediction based on actual needs and the development of related technologies.

[0193] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0194] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0195] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0196] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0197] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0198] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.< / x>

Claims

1. A medium- to long-term stochastic production simulation method, characterized in that, Includes the following steps: S1. Obtain basic technical data of power systems containing renewable energy and predicted distribution data of renewable energy; S2. Based on the basic technical data and predicted distribution data of the power system obtained in step S1, and with the goals of achieving system power supply security, renewable energy consumption, and reducing the total system cost, a Markov decision process is constructed under complex constraints, specifically as follows: Under the premise of adhering to the unit operating constraints, the dispatch center uses operating condition variables, including the predicted results of new energy output, water inflow, and system load, to determine the optimal conditions for operation. Simultaneously, it incorporates system state variables including unit maintenance status and reservoir water status. Provide operational mode decisions ;based on and The result is obtained through the PS calculation process. The system state for the new time period is derived. The system balance index and power generation cost are used to quantitatively analyze the cost of this round of decision-making; the above calculations are repeated until the system state changes from... Transfer to , and These represent the initial and final states of the system. The ADP search is used to find the decision sequence that minimizes the total cost of the power system, thus forming the operating mode. The dispatch center operates within a specific time period. Based on state variables and operating condition variables Make a decision Then, power system balance calculations are performed using probabilistic sequence arithmetic rules to derive the time periods. System status The state variables of the Markov decision process include the maintenance status of thermal power units and hydropower units, and the water and electricity distribution of the reservoir; the decision variables include the maintenance decisions of thermal power units and hydropower units, and the water and electricity storage decisions of the reservoir. S3. The Markov decision process obtained in step S2 is iteratively trained using an approximate dynamic programming method to search for the optimal set of decisions as the medium- to long-term operating mode. The scheduling center first determines an initial solution in the MDP framework and obtains the corresponding initial state sequence through simple methods or historical experience. Then, it changes a small increment above and below the state sequence to form a corridor. Then, it searches within the corridor using conventional dynamic programming and uses the result of this optimization as the initial solution for the next search. The process is repeated until the convergence condition is met. ADP based on a table model can solve the original problem in polynomial time complexity. By recording the solutions to subproblems, it avoids repeated calculations of subproblems. If the dispatch center is in a state variable state in the power system... Make decisions in a timely manner After actual execution, the new state of the system in the next time period and the related discrete sequence are uniquely determined; The optimal operating mode of the system is as follows: in, and They are the first Second and third The state obtained in the next iteration The corresponding value function; It is a constant whose value ranges from 0 to 1.

2. The medium- to long-term stochastic production simulation method according to claim 1, characterized in that, Basic technical data for the power system includes unit parameters, and the predicted distribution of renewable energy and load.

3. The medium- to long-term stochastic production simulation method according to claim 1, characterized in that, The power system balance calculation is specifically as follows: By sorting and renumbering the generating units according to their power generation economics, an ordered matching set can be obtained. If the economic order is The units during the period The amount of electricity that can be generated is ,and and The distribution of local residual load and the distribution of power consumption after the unit participates in balancing are respectively introduced, and convolution calculation is performed. and interleaving operations , obtain time period The internal order is Unit load balancing calculation .

4. The medium- to long-term stochastic production simulation method according to claim 1, characterized in that, For new energy generating units, directly use the operating condition variables. Inside and Transform into a discrete sequence. It is a wind turbine. During the period Predicted output distribution; It is a photovoltaic unit During the period The predicted power output distribution; for thermal power resources, the impact of random unit failures and maintenance decisions on available capacity is fully considered to obtain the sequence. Regarding hydropower resources, considering the time period State variables Operating condition variables Decision variables Determine the time period of the reservoir Water storage power sequence If the reservoir Corresponding to There are [number] hydropower units, which are simultaneously ranked [number] within the entire system. The hydroelectric units here The economic order among the individual units is as follows: The hydropower unit was obtained during the time period Discrete sequence of power generation The hydropower unit participates in system balancing, and the discrete sequence of the unit's power generation is calculated. Then solve for the reservoir The total generating capacity of other units not involved in the balancing process, and the dispatch center's guarantee period If the total capacity of units not undergoing maintenance exceeds the maximum load power value for that period and has a preset margin, the economic priority is as follows: The units during the period Electricity generation cost Calculations are performed, and then the results of all units are summed to obtain the system's performance during the time period. The cost of generating electricity.

5. The medium- to long-term stochastic production simulation method according to claim 1, characterized in that, The evaluation indicators for system balance include time periods. The expected energy deficit (EUE) and the load factor (LOLR) are as follows: in, and These are discrete sequences The state sequence number and the number of states; and It is a discrete sequence The The state outcome and probability of occurrence for each state; The function evaluates the input value; if it is positive, it returns 0; otherwise, it returns the input value. and These are the scene number and the total number of scenes; and The first Power loss and total load power in each scenario; State variables : in, and It is a thermal power unit and hydroelectric generator units During the period The remaining number of downtime days, It is a reservoir During the period Distribution of water storage and electricity; It is the corresponding time period number; Operating condition variables express: in, It is a reservoir During the period The predicted distribution of water and electricity flow; Is the load during the time period The predicted distribution; It is the time period number; Decision variables : in, and It is a thermal power unit and hydroelectric generator units During the period The maintenance decision result is 0 if maintenance is decided and 1 otherwise. and It is the unit's original serial number; It is a reservoir At any moment The decision-making outcome regarding the amount of water stored; State variables The optimal total cost function of the system under the following conditions : in, Let it be the expected function; A return factor ranging from 0 to 1; State variables Decision execution And transferred to Instantaneous costs generated : in, , , and These are the power generation cost function, the hydropower curtailment cost function, the renewable energy curtailment cost function, and the load shedding cost function. , and That is the corresponding cost coefficient.

6. A medium- to long-term stochastic production simulation system, characterized in that, include: The data module acquires basic technical data of power systems containing renewable energy and predicted distribution data of renewable energy. The construction module, based on fundamental technical data and predicted distribution data of the power system, aims to achieve system power supply security, renewable energy consumption, and reduce total system cost. It constructs a Markov decision process under complex constraints, specifically as follows: Under the premise of adhering to the unit operating constraints, the dispatch center uses operating condition variables, including the predicted results of new energy output, water inflow, and system load, to determine the optimal conditions for operation. Simultaneously, it incorporates system state variables including unit maintenance status and reservoir water status. Provide operational mode decisions ;based on and The result is obtained through the PS calculation process. The system state for the new time period is derived. The system balance index and power generation cost are used to quantitatively analyze the cost of this round of decision-making; the above calculations are repeated until the system state changes from... Transfer to , and These represent the initial and final states of the system. The ADP search is used to find the decision sequence that minimizes the total cost of the power system, thus forming the operating mode. The dispatch center operates within a specific time period. Based on state variables and operating condition variables Make a decision Then, power system balance calculations are performed using probabilistic sequence arithmetic rules to derive the time periods. System status The state variables of the Markov decision process include the maintenance status of thermal power units and hydropower units, and the water and electricity distribution of the reservoir; the decision variables include the maintenance decisions of thermal power units and hydropower units, and the water and electricity storage decisions of the reservoir. The simulation module uses an approximate dynamic programming method to iteratively train the Markov decision process and search for the optimal set of decisions as the medium- to long-term operating mode. The scheduling center first determines an initial solution in the MDP framework and obtains the corresponding initial state sequence through simple methods or historical experience. Then, it changes a small increment above and below the state sequence to form a corridor. Then, it searches within the corridor using conventional dynamic programming and uses the result of this optimization as the initial solution for the next search. The process is repeated until the convergence condition is met. ADP based on a table model can solve the original problem in polynomial time complexity. By recording the solutions to subproblems, it avoids repeated calculations of subproblems. If the dispatch center is in a state variable state in the power system... Make decisions in a timely manner After actual execution, the new state of the system in the next time period and the related discrete sequence are uniquely determined; The optimal operating mode of the system is as follows: in, and They are the first Second and third The state obtained in the next iteration The corresponding value function; It is a constant whose value ranges from 0 to 1.

7. A computer-readable storage medium for storing one or more programs, characterized in that, The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform the method of any one of claims 1 to 5.

8. A computing device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including steps for performing the method of any one of claims 1 to 5.

Images