A new power system planning method based on the integration of knowledge rules and evolutionary algorithms

Abstract

The present invention proposes a new power system planning method based on the fusion of knowledge rules and evolutionary algorithms, including: establishing a new power system planning model; specifically: determining the planning boundary parameters of the new power system, constructing the objective function of the power system planning model, and constructing the planning constraints and operation constraints of the power system planning model; utilizing the relative independence of the planning layer and the operation layer to transform the new power system planning model into a planning and operation double-layer sub-model to reduce the complexity of a single new power system planning model; utilizing a genetic algorithm guided by knowledge rules to solve the planning and operation double-layer sub-model to output the optimal planning scheme for the new power system. The present invention ensures that the planning model is closely integrated with the system operation requirements to avoid separation, and at the same time combines the large-scale optimization capability of the genetic algorithm with the rapid search capability guided by knowledge rules in the model solution, achieving a balance between comprehensiveness and rapidity in optimization and supporting large-scale new power system planning.

Description

Technical Field

[0001] The present invention relates to the technical field of novel power system planning, and in particular to a novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms. Background Art

[0002] Renewable energy sources such as wind power and photovoltaics have seen large-scale development in recent years. However, the volatility and randomness of wind and solar power output have led to complex and variable operating modes for new power systems. Traditional new power system planning methods use typical methods to evaluate operational indicators such as power generation costs and safety and stability levels, and the lack of coordination between planning and operation has become increasingly prominent. Therefore, when planning systems such as power optimization layout, grid upgrades, and energy storage configuration, comprehensively evaluating the operational performance of planning schemes based on multiple operating modes to ensure that the planning schemes meet the requirements of complex and variable operating modes has become the development direction of new power system planning.

[0003] Time series operation simulation can finely model the operation process of the power system and accurately quantify operational indicators such as unit operation, grid flow and power curtailment. Embedding it into the new power system planning model can consider the operation scenarios in detail during the planning stage, thereby improving the applicability of the planning results. However, the integration of planning models and operation models will significantly increase the number of decision variables and constraints, and the correlation between planning decision variables and operation decision variables will also introduce nonlinear factors, resulting in a significant increase in the complexity of the planning model. In particular, when planning the source-grid-storage collaborative planning of the new power system, the joint optimization of the site selection and sizing of power sources and energy storage and grid upgrades may result in the number of variables to be optimized reaching thousands of dimensions, becoming a large-scale nonlinear optimization problem. The solution methods for power system planning models include mathematical optimization algorithms based on optimization theory and evolutionary algorithms. Mathematical optimization methods usually require simplifying actual problems and converting them into solvable linear forms. This simplification may cause the solution to deviate from the actual optimal value. Evolutionary algorithms, inspired by the laws of evolution in nature and swarm intelligence phenomena, are suitable for solving complex problems such as nonlinear, non-convex, discrete or mixed variables, and are not restricted by the differentiability of the model. However, when directly used for source-grid-storage coordinated planning problems, the solution time increases due to the increase in variable dimensions. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a new power system planning method based on the fusion of knowledge rules and evolutionary algorithms. Through the improvements in two aspects, namely, embedded operation simulation in the planning model and rule-guided genetic algorithm, it can ensure that the planning model is closely integrated with the system operation requirements to avoid separation. At the same time, in the model solution, the large-scale optimization capability of the genetic algorithm and the rapid search capability guided by knowledge rules are combined to achieve both comprehensiveness and rapidity of optimization and support large-scale new power system planning.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A new power system planning method based on the fusion of knowledge rules and evolutionary algorithms includes the following steps:

[0007] Establish a new power system planning model; specifically, first determine the planning boundary parameters of the new power system, then construct the objective function of the power system planning model with the goal of optimizing the comprehensive combination of new equipment investment and system operating costs, and finally establish planning and operating constraints of the power system planning model;

[0008] Taking advantage of the relative independence of the planning layer and the operation layer, the new power system planning model is transformed into a planning and operation two-layer sub-model to reduce the complexity of a single new power system planning model;

[0009] A genetic algorithm guided by knowledge rules is used to solve the planning and operation double-layer sub-model to output the optimal planning scheme for the new power system; specifically: the solution parameters of the genetic algorithm and the physical parameters of the planning object are set; the search space of power supply equipment, energy storage equipment and transmission lines is narrowed in turn; the optimization variable of the power supply equipment is used as the main variable, and the energy storage equipment and transmission lines are used as dependent variables, and the initial value of each optimization variable is generated in turn according to the dependency relationship between the decision variables; various variables are combined to form an initial population, and the population is verified to meet the overall installation goal; the operation performance indicators of each planning scheme in the population are evaluated based on the time series operation simulation, and the fitness of the planning scheme is calculated by combining the operation indicators and investment indicators; under the guidance of knowledge rules, the crossover and mutation of the planning scheme promote the evolution of the planning scheme population, and the convergence is checked according to the convergence conditions to end the evolution or enter the next generation evolution process, and the individual with the best fitness is selected from the current generation of schemes as the optimization result, and the optimal planning scheme for the new power system is output.

[0010] The effects provided in the summary of the invention are only the effects of the embodiments, not all the effects of the invention. One of the above technical solutions has the following advantages or beneficial effects:

[0011] The present invention proposes a new power system planning method based on the fusion of knowledge rules and evolutionary algorithms, comprising the following steps: establishing a new power system planning model; specifically: first determining the planning boundary parameters of the new power system, then constructing the objective function of the power system planning model with the comprehensive optimization of new equipment investment and system operation costs as the goal, and finally constructing planning constraints and operation constraints of the power system planning model; utilizing the relatively independent characteristics of the planning layer and the operation layer to transform the new power system planning model into a planning and operation double-layer sub-model to reduce the complexity of a single new power system planning model; utilizing a genetic algorithm guided by knowledge rules to solve the planning and operation double-layer sub-model and output the optimal planning scheme for the new power system; specifically: setting the solution parameters and The physical parameters of the planning object; shrink the search space of power supply equipment, energy storage equipment and transmission lines in turn; take the power supply equipment optimization variable as the main variable, energy storage equipment and transmission lines as the dependent variables, and generate the initial value of each optimization variable in turn according to the dependency relationship between decision variables; combine various variables to form an initial population, and verify that the population meets the overall installation goal; evaluate the operating performance indicators of each planning scheme in the population based on the time series operation simulation, and calculate the fitness of the planning scheme by combining the operating indicators and investment indicators; under the guidance of knowledge rules, the crossover and mutation of planning schemes drive the evolution of the planning scheme population, and check the convergence according to the convergence conditions to end the evolution or enter the next generation of evolution process, select the individual with the best fitness from the current generation of schemes as the optimization result, and output the optimal planning scheme for the new power system. The present invention first constructs a source-grid-storage collaborative planning model with an embedded refined operation process to adapt to the volatility and intermittent characteristics of wind and solar energy. Then, based on the genetic algorithm with elite retention, it designs and improves the solution process of the evolutionary algorithm. In the process of initialization and cross-mutation to generate offspring solutions, it incorporates knowledge rules based on engineering experience, and guides the optimization direction of the evolutionary process from three aspects: decision variable reduction, feasible set contraction, and evolutionary rule improvement. The present invention proposes a new power system planning method based on the fusion of knowledge rules and evolutionary algorithms. Through the two improvements of embedded operation simulation and rule-guided genetic algorithm in the planning model, it can ensure that the planning model is closely integrated with the system operation requirements to avoid separation. At the same time, in the model solution, it combines the large-scale optimization capability of the genetic algorithm with the rapid search capability guided by knowledge rules to achieve both comprehensiveness and rapidity in optimization and support large-scale new power system planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a flow chart of a new power system planning method based on the fusion of knowledge rules and evolutionary algorithms proposed in Example 1 of the present invention;

[0013] Figure 2 This is a flowchart of constructing a new power system planning model proposed in Example 1 of the present invention;

[0014] Figure 3 Flowchart for dividing the source-network-storage planning and operation planning sub-problems proposed in Example 1 of the present invention;

[0015] Figure 4 This is a schematic diagram of the division results of the source-network-storage planning and operation planning sub-problems proposed in Example 1 of the present invention;

[0016] Figure 5 This is a flowchart for solving the source-network-storage collaborative planning model proposed in Example 1 of the present invention;

[0017] Figure 6 This is a process diagram for generating dependent variables based on main variables proposed in Example 1 of the present invention;

[0018] Figure 7 A diagram of the process of cross-generation of offspring by parent generation under the guidance of the knowledge rule proposed in Example 1 of the present invention;

[0019] Figure 8 This is the process of generating offspring by mutation of the parent generation under the guidance of the knowledge rule proposed in Example 1 of the present invention;

[0020] Figure 9 This is a topological diagram of the starting point system of a regional power grid;

[0021] Figure 10 This is a schematic diagram of a typical daily load curve for a certain area;

[0022] Figure 11 is the typical wind power maximum output curve;

[0023] Figure 12 is the photovoltaic maximum output curve;

[0024] Figure 13 This is the relationship diagram between the optimal comprehensive cost and the number of scanning solutions. DETAILED DESCRIPTION

[0025] In order to clearly illustrate the technical features of this solution, the present invention is described in detail below through specific implementation methods and in conjunction with the accompanying drawings. The disclosure below provides many different embodiments or examples for realizing different structures of the present invention. In order to simplify the disclosure of the present invention, the components and settings of specific examples are described below. In addition, the present invention may repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplicity and clarity and does not itself indicate the relationship between the various embodiments and / or settings discussed. It should be noted that the components illustrated in the accompanying drawings are not necessarily drawn to scale. The present invention omits descriptions of well-known components and processing technologies and processes to avoid unnecessary limitations on the present invention.

[0026] Example 1

[0027] Embodiment 1 of the present invention proposes a novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms to address the technical issues existing in conventional planning methods. Specifically, it includes three steps: constructing a source-grid-storage collaborative planning model with embedded time-series operation simulation; decomposing an optimization model based on the partitioning of planning and operation subproblems; and solving the model based on the fusion of knowledge rules and evolutionary algorithms.

[0028] Figure 1 This is a flow chart of a new power system planning method based on the fusion of knowledge rules and evolutionary algorithms proposed in Example 1 of the present invention;

[0029] In step 1, a source-grid-storage collaborative planning model with an embedded time-series operation simulation is established, that is, a new power system planning model is established. Specifically, the planning boundary parameters of the new power system are first determined. Then, the objective function of the power system planning model is constructed with the goal of achieving the optimal combination of new equipment investment and system operating costs. Finally, the planning and operation constraints of the power system planning model are constructed.

[0030] Figure 2 This is a flowchart of constructing a new power system planning model proposed in Example 1 of the present invention;

[0031] Step 1.1: Determine the planning boundary parameters of the new power system based on research or data extrapolation. Specifically, this includes: researching reports published by social and economic research institutions, collecting regional economic and population forecast data published in the planning target year, and forming the maximum load of each node using formula (1):

[0032] P s L =a0+a1×E s +a2×N s ; (1)

[0033] Among them, P s L E is the forecast result of the maximum power load in the planning year; s is the economic total forecast data corresponding to the maximum load; N s is the population forecast data corresponding to the maximum load; a1 is the correlation coefficient of the economic aggregate forecast data; a1 is the correlation coefficient of the population forecast data; a0 is a constant term; a1, a1 and a0 are obtained by fitting the historical data of the maximum power load, economy and population in previous years.

[0034] Based on the typical daily load curve of power load and the annual curve of daily maximum load released by the power department, the hourly 8760h load curve of last year was formed.

[0035] Based on the wind and solar resource distribution map released by the Ministry of Land and Resources, the average density of wind and solar resources in each region is calculated, and the boundary of regional exploitable resources is determined by formula (2):

[0036]

[0037] in, is the wind and solar resource density in the area where node i is located; S i is the area of ​​the region where node i is located; is the upper limit of resource development in the region where node i is located; η wf,pv The maximum conversion efficiency of wind and solar power related to the power generation technology level can be selected from the data published by the wind and solar generator manufacturer.

[0038] The wind and solar equipment models and prices released by mainstream manufacturers were investigated to determine the unit construction cost benchmark of wind and solar power. The benchmark was further modified using formula (3) based on the ratio of the installed capacity of wind and solar power in the region to the upper limit of development:

[0039]

[0040] in, The unit construction cost of wind and solar power equipment at each node in the planning year; The unit cost benchmark value of wind and solar power equipment at each node depends on the technical level of the equipment manufacturer; is the installed wind and solar power capacity at each node; F(·) is a non-decreasing function; it reflects the rule that the development cost of wind and solar power increases with the degree of development, and can typically adopt a linear, exponential or hyperbolic function.

[0041] Determine the construction cost of transmission lines with different voltage levels and cross-sectional areas and its corresponding rated transmission capacity Investigate mainstream transmission line models and generate construction costs for transmission lines with different cross-sectional areas at voltage levels such as 500kV and 1000kV and its corresponding rated transmission capacity

[0042] Step 1.2: Construct the objective function of the power system planning model with the goal of optimizing the comprehensive combination of new equipment investment and system operating costs.

[0043] Using the unit installed cost and discount rate of each type of equipment as input, the annualized investment cost of all equipment in the entire system is constructed;

[0044]

[0045] Among them, r0 is the discount rate during the planning period, which is generally the bank loan interest rate for the same period; H type is the service life of type type equipment, ntype A set of locations for type type equipment. is the unit cost of type type equipment at node i, is the newly installed capacity of type type equipment at node i;

[0046] Taking into account the fuel costs, carbon emission fees, standby costs of thermal power units, as well as the load shedding and power curtailment penalties of the entire power system, an expression for the annual operating cost of the entire system is constructed;

[0047] The annual operating cost of thermal power units is:

[0048]

[0049] Among them, F op is the annual operating cost of the thermal power unit, T is the operating cycle of the system, λ coal 、 is the unit coal price and carbon emission cost; P i coal (t) is the actual active power output of the unit at node i at time t, p i The quadratic coefficient of actual active output; q i is the first-order coefficient of actual active power output; l i is the constant term of actual active power output; are the unit positive and negative standby costs, are the positive and negative backups provided by the thermal power unit at node i at time t, are the single start-up and shutdown costs respectively; are the start-up and shutdown indicator variables of the unit, 1 means the unit is on. A value of 1 indicates shutdown.

[0050] Determine load shedding penalties and power curtailment penalties;

[0051]

[0052] Among them, F dc is the load shedding penalty for the system, is the penalty cost per unit load shedding power, is the load removed at node i at time t, F drop Fines for abandoned power from wind and solar generators, is the penalty cost of the unit power curtailment, is the power curtailment of the unit at node i at time t;

[0053] The annual operating cost of the thermal power unit F op , System load shedding penalty F dc and power curtailment penalty Fdrop Add together to construct the expression for the annual operating cost of the entire system:

[0054] F ot =F op +F dc +F drop ; (7)

[0055] Among them, F ot It is the annual operating fee of the entire system;

[0056] The annualized cost of all equipment investment in the system is I inv and the annual operating cost of the entire system F ot Adding together, the annual comprehensive cost of constructing the planning scheme is: min(I inv +F ot );(8).

[0057] Step 1.3: Constructing the planning constraints of the power system planning model. The process includes:

[0058] Calculate the sum of the installed capacity of the entire system to form the total installed capacity constraint of the source and storage; specifically:

[0059]

[0060] The target construction total capacity for type type equipment, ξ type The maximum allowable error in the installed capacity of type type equipment;

[0061] Based on the total installed capacity constraints of source and storage, the installed capacity of various types of wind and solar power sources at each node is constructed to meet the upper limit constraints of local resource development; specifically:

[0062]

[0063] in, The upper limit of the developable capacity of each type of equipment located at node i.

[0064] Step 1.4: Construct operational constraints for the power system planning model.

[0065] Taking into account the power balance between source and load, backup requirements, line transmission capacity, and equipment operating characteristics, the operating constraints of the source-grid-storage collaborative planning model are constructed.

[0066] Calculate the sum of total power generation and total load power to form a source-load power balance constraint where total power generation and total load are equal;

[0067]

[0068] Among them, P icoal (t) is the actual active power output of the unit at node i at time t; P i RE (t) is the active power output of the wind and solar generator at node i at time t; is the power curtailment of the unit at node i at time t; is the charging power of the energy storage device at node i at time t; is the discharge power of the energy storage device at node i at time t; P i load (t) is the active power of the load at node i at time t; is the load removed at node i at time t;

[0069] Taking the load of each node and the output of wind and solar power units, as well as the load shedding power and power abandonment power as input, the reserve demand constraint of thermal power units is constructed; specifically:

[0070]

[0071] in, L is the upper and lower reserve power set for the thermal power unit at time t; + % is the first reserve requirement coefficient of the load; L - % is the second reserve requirement coefficient of the load; W + % is the first reserve requirement coefficient reserved for wind and solar power unit output; W - % is the second reserve requirement coefficient reserved for wind and solar power unit output;

[0072] The DC power flow is used to linearize the relationship between the line power flow and the node injection power flow to form the line power flow constraint, which is specifically:

[0073]

[0074] Among them, x ij is the reactance between node i and node j; M is the branch correlation matrix; X is the node impedance matrix; P i,in (t) is the net injected power at node i at time t; P j,in (t) is the net injected power at node j at time t; is the upper limit of the reverse transmission capacity of the line between node i and node j; is the upper limit of the forward transmission capacity of the line between node i and node j;

[0075] Taking into account the active power output constraints of thermal power units, the active power output constraints considering the standby output, the ramping constraints, and the minimum start-stop time constraints, the active power output constraints of thermal power units are constructed as follows:

[0076]

[0077] is the start / stop status of the i-th node unit at time t, is the maximum output ratio of the i-th node unit; is the minimum output ratio of the i-th node unit; is the newly installed capacity of thermal power units at node i;

[0078] Taking the active power output of the unit and its upper limit as input, the active power output constraint of the thermal power unit considering the reserve output is constructed:

[0079]

[0080] The lower reserve power set for the thermal power unit at time t; The upper reserve power set for the thermal power unit at time t;

[0081] The ramping constraint of the thermal power unit is constructed by taking the active power output of the unit as input:

[0082]

[0083] is the maximum climbing rate of the unit; is the maximum descending and climbing rate of the unit; P i coal (t+1) is the actual active power output of the unit at node i at time t+1;

[0084] Taking the start and stop indicator variables of the unit as input, the minimum start and stop time constraint of the thermal power unit is constructed:

[0085]

[0086] Where, T ON is the minimum startup time of the unit; T OFF The minimum stop time of the unit;

[0087] After normalizing the wind-solar output curve, the wind-solar output curve and wind-solar installed capacity are used as input to construct the active output constraint of the wind-solar generator set:

[0088]

[0089] Among them, P i wf (t) is the actual active power output of the wind turbine at node i at time t; is the maximum value of the normalized wind power curve at time t, that is, the maximum power generation capacity percentage of wind power in the maximum power point tracking mode; is the newly installed capacity of wind turbines at node i; P ipv (t) is the actual active power output of the PV unit at node i at time t; is the maximum value of the normalized photovoltaic power generation curve at time t, that is, the percentage of the maximum power generation capacity of the photovoltaic power source in the maximum power point tracking mode; is the newly installed capacity of the PV unit at node i;

[0090] The operating characteristic constraints of energy storage equipment are constructed by integrating the charging and discharging power constraints, state of charge constraints, daily clearance constraints, and energy balance constraints of energy storage equipment;

[0091] Among them, the installed capacity and charging and discharging power of the energy storage device are used as input, and the charging and discharging power constraint of the energy storage device is constructed as follows:

[0092]

[0093] in, Indicates that the energy storage device is in operation at time t;

[0094] The charge and discharge power constraints of the energy storage device are used to construct the state of charge constraints of the energy storage device:

[0095]

[0096] in, is the state of charge of the energy storage device at time t; is the maximum storage ratio of the energy storage device; is the minimum storage ratio of the energy storage device;

[0097] Construct the daily clearance constraints of the energy storage equipment based on the state of charge constraints of the energy storage equipment:

[0098] SOC i ess (t=0h)=SOC i ess (t=24h); (22)

[0099] Among them, SOC i ess (t=0h) is the state of charge of the energy storage at 0h every day; SOC i ess (t = 24h) is the state of charge of the energy storage at 24h per day;

[0100] Taking the state of charge and charge and discharge power of the energy storage device as input, the energy balance constraint expression of the energy storage device is constructed:

[0101]

[0102] in, is the state of charge of the energy storage device at time t-1; P i ess is the charging and discharging power of the energy storage device at node i at time t; is the rated capacity of the energy storage at node i; ΔT is the time step of energy storage charging and discharging.

[0103] In step 2, the new power system planning model is transformed into a planning and operation two-layer sub-model by taking advantage of the relative independence of the planning layer and the operation layer to reduce the complexity of a single new power system planning model. Figure 3 This is a flowchart for dividing the source-network-storage planning and operation planning sub-problems proposed in Example 1 of the present invention.

[0104] Analyze the relationships among optimization variables in the source-grid-storage collaborative planning model and determine the handover variables based on the closeness between the variables. Specifically, source-storage installed capacity and line upgrade capacity are selected as handover variables transferred from the planning layer to the operation layer. Economic indicators such as fuel cost, load shedding penalties, and total system operating cost, as well as operational indicators such as load shedding rate, curtailment rate, and renewable energy power ratio, are selected as handover variables transferred from the operation layer to the planning layer.

[0105] Construct a planning-level optimization subproblem. Select the installed capacity of each type of power source and load at each node and the upgrade capacity of each line as the decision variables of the planning level. Use the total installed capacity and resource upper limit constraints as constraints, and minimize the combined investment and operating costs as the optimization goal. This forms a nonlinear continuous optimization subproblem of the planning level that depends on the external input of operating costs:

[0106] The nonlinear continuous optimization subproblem at the planning level is described as:

[0107]

[0108] On this basis, the planning layer generates a construction plan for the source, network and storage installed capacity and passes it to the operation layer.

[0109] Construct the runtime optimization subproblem: Figure 4 This is a schematic diagram of the division results of the source-grid-storage planning and operation planning sub-problems proposed in Example 1 of the present invention; the start and stop status of the thermal power units at each node and the generated power of various power sources, the energy storage charging and discharging power, and the load shedding power are selected as operation-layer decision variables, with the power balance of the entire system, the backup demand, and the operating boundaries of each equipment as constraints, and the minimization of the operating cost taking into account fuel, carbon emissions, and load shedding penalties as the optimization goal, forming an operation-layer linear mixed integer optimization sub-problem that depends on the source and storage installed capacity of each node and the line reinforcement capacity as external inputs.

[0110] The run-level linear mixed integer optimization subproblem is described as:

[0111]

[0112] On this basis, the operation layer performs production simulation calculations according to the given construction plan to solve the operation sub-problem, and thereby feeds back the operation indicators of its construction plan to the planning layer.

[0113] In step 3, a genetic algorithm guided by knowledge rules is used to solve the planning and operation double-layer sub-model to output the optimal planning scheme for the new power system. Figure 5 This is a flowchart for solving the source-network-storage collaborative planning model proposed in Example 1 of the present invention.

[0114] Specifically, the solution parameters of the genetic algorithm and the physical parameters of the planning object are set; the search space of power supply equipment, energy storage equipment and transmission lines is narrowed in turn; the optimization variable of the power supply equipment is the main variable, and the energy storage equipment and transmission line are the dependent variables, and the initial value of each optimization variable is generated in turn according to the dependency relationship between the decision variables; various variables are combined to form an initial population, and the population is verified to meet the overall installation goal; the operating performance indicators of each planning scheme in the population are evaluated based on the time series operation simulation, and the fitness of the planning scheme is calculated by combining the operating indicators and investment indicators; under the guidance of knowledge rules, the crossover and mutation of planning schemes drive the evolution of the planning scheme population, and the convergence is checked according to the convergence conditions to end the evolution or enter the next generation of evolution process, and the individual with the best fitness is selected from the current generation of schemes as the optimization result, and the optimal planning scheme for the new power system is output.

[0115] Step 3.1: Set the solution parameters of the evolutionary algorithm, mainly including the population size N Chrome , maximum number of iterations N iterMax , crossover probability P cross , mutation probability P mut Etc., generally determined based on historical experience or adjusted based on actual experimental needs.

[0116] Set the physical parameters of the planning object, including the topological structure information of each node and the grid architecture, the total installed capacity target of each generator set and energy storage equipment, the upper limit of the developable capacity of the new energy power source at each node and the typical maximum output curve, the maximum load of each node in the planning year and the typical daily load curve.

[0117] Step 3.2: Generate the initial population based on the planning layer model of search space contraction. This mainly reduces the search space of the planning subproblem according to the knowledge rule and uses random numbers to generate a set number of initial planning solutions. Specifically, it includes:

[0118] (1) According to the order of power supply, energy storage, and circuit, the search space of the optimization scheme is narrowed based on the key characteristic indicators of each type of equipment.

[0119] The process of shrinking the search space for power devices includes:

[0120] Sort by the unit development cost of each power source at each node in ascending order. The nodes of the bits form a power supply addressable space; The number of nodes planned for the type of power supply is: The upper limit of the installed capacity of the selected node corresponding to the type of power supply is calculated based on the remaining developable capacity of the selected node, the planned installation target, and the planned target magnification factor:

[0121]

[0122] in,

[0123] in, is the upper limit of the installable capacity of the type power supply of the i-th node; is the upper limit of the remaining development capacity of the type power source at the i-th node; is the installed capacity of type power supply at node i; type The target capacity magnification factor of the entire system for type type power supply; Plan the total installed capacity target for type power supply in the year;

[0124] The process of shrinking the search space for energy storage devices includes:

[0125] Calculate the difference in installed energy storage capacity based on the installed capacity of various types of wind and solar power sources and energy storage at each node; sort the nodes in descending order based on the difference in installed energy storage capacity as a characteristic indicator, and select the node before sorting. The nodes of the bit form the energy storage addressable space; among them, The number of nodes planned for energy storage equipment investment and construction; the upper limit of the energy storage capacity that can be installed at each node is calculated based on the energy storage installed capacity difference of the selected nodes, the planned installed capacity target, and the planned target capacity magnification factor, specifically:

[0126]

[0127] in,

[0128] in, is the upper limit of the installable capacity of energy storage at the i-th node; is the difference in installed capacity of energy storage at the i-th node; is the installed capacity of energy storage for node i; is the target capacity of the entire energy storage system; ess is the target capacity magnification factor; is the installed capacity of the wind turbine at node i; η wf The wind power configuration ratio set for the entire system; is the installed capacity of the PV system at node i; η pvThe photovoltaic power configuration ratio set for the entire system;

[0129] The search space process of shrinking the transmission line includes: calculating the comprehensive load rate according to the load rate of each line at each time using the following formula:

[0130]

[0131] Among them, Quartile(·) is the upper quartile function; is the load rate of the jth line at time t; is the comprehensive load rate of the line;

[0132] Sort the lines in descending order based on the comprehensive load rate as the characteristic index, and select The lines of bits form a line set to be upgraded, where Plan the number of upgrades for line equipment; determine the maximum number of lines of the same model that can be connected in parallel based on the line's comprehensive load rate, and form a line upgrade search space, specifically:

[0133]

[0134] in, is the maximum upgrade capacity of the k-th line; is the original rated capacity of the kth line; It is the first level threshold of line load factor; is the line load rate classification threshold of the kth line; is the nth level threshold of line load rate; The total number of levels required for line upgrades; The maximum number of lines of the same model that can be connected in parallel for line upgrade; line (·) is the function of the number of line upgrades based on the load factor input.

[0135] Taking the power supply optimization variables as the main variables and the energy storage and line optimization variables as the dependent variables, the initial values ​​of each optimization variable are generated in sequence according to the dependency relationship between the decision variables. Figure 6 This is a process diagram for generating dependent variables based on main variables proposed in Example 1 of the present invention.

[0136] a. Use the random variable method to generate the initial value of power supply installed capacity:

[0137] According to the number of installable locations of various power search spaces generated above and the upper limit of installation capacity at each point The Monte Carlo random number generator function rand(·) generates the number Range uniformly distributed random numbers to form a single planning scheme power installation initial plan; repeat the above process NChrome times, forming N Chrome Group power supply installation plan.

[0138] b. Use the following rule 1 to N Chrome The power supply installation plan determines the energy storage plan one by one

[0139] The initial installation plan is input into the operation sub-problem to calculate the system operation status, and the annual power curtailment rate of each node is calculated based on the following formula:

[0140]

[0141] Where: are the annual wind and solar curtailment rates at the i-th node; are the wind power and photovoltaic power generation curtailment at the i-th node at time t; are the wind and solar power generation at the i-th node at time t respectively.

[0142] The annual wind and solar power curtailment rates are sorted in descending order, and the energy storage installation demand is divided into grades based on the curtailment rate. The energy storage node installation coefficients are assigned according to the graded levels, and the total energy storage installation target is decomposed by weighted average of the node coefficients.

[0143] c. Use the following rule 2 to N Chrome Group power installation plan to determine line upgrade plan one by one

[0144] Based on the above known system operation status, the comprehensive load rate of the line is calculated using formula (28). According to the line load rate, it is substituted into formula (29) to calculate the line upgrade demand classification, and the top is selected again from high to low. Level circuit, each circuit is connected in parallel Lines of the same model are upgraded to generate an initial line plan.

[0145] Various variables are combined to form an initial population, and the population is verified to meet the overall installation target. Specifically, the power supply, energy storage, and line upgrade plans obtained in the previous process are merged to form the initial plan set of the planning model, namely the initial population; the total installed capacity of various types of power supplies and energy storage is verified. If the total installed capacity requirements are not met, the capacity is increased or decreased at randomly selected locations until the total installed capacity target requirements of all types of power supplies and energy storage are met.

[0146] Step 3.3: Based on the time series operation simulation, evaluate the operation performance indicators of each planning scheme in the population, and calculate the fitness of the planning scheme by combining the operation indicators and investment indicators. Specifically:

[0147] Each planning scheme in the planning scheme set is passed to the lower layer. Based on the typical daily load curve, multiple 24-hour operation optimization sub-problems are constructed, or based on the annual 8760-hour load curve, a 24-hour operation optimization sub-problem is constructed in a rolling manner. By solving the operation optimization sub-problems, a refined time series production simulation is realized, and the operation status of the system at each moment is calculated. Based on the state of the power system at each moment, including the output of thermal power units, wind and solar power output, energy storage charging and discharging power, line transmission power, and actual node power, etc., the annual operation cost F is calculated by substituting them into equations (5) to (7). ot , annual wind curtailment rate, annual solar curtailment rate, comprehensive line load rate and other operating indicators;

[0148] For each planning scheme, the annualized investment cost I corresponding to the planning scheme is calculated based on formula (4): inv ;

[0149] Comprehensive operation economic indicators F ot and investment indicators I inv A comprehensive cost index F, also known as the fitness index, is formed to evaluate the pros and cons of planning schemes.

[0150] Step 3.4: Under the guidance of knowledge rules, the crossover and variation of planning schemes drive the evolution of the planning scheme population.

[0151] First, the importance and fitness of each planning scheme are ranked through non-dominated sorting and congestion calculation. Then, the existing scheme set is optimized using a tournament method to form a parent generation with elite retention characteristics. Further cross-mutation guided by knowledge rules is performed to generate a new generation of planning scheme sets. Through the generation of new schemes, the planning schemes are gradually updated and evolved towards a planning scheme set with higher fitness.

[0152] Based on the current and previous generation populations, individuals in the population are sorted according to the non-dominated sorting principle and fitness index to form a hierarchical sorting of dominated solutions and non-dominated solutions; the crowding degree of individuals in each population is calculated to evaluate their relative distribution density in the solution space;

[0153] The parent population is formed from the sorted individuals according to the tournament strategy; the parent population is crossed to generate the child population under the guidance of the knowledge rule; that is, according to the tournament selection strategy, two individuals are randomly selected from the above population for comparison, and the individuals with high non-dominated level and low crowding degree are selected to form the parent population; repeat N Chrome The tournament selected the scheme, including N Chrome The parent population of individuals;

[0154] (2) Perform N Chrome / 2 knowledge rule guided crossover and N Chrome / 2 times of knowledge rule-guided mutation, generating a maximum of N ChromeThe offspring population of each scheme. Specifically:

[0155] The process of crossover generation between parent and offspring under the guidance of knowledge rules

[0156] Call the random number generation function to generate a real number p randomly distributed between 0 and 1 cross , if p cross <P cross Then the subsequent crossover steps are executed to generate offspring individuals, otherwise the crossover ends without generating offspring;

[0157] Randomly select two parent individuals X1 and X2 from the parent population;

[0158] For the power supply variables in individuals, nodes with high power abandonment rates are selected as gene fragment intersections, exchange ratio coefficients are randomly generated, and the value of the gene fragment intersection at the power supply intersection is calculated according to the following formula:

[0159]

[0160] in, The first gene segment value of the node with high power abandonment rate of the power type variable to be exchanged is selected; The second gene segment value of the node with high power abandonment rate of the power type variable to be exchanged is selected; is the randomly generated 0-1 first exchange ratio coefficient; The values ​​of the chromosomes with larger and smaller power abandonment rates are taken after crossing respectively;

[0161] For the energy storage variables in individuals, nodes with high power abandonment rates and the corresponding low power abandonment rate nodes in reverse order are selected as gene fragment intersections, exchange ratio coefficients are randomly generated, and the gene exchange values ​​between energy storage intersections are calculated: the specific formula is:

[0162]

[0163] These are the values ​​of the corresponding gene fragments before chromosome crossover for energy storage variables with larger and smaller abandonment rates, is the randomly generated 0-1 second exchange ratio coefficient;

[0164] For the line-type variables in an individual, the line with high load rate and the line with low load rate corresponding to the reverse order are preferred as the gene segment intersection points, and the gene exchange value between the line-type intersection points is calculated according to the form of reference (32).

[0165] Combining the changes of three crossover points to form new offspring individuals Figure 7 This is a diagram of the process of generating offspring by crossover of the parent generation under the guidance of the knowledge rule proposed in Example 1 of the present invention.

[0166] The process of generating offspring from parent generation through mutation guided by knowledge rules

[0167] Call the random number generation function to generate a real number p randomly distributed between 0 and 1 mut , if p mut <P mut Then the subsequent mutation steps are executed to generate offspring individuals, otherwise the crossover ends without generating offspring;

[0168] Randomly select two parent individuals X1 from the parent population;

[0169] For the power supply variables in individuals, nodes with high power abandonment rates and nodes with low power abandonment rates corresponding to the reverse order are selected as gene fragment mutation points. Exchange ratio coefficients are randomly generated, and the gene exchange values ​​between power supply mutation points are calculated. Specifically,

[0170]

[0171] The first gene segment value of the low power abandonment rate node of the power type variable to be exchanged is selected. The second gene segment value of the low power abandonment rate node of the power type variable to be exchanged is selected; is the randomly generated 0-1 first variation proportional coefficient; The larger value of the gene fragment at the mutation point after the power variable mutates; The smaller value of the gene fragment at the mutation point after the power supply variable mutates;

[0172] For the energy storage variables in individuals, nodes with high power abandonment rates and the corresponding low power abandonment rate nodes in reverse order are selected as gene fragment mutation points, exchange ratio coefficients are randomly generated, and the gene exchange value between energy storage mutation points is calculated:

[0173]

[0174] Where, is a randomly generated second exchange ratio coefficient of 0 to 1. The above rules ensure that the energy storage installed capacity of the corresponding gene site will inevitably increase after the chromosome with a large power curtailment rate crosses.

[0175] For the line-type variables in individuals, the lines with high load rates and the lines with low load rates corresponding to the reverse order are selected as the gene segment variation points, and the gene exchange values ​​between the line-type variation points are calculated according to formula (34).

[0176] Combining the changes of six mutation points to form new offspring individuals Figure 8 This is the process of generating offspring by mutation of the parent generation under the guidance of the knowledge rule proposed in Example 1 of the present invention.

[0177] The operational performance indicators and comprehensive indicators of each planning scheme within the population are evaluated based on time series operation simulation.

[0178] Step 3.5: Convergence check is performed based on the convergence conditions to end the evolution or enter the next evolutionary process, specifically including:

[0179] (1) The optimal comprehensive fitness of the contemporary planning scheme set is calculated using the following formula;

[0180] F min =min(F k ),k=1...N Chrome (35)

[0181] Where, F min 、F k is the optimal comprehensive fitness of the current generation and the fitness of each individual.

[0182] (2) Compare the current optimal fitness with the optimal fitness of the previous three generations. If the difference between the two is continuously less than the set threshold, the genetic algorithm optimization process converges and jumps to step 3.6. Otherwise, proceed to the next step.

[0183] (3) Compare whether the genetic algorithm evolutionary generations have reached the maximum value N iterMax If yes, the optimization process ends and jumps to step 3.6. If no, it returns to step 3.4 and continues the evolution process.

[0184] Step 3.6: Select the individual with the best fitness from the current generation of solutions as the optimization result, and output it as the source-grid-storage collaborative planning solution.

[0185] Embodiment 1 of the present invention proposes a new power system planning method based on the fusion of knowledge rules and evolutionary algorithms. Through the improvements of embedded operation simulation in the planning model and rule-guided genetic algorithm, it can ensure that the planning model is closely integrated with the system operation requirements to avoid separation. At the same time, in the model solution, the large-scale optimization capability of the genetic algorithm and the rapid search capability guided by knowledge rules are combined to achieve comprehensiveness and rapidity of optimization and support large-scale new power system planning.

[0186] In order to fully illustrate the new power system planning method based on the fusion of knowledge rules and evolutionary algorithms proposed in Example 1 of the present invention, a simplified system of a regional power grid is taken as an example to verify the effect of Example 1 of the present invention. Figure 9 This is the topological structure diagram of the starting system of a regional power grid. The system has a total of 90 nodes, 200 lines, 83.2GW thermal power plants, 13GW wind power plants, 18GW photovoltaic power plants, and a total of 2GW energy storage stations. It also includes two voltage levels of 500kV and 100kV, and the maximum initial load is 100GW. Figure 10This is a schematic diagram of a typical daily load curve in a certain area.

[0187] First, follow step 1 to build a source-grid-storage collaborative planning model with embedded time series operation simulation. Investigating the population and economic data of the region, it is predicted that the maximum load in the planned year will increase to 209.4GW. If the typical load curve is still used Figure 10 The graph shown in the figure is combined with the wind and solar resource distribution map of the region to calculate the upper limit of wind and solar resource development in each node area as shown in Table 1. Figure 11 is the typical wind power maximum output curve, Figure 12 is the maximum photovoltaic output curve. Considering the current technical level of mainstream wind power and photovoltaic equipment manufacturers and future technological advances, the unit construction cost and lifespan of wind, solar, and storage equipment and lines in various regions are shown in Tables 1 and 2. Based on the above data, the optimization planning model is formed by substituting it into Equations (4) to (23). At the same time, assuming that the renewable energy penetration rate reaches 50% in the planned year, the installed capacity target calculated using the source-storage planning program without considering network constraints is: 80GW of new wind power generation, 290GW of photovoltaic power generation, and 161.6GW of energy storage stations.

[0188] Table 1: Wind and solar resource development ceiling and construction cost at each node

[0189]

[0190]

[0191]

[0192] Table 2: Unit construction cost and lifespan of lines

[0193] Line Model Maximum forward power / MW Maximum reverse power / MW Cost per unit length (yuan / km) Use life Line Model 51 4300 -4300 2800000 30 51 52 3000 -3000 2700000 30 52 53 2500 -2500 2600000 30 53 71 12300 -12300 6757200 40 71 91 3000 -3000 282000 25 91 92 6000 -6000 282000 25 92

[0194] Use the model decomposition method given in step 2 to divide the original optimization model into planning and operation sub-problems, as expressed in Equations (24) and (25).

[0195] Next, we solve the model using step 3. First, we set the algorithm parameters, including a population size of 20, 5 iterations, a crossover probability of 0.9, and a mutation probability of 0.3. We also set the total installed capacity target as in step 1. Next, we set the target capacity multiplier to 2 and narrowed the optimization space for wind power, photovoltaic power, energy storage, and transmission lines to 9, 13, and 21 nodes, respectively, and 24 transmission lines. We randomly generated the initial population, and after only nine rule-guided generations, we converged to the optimal plan.

[0196] As a comparison, the planning model solution results of the traditional NSGA-Ⅱ algorithm are given, and the convergence curve compared with the method proposed in the patent is as follows: Figure 13 As shown, Figure 13The graph shows the relationship between the optimal comprehensive cost and the number of scanned solutions. The results show that compared to the traditional NSGA-II algorithm, the improved method reduces the number of solutions scanned before convergence from 2,055 to 279, a reduction of 86.4%. The planning time is reduced from 250,358 seconds to 37,569 seconds, a reduction of 85%. The overall cost of the planning solutions is reduced from 62.64 billion yuan to 62.15 billion yuan, a reduction of 1%. This shows that the planning method proposed in this patent only takes about 15% of the time of the original method to obtain a planning solution that is similar to or better than the original method.

[0197] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device comprising a series of elements are inherent to the elements. In the absence of further restrictions, the elements limited by the statement "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device comprising the elements. In addition, the above-mentioned technical solutions provided in the embodiments of the present application are not described in detail in accordance with the corresponding technical solutions in the prior art to achieve the same principle, so as to avoid excessive elaboration.

[0198] Although the above description is of specific embodiments of the present invention in conjunction with the accompanying drawings, it does not limit the scope of protection of the present invention. For those skilled in the art, other different forms of modifications or variations can be made based on the above description. It is not necessary and impossible to list all embodiments here. Based on the technical solution of the present invention, various modifications or variations that can be made by those skilled in the art without expending creative effort are still within the scope of protection of the present invention.

Claims

1. A new power system planning method based on the fusion of knowledge rules and evolutionary algorithms, characterized by: The following steps are involved: Establish a new power system planning model; specifically, first determine the planning boundary parameters of the new power system, then construct the objective function of the power system planning model with the goal of optimizing the comprehensive combination of new equipment investment and system operating costs, and finally establish planning and operating constraints of the power system planning model; Utilizing the relatively independent characteristics of the planning layer and the operation layer, the new power system planning model is transformed into a planning and operation two-layer sub-model to reduce the complexity of a single new power system planning model; including: selecting source storage installed capacity and line upgrade capacity as the handover variables transferred from the planning layer to the operation layer, and selecting economic indicators and operation indicators as the handover variables transferred from the operation layer to the planning layer; the economic indicators include fuel cost, load shedding penalty and total system operation cost; the operation indicators include load shedding rate, power abandonment rate and proportion of renewable energy power; selecting various power sources of each node, load installed capacity and each line upgrade capacity as decision variables of the planning layer; taking the total installed capacity and resource upper limit constraints as constraints, and minimizing the comprehensive investment and operation costs as the optimization goal, a model that depends on operation is formed. The nonlinear continuous optimization subproblem of the planning layer with external input of the line cost; on this basis, the planning layer generates a construction plan for the source, grid and storage installed capacity and passes it to the operation layer; the start and stop status of the thermal power units at each node, the power generation power of various power sources, the charging and discharging power of the energy storage, and the load shedding power are selected as the decision variables of the operation layer, with the power balance of the entire system, the backup demand and the operating boundaries of each equipment as constraints, and the minimization of the operating cost taking into account the fuel, carbon emissions and load shedding penalties as the optimization goal, forming a linear mixed integer optimization subproblem of the operation layer that depends on the source and storage installed capacity of each node and the line reinforcement capacity as external input; on this basis, the operation layer performs production simulation calculations according to the given construction plan to solve the operation subproblem, and thus feeds back the operation indicators of its construction plan to the planning layer; A genetic algorithm guided by knowledge rules is used to solve the planning and operation double-layer sub-model and output the optimal planning scheme for the new power system; specifically: setting the solution parameters of the genetic algorithm and the physical parameters of the planning object; shrinking the search space of power supply equipment, energy storage equipment and transmission lines in turn; taking the optimization variable of the power supply equipment as the main variable, and the energy storage equipment and transmission lines as the dependent variables, generating the initial value of each optimization variable in turn according to the dependency relationship between the decision variables; combining various variables to form an initial population, and verifying that the population meets the overall installation goal; evaluating the operating performance indicators of each planning scheme in the population based on the time series operation simulation, and calculating the fitness of the planning scheme by combining the operating indicators and the investment indicators; promoting the evolution of the planning scheme population through the crossover and mutation of the planning scheme under the guidance of knowledge rules, checking the convergence check according to the convergence conditions, and then ending the evolution process or entering the next generation of evolution, selecting the individual with the best fitness from the current generation of schemes as the optimization result, and outputting the optimal planning scheme for the new power system; The crossover and variation of planning schemes under the guidance of knowledge rules to promote the evolution of planning scheme populations specifically includes: sorting individuals in the population according to the non-dominated sorting principle and fitness index based on the current and previous generation populations to form a hierarchical sorting of dominated solutions and non-dominated solutions; calculating the crowding degree of individuals in each population; forming a parent population from the sorted individuals according to the tournament strategy; crossover of the parent population to generate a child population under the guidance of knowledge rules; and evaluating the operation performance indicators and comprehensive indicators of each planning scheme in the population based on time series operation simulation; The process of generating offspring populations by crossover of parent populations under the guidance of knowledge rules includes: For the power supply variables in individuals, nodes with high power abandonment rates are selected as gene fragment intersections, exchange ratio coefficients are randomly generated, and the values ​​of the gene fragments after intersection at the power supply intersections are calculated; the specific formula is: in, The first gene segment value of the node with high power abandonment rate of the power type variable to be exchanged is selected; The second gene segment value of the node with high power abandonment rate of the power type variable to be exchanged is selected; is the randomly generated 0-1 first exchange ratio coefficient; The values ​​of the chromosomes with larger and smaller power abandonment rates are taken after crossing respectively; For the energy storage variables in individuals, nodes with high power abandonment rates and the corresponding low power abandonment rate nodes in reverse order are selected as gene fragment intersections, exchange ratio coefficients are randomly generated, and the gene exchange values ​​between energy storage intersections are calculated: the specific formula is: These are the values ​​of the corresponding gene fragments before chromosome crossover for energy storage variables with larger and smaller abandonment rates, is the randomly generated 0-1 second exchange ratio coefficient; For the power supply variables in individuals, nodes with high power abandonment rates and nodes with low power abandonment rates corresponding to the reverse order are selected as gene fragment mutation points. Exchange ratio coefficients are randomly generated, and the gene exchange values ​​between power supply mutation points are calculated. Specifically, The first gene segment value of the low power abandonment rate node of the power type variable to be exchanged is selected. The second gene segment value of the low power abandonment rate node of the power type variable to be exchanged is selected; is the randomly generated 0-1 first variation proportional coefficient; The larger value of the gene fragment at the mutation point after the power variable mutates; The smaller value of the gene fragment at the mutation point after the power supply variable mutates; For the energy storage variables in individuals, nodes with high power abandonment rates and the corresponding low power abandonment rate nodes in reverse order are selected as gene fragment mutation points, exchange ratio coefficients are randomly generated, and the gene exchange value between energy storage mutation points is calculated: Where, The second exchange ratio coefficient is randomly generated from 0 to 1.

2. The novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms according to claim 1 is characterized in that: The process of determining the planning boundary parameters of the new power system includes: Determine the maximum load of each node: in, E is the forecast result of the maximum power load in the planning year; s is the economic total forecast data corresponding to the maximum load; N s is the population forecast data corresponding to the maximum load; a1 is the correlation coefficient of the economic aggregate forecast data; a2 is the correlation coefficient of the population forecast data; a0 is the constant term; Determine the boundaries of regional developable resources: in, is the wind and solar resource density in the area where node i is located; S i is the area of ​​the region where node i is located; is the upper limit of resource development in the region where node i is located; η wf,pv The highest conversion efficiency of wind and solar power related to the level of power generation technology; Determine the unit construction cost of wind and solar power equipment at each node in the planning year: in, The unit construction cost of wind and solar power equipment at each node in the planning year; The unit cost benchmark value of wind and solar power equipment at each node; is the installed wind and solar capacity of each node; F(·) is a non-decreasing function; Determine the construction cost of transmission lines with different voltage levels and cross-sectional areas and its corresponding rated transmission capacity 3. The novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms according to claim 2 is characterized in that: The process of constructing the objective function of the power system planning model with the goal of optimizing the comprehensive combination of new equipment investment and system operating costs includes: Using the unit installed cost and discount rate of each type of equipment as input, the annualized investment cost of all equipment in the entire system is constructed; Among them, r0 is the discount rate during the planning period; H type is the service life of type type equipment, n type A set of locations for type type equipment. is the unit cost of type type equipment at node i, is the newly installed capacity of type type equipment at node i; Based on the annual operating costs of thermal power units and the load shedding and power curtailment penalties of the entire power system, the annual operating costs of the entire system are constructed; The annual operating cost of thermal power units is: Among them, F op is the annual operating cost of the thermal power unit, T is the operating cycle of the system, λ coal 、 is the unit coal price and carbon emission cost; P i coal (t) is the actual active power output of the unit at node i at time t, p i The quadratic coefficient of actual active output; q i is the first-order coefficient of actual active power output; l i is the constant term of actual active power output; are the unit positive and negative standby costs, are the positive and negative backups provided by the thermal power unit at node i at time t, are the single start-up and shutdown costs respectively; are the start-up and shutdown indicator variables of the unit, 1 means the unit is on. If it is 1, it means shutdown; Determine load shedding penalties and power curtailment penalties; Among them, F dc is the load shedding penalty for the system, is the penalty cost per unit load shedding power, is the load removed at node i at time t, F drop Fines for abandoned power from wind and solar generators, is the penalty cost of the unit power curtailment, is the power curtailment of the unit at node i at time t; The annual operating cost of the thermal power unit F op , System load shedding penalty F dc and power curtailment penalty F drop Add together to construct the expression for the annual operating cost of the entire system: F ot =F op +F dc +F drop ; (7) Among them, F ot It is the annual operating fee of the entire system; The annualized cost of all equipment investment in the system is I inv and the annual operating cost of the entire system F ot Adding together, the annual comprehensive cost of constructing the planning scheme is: min(I inv +F ot );(8).

4. The novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms according to claim 3 is characterized in that: The process of constructing planning constraints for the power system planning model includes: Calculate the sum of the installed capacity of the entire system to form the total installed capacity constraint of the source and storage; specifically: in, The target construction total capacity for type type equipment, ξ type The maximum allowable error in the installed capacity of type type equipment; Based on the total installed capacity constraints of source and storage, the installed capacity of various types of wind and solar power sources at each node is constructed to meet the upper limit constraints of local resource development; specifically: in, The upper limit of the developable capacity of each type of equipment located at node i.

5. The novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms according to claim 4 is characterized in that: The process of constructing the operational constraints of the power system planning model includes: calculating the sum of the total generation power and the total load power to form a source-load power balance constraint in which the total generation and the total load are equal; Among them, P i coal (t) is the actual active power output of the unit at node i at time t; P i RE (t) is the active power output of the wind and solar generator at node i at time t; is the power curtailment of the unit at node i at time t; is the charging power of the energy storage device at node i at time t; is the discharge power of the energy storage device at node i at time t; P i load (t) is the active power of the load at node i at time t; is the load removed at node i at time t; Taking the load of each node and the output of wind and solar power units, as well as the load shedding power and power abandonment power as input, the reserve demand constraint of thermal power units is constructed; specifically: in, The upper reserve power set for the thermal power unit at time t; L is the lower reserve power set for the thermal power unit at time t; + % is the first reserve requirement coefficient of the load; L - % is the second reserve requirement coefficient of the load; W + % is the first reserve requirement coefficient reserved for wind and solar power unit output; W - % is the second reserve requirement coefficient reserved for wind and solar power unit output; The power flow constraints of the line are as follows: Among them, x ij is the reactance between node i and node j; M is the branch correlation matrix; X is the node impedance matrix; P i,in (t) is the net injected power at node i at time t; P j,in (t) is the net injected power at node j at time t; is the upper limit of the reverse transmission capacity of the line between node i and node j; is the upper limit of the forward transmission capacity of the line between node i and node j; The active output constraint of the thermal power unit is constructed as follows: is the start / stop status of the i-th node unit at time t, is the maximum output ratio of the i-th node unit; is the minimum output ratio of the i-th node unit; is the newly installed capacity of thermal power units at node i; Taking the active power output of the unit and its upper limit as input, the active power output constraint of the thermal power unit considering the reserve output is constructed: The lower reserve power set for the thermal power unit at time t; The upper reserve power set for the thermal power unit at time t; The ramping constraint of the thermal power unit is constructed by taking the active power output of the unit as input: is the maximum climbing rate of the unit; is the maximum descending and climbing rate of the unit; P i coal (t+1) is the actual active power output of the unit at node i at time t+1; Taking the start and stop indicator variables of the unit as input, the minimum start and stop time constraint of the thermal power unit is constructed: Where, T ON is the minimum startup time of the unit; T OFF The minimum stop time of the unit; Taking the wind and solar power output curve and wind and solar power installed capacity as input, the active power output constraint of the wind and solar power generator set is constructed: Among them, P i wf (t) is the actual active power output of the wind turbine at node i at time t; is the maximum value of the normalized wind power curve at time t, that is, the maximum power generation capacity percentage of wind power in the maximum power point tracking mode; is the newly installed capacity of wind turbines at node i; P i pv (t) is the actual active power output of the PV unit at node i at time t; is the maximum value of the normalized photovoltaic power generation curve at time t, that is, the percentage of the maximum power generation capacity of the photovoltaic power source in the maximum power point tracking mode; is the newly installed capacity of the PV unit at node i; The operating characteristic constraints of energy storage equipment are constructed by integrating the charging and discharging power constraints, state of charge constraints, daily clearance constraints, and energy balance constraints of energy storage equipment; Among them, the charge and discharge power constraints of the energy storage equipment are: in, Indicates that the energy storage device is in operation at time t; The charge and discharge power constraints of the energy storage device are used to construct the state of charge constraints of the energy storage device: in, is the state of charge of the energy storage device at time t; is the maximum storage ratio of the energy storage device; is the minimum storage ratio of the energy storage device; Construct the daily clearance constraints of the energy storage equipment based on the state of charge constraints of the energy storage equipment: SOC i ess (t=0h)=SOC i ess (t=24h); (22) Among them, SOC i ess (t=0h) is the state of charge of the energy storage at 0h every day; SOC i ess (t = 24h) is the state of charge of the energy storage at 24h per day; Taking the state of charge and charge and discharge power of the energy storage device as input, the energy balance constraint expression of the energy storage device is constructed: in, is the state of charge of the energy storage device at time t-1; P i ess is the charging and discharging power of the energy storage device at node i at time t; is the rated capacity of the energy storage at node i; ΔT is the time step of energy storage charging and discharging.

6. The novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms according to claim 5 is characterized in that: The nonlinear continuous optimization subproblem of the planning layer is described as: The runtime linear mixed integer optimization subproblem is described as:

7. The novel power system planning method based on the fusion of knowledge rules and evolutionary algorithms according to claim 6 is characterized in that: The process of shrinking the search space for power devices includes: Sort by the unit development cost of each power source at each node in ascending order. The nodes of the bits form a power supply addressable space; The number of nodes planned for the type of power supply is: The upper limit of the installed capacity of the selected node corresponding to the type of power supply is calculated based on the remaining developable capacity of the selected node, the planned installation target, and the planned target magnification factor: in, is the upper limit of the installable capacity of the type power supply of the i-th node; is the upper limit of the remaining development capacity of the type power source at the i-th node; is the installed capacity of type power supply at node i; type The target capacity magnification factor of the entire system for type type power supply; Plan the total installed capacity target for type power supply in the year; The process of shrinking the search space for energy storage devices includes: Calculate the difference in installed energy storage capacity based on the installed capacity of various types of wind and solar power sources and energy storage at each node; sort the nodes in descending order based on the difference in installed energy storage capacity as a characteristic indicator, and select the node before sorting. The nodes of the bit form the energy storage addressable space; among them, The number of nodes planned for energy storage equipment investment and construction; the upper limit of the energy storage capacity that can be installed at each node is calculated based on the energy storage installed capacity difference of the selected nodes, the planned installed capacity target, and the planned target capacity magnification factor, specifically: in, is the upper limit of the installable capacity of energy storage at the i-th node; is the difference in installed capacity of energy storage at the i-th node; The installed capacity of energy storage for node i; is the target capacity of the entire energy storage system; ess is the target capacity magnification factor; is the installed capacity of the wind turbine at node i; η wf The wind power configuration ratio set for the entire system; is the installed capacity of the PV system at node i; η pv The photovoltaic power configuration ratio set for the entire system; The search space process of shrinking the transmission line includes: calculating the comprehensive load rate according to the load rate of each line at each time using the following formula: Among them, Quartile(·) is the upper quartile function; is the load rate of the jth line at time t; is the comprehensive load rate of the line; Sort the lines in descending order based on the comprehensive load rate as the characteristic index, and select The lines of bits form a line set to be upgraded, where Plan the number of upgrades for line equipment; determine the maximum number of lines of the same model that can be connected in parallel based on the line's comprehensive load rate, and form a line upgrade search space, specifically: in, is the maximum upgrade capacity of the k-th line; is the original rated capacity of the kth line; It is the first level threshold of line load factor; is the line load rate classification threshold of the kth line; is the nth level threshold of line load rate; The total number of levels required for line upgrades; The maximum number of lines of the same model that can be connected in parallel for line upgrade; line (·) is the function of the number of line upgrades based on the load factor input.

Images