Long-term planning method for power supply structure in electric power system

Abstract

The invention provides a long-term planning method for a power supply structure in an electric power system. The method comprises the steps of establishment of a low-carbon power supply structure long-term planning model containing a staged target, establishment of an electric power-energy storage-low-carbon long-term planning ETEM model, energy storage technology development uncertainty description and Bayesian stochastic optimization thereof. The invention provides an energy-technology-environment model for long-term planning of a power supply structure in a novel electric power system, combines factors such as electric power planning, energy storage technology development randomness, dual-carbon constraint and the like, constructs a long-term planning model, considers a staged planning target and constraint, is beneficial to planning implementation and path optimization, and improves the planning efficiency. The ever-increasing power supply demand of the planning area is jointly guaranteed; the ETEM and Bayesian stochastic optimization method combining a long-term planning model and stochastic problem solution for the first time in the planning field has good theoretical development and application prospects in the aspects of meeting energy requirements, guaranteeing environmental constraints, adapting to new technology development, even long-term planning of other industries and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of power monitoring, and in particular to a responsive power monitoring real-time calculation method. Background Art

[0002] Traditional power planning methods are mostly based on single-stage optimization models under deterministic conditions, with the goal of economic optimization. They use linear programming or mixed integer programming to determine the installed capacity and layout of conventional power sources such as thermal power and hydropower. However, with the rapid increase in renewable energy penetration, the large-scale integration of power electronic equipment, and the increasingly complex demand-side response mechanisms, existing technologies have shown significant limitations:

[0003] 1. Imbalance in multi-objective trade-offs: Existing planning methods mostly use weighted summation to address multi-objective optimization problems such as economic efficiency, environmental protection, and reliability. This makes it difficult to accurately characterize carbon constraints, ignores key factors such as power planning and the randomness of energy storage technology development, and the importance of thermal power and external power supply. They also fail to consider phased planning objectives and constraints, hindering planning implementation and path optimization.

[0004] 2. Insufficient dynamic adaptability: Due to the uncertainty of large-scale and long-term problems, the selection of various influencing factors and random variables of the planning objects are not fully considered. Random problems in long-term planning will become more and more numerous, making it difficult to meet energy needs and environmental constraints at the same time.

[0005] Therefore, proposing a long-term plan for the power system power structure is a difficult problem that technicians in this field urgently need to solve. Summary of the Invention

[0006] In order to solve the above technical problems, the present invention proposes a long-term planning method for power supply structure in a power system;

[0007] To achieve the above object, the technical solution adopted by the present invention is:

[0008] A method for long-term planning of power supply structure in a power system, the method comprising: S1, establishing a long-term planning model of a low-carbon power supply structure with phased targets; establishing a long-term planning model with phased targets; S2. Establishment of an ETEM model for long-term planning of power, energy storage, and low carbon. An ETEM model with phased targets is established, and based on this, an ETEM model for long-term planning of power, energy storage, and low carbon is established. S3. Characterization of uncertainty in the development of energy storage technology and its Bayesian stochastic optimization The random variables of uncertainty in the development of energy storage technology are characterized and Bayesian stochastic optimization is performed on the random variables.

[0009] Furthermore, the specific method of step S1 is as follows: S1-1, establishing an investment configuration model with the goal of minimizing the unit configuration cost, and the objective function is as follows:

[0010] Where F represents the total cost of the entire planning period, s is the s-th planning period, r represents the annual discount rate, y is the y-th planning year in a planning stage, Finv s is the investment cost, Fmain s is the operation and maintenance cost, Fop s is the energy purchase cost, FC s is the carbon emission cost, and Fgen s is the power generation cost;

[0011] S1-2. The goal of minimizing operation and dispatching costs is used as a constraint, including power balance constraints including typical daily scenarios, electricity balance constraints, energy storage constraints including new technology energy storage, thermal power generation constraints, hydropower generation constraints, photovoltaic power generation constraints, wind power generation constraints, power transmission constraints, peak regulation constraints, carbon emission constraints, CCER equivalent carbon absorption constraints, and phased carbon constraints.

[0012] Furthermore, the investment cost Finvs in step S1-1 is calculated as follows:

[0013]

[0014] Where, The newly installed capacity of equipment i in the planning stage s, including photovoltaic capacity , wind power , existing technology energy storage , new technology energy storage , thermal power and hydropower , the set I of device type i is {i PV ,i WPG ,i ETES ,i NTES ,i TP ,i HP}, ci,ad s is the unit capacity cost of equipment i in stage s;

[0015] The operation and maintenance cost Fmain s is calculated as follows:

[0016]

[0017] Where, in the sth planning stage, yth year, dth typical day scenario, and time t, Pi s,y,d,t is the power output value of device i, ci,main s,y is the operation and maintenance unit price of device i, and γ d is the proportion of the duration of a typical day in a year;

[0018] The purchased energy cost Fop s is calculated as follows:

[0019]

[0020] Where Ptran,in s,y,d,t is the power imported from outside the district, Ptran,out s,y,d,t is the power transmitted to outside the district, cele,in s,y,d,t and cele,out s,y,d,t are the electricity prices corresponding to the power imported from outside the district and the power transmitted to outside the district, respectively.

[0021] The carbon emission cost FCs is calculated as follows:

[0022]

[0023] Where c C is the unit price of carbon emissions, CP s,y is the carbon emissions, Cab s,y is the equivalent carbon absorption, and the difference between the two is the equivalent carbon emissions;

[0024] The power generation cost Fgen s is calculated as follows:

[0025]

[0026] Where Fgen,TPs and Fgen,HPs are the total cost of thermal power generation and hydropower generation, cgen,TPs,y and cgen,HPs,y are the unit power cost of thermal power generation and unit power cost of hydropower generation, respectively; PTPs,y,d,t and PHPs,y,d,t are the output power of the thermal power unit and the hydropower unit, respectively.

[0027] Furthermore, the 365 days of the year are clustered into D typical day scenarios using methods such as K-means clustering. The power balance at time t in the sth planning phase, the yth year, d typical day scenarios, and the power balance constraint for the typical day scenario in step S1-2 is as follows:

[0028]

[0029] Where, is the total power generation, is the total power demand, PPV s,y,d,t is the power generation power of photovoltaic equipment, PWPG s,y,d,t is the power generation power of wind power equipment, PTP s,y,d,t is the power generation power of thermal power equipment, PHP s,y,d,t is the power generation power of hydropower equipment, Ptran,in s,y,d,t is the power from outside the area, PES,dis s,y,d,t is the energy storage discharge power, which is composed of the existing energy storage discharge power PETES,dis s,y,d,t and the new technology energy storage discharge power PNTES,diss,y,d,t, PES,ch s,y,d,t is the energy storage charging power, which is composed of the existing energy storage charging power PNTES,ch s,y,d,t and the new technology energy storage charging power PLTES,ch s,y,d,t, Ptran,out s,y,d,t is the power transmitted to outside the area, and Pload s,y,d,t is the load demand power;

[0030] In long-term planning, the power balance considers the annual balance relationship. In the annual balance, it is necessary to ensure the source-load balance relationship. The power balance constraint is as follows:

[0031]

[0032] Where Ei s,y,d is the amount of electricity absorbed or emitted by device i in a day, including photovoltaic power generation EPV s,y,d, wind power generation EWPG s,y,d, thermal power generation ETP s,y,d, the amount of electricity flowing into the regional system through the tie line Etran,in s,y,d, the amount of electricity flowing out of the regional system through the tie line Etran,out s,y,d, and the load demand Eload s,y,d. The relationship between electricity and power is shown in the second row of the equation.

[0033] In long-term planning, the configuration of energy storage and power structure planning will change with the development of new technology energy storage. The constraint expressions for new technology energy storage (NTES) and existing technology energy storage (ETES) in the entire long-term planning are as follows:

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041] Where, in stage s, EES s,y,d,t, EETES s,y,d,t and ENTES s,y,d,t are the total energy storage capacity, existing technology energy storage capacity and new technology energy storage capacity at time t, respectively. is the current capacity of energy storage using existing technologies, The added capacity of existing technology energy storage, Current capacity for new technology storage, Increased capacity for new technology energy storage, is the total energy storage capacity, e ETES,self and e NTES,self is the self-discharge coefficient of existing energy storage technology and new energy storage technology, e ETES,ch 、e ETES,dis are the energy storage charging and discharging efficiencies of existing technologies, e NTES,ch 、e NTES,dis are the new technology energy storage charging and discharging efficiencies, PETES,ch s,y,d,t and PETES,dis s,y,d,t are the existing technology energy storage charging and discharging powers, PNTES,ch s,y,d,t and PNTES,dis s,y,d,t are the new technology energy storage charging and discharging powers, PES,ch s,y,d,t and PES,dis s,y,d,t are the total energy storage charging and discharging powers, XES s is a 0-1 variable of electrical energy storage, and β ES are the upper and lower limits of the energy storage power capacity ratio, EES s,y,d,24 is the energy storage end power after 24 hours a day, and EES s,y,d,0 is the energy storage initial power before the start of the day;

[0042] The thermal power generation constraints are as follows:

[0043]

[0044]

[0045] Where, PTP s,y,d,t is the thermal power output, is the existing thermal power installed capacity in stage s, The installed capacity of thermal power generated during the s phase;

[0046] Hydropower generation needs to take into account seasonal variations, with output power and power generation varying between the dry and wet seasons. Specifically, full power generation is maintained during the wet season to bear the base load, while during the dry season it primarily undertakes peak load regulation. The hydropower generation constraints are shown in the following equation:

[0047]

[0048]

[0049] Where PHP s,y,d,ts is the hydropower output, η HP is the upper limit of hydropower output ratio in dry season, dk and df are typical days in dry season and wet season respectively, γ dk is the proportion of dry season in the whole year, γ df is the proportion of the flood season to the number of days in the whole year, η E is the ratio of dry season power generation to wet season power generation, is the existing hydropower installed capacity in stage s, The hydropower installed capacity added in the s phase;

[0050] The photovoltaic power generation constraint is as follows:

[0051]

[0052]

[0053] Where η PV is the photovoltaic available capacity factor, is the existing photovoltaic installed capacity in stage s, is the photovoltaic installed capacity added in stage s, and PV is the actual power consumption during this period, which cannot exceed its predicted maximum output. To simplify the calculation, the PV output upper limit during this period can be calculated by multiplying the installed capacity and the output coefficient, assuming that all parameters remain unchanged and only the photovoltaic installed capacity changes.

[0054] The wind power generation constraints are as follows:

[0055]

[0056]

[0057] Where η WPG is the wind power available capacity factor, is the existing wind power installed capacity in stage s, is the wind power installed capacity added in stage s; similar to photovoltaic power, assuming that all parameters remain unchanged and only the wind turbine installed capacity changes;

[0058] The power transfer constraint is expressed as follows:

[0059]

[0060]

[0061] Where Ptran,in,max s and Ptran,out,max s are the maximum values ​​of the power flowing into and out of the system through the tie line in stage s, respectively;

[0062] The peak-shaving constraint takes into account energy storage, thermal power, and hydropower in the dry season. The peak-shaving constraint is as follows:

[0063]

[0064] Where η s Predict the disturbance of output for new energy;

[0065] The carbon emission constraint is as follows:

[0066]

[0067] Where EP s,y is the carbon emissions in a planning year, ω ep is the grid emission factor, and is updated in real time based on the latest values ​​released by the Ministry of Ecology and Environment. TPp is the carbon emission coefficient of thermal power;

[0068] The CCER equivalent carbon absorption constraint is as follows:

[0069]

[0070] Where Eab s,y is the carbon absorption amount of renewable energy output replacing thermal power output in a planning year, ω ab is the carbon equivalent absorption factor;

[0071] The phased carbon constraint is as follows:

[0072]

[0073] .

[0074] Furthermore, the specific method of step S2 is as follows:

[0075] S2-1. The Energy-Technology-Environment Model (ETEM) is a large-scale linear long-term dynamic programming model established from the perspective of mathematical modeling. It features multiple regions, multiple energy types, long time scales, and linear constraints. From a macro perspective, this model establishes a mathematical model of regional or inter-regional energy allocation, energy demand, and resource and environmental factors. It aims to characterize the evolution of energy demand, energy production and use, and the development of new energy through mathematical functions. The deterministic expression characterized by the ETEM is as follows:

[0076]

[0077]

[0078] Where x represents the decision variable for equipment installation in all phases, u represents the decision variable for equipment operation in all time periods, and J is the set of all time segments, including hours t, scenarios s, years y, and phases p. A, B, O, and V are coefficient matrices under the constraints, and b, o, and v are constant column vectors.

[0079] In the long-term planning of power structure, the total planning year is divided into S stages, each stage is Y years, and each year is clustered into D typical day scenarios. Each typical day scenario contains T moments. The ETEM model with staged targets is expressed as follows:

[0080] Where S, Y, D, and T are the sets of planning stage s, planning year y, typical planning day d, and time t within a typical planning day, respectively; A1, A2, B1, and B2 are the coefficient matrices of equality constraints and inequality constraints, respectively;

[0081] S2-2. ETEM provides a mathematical modeling framework. In actual planning, differentiated modeling is required based on specific factors, physical characteristics, different regions, and different objects. Phased long-term power structure planning refines the energy, technology, and environment components of ETEM, thereby constructing the Electricity-Storage-Carbon (ESC) ETEM, or ESC-ETEM. This can be expressed as follows:

[0082]

[0083]

[0084] Where, PNs is the column vector of newly installed equipment capacity, P s,y,d,t The column vector of the output power of the device during operation;

[0085] S2-3. In the long-term planning of power structure, there will be different phased changes and goals in each planning stage. Therefore, based on ESC-ETEM, an ESC-ETEM model with phased goals is constructed. The expression is as follows:

[0086]

[0087]

[0088]

[0089]

[0090] Where, P s,y,d,t =[P(1) s,y,d,t;P(2) s,y,d,t], P(1) Ns is obtained by adding PETES,ad Ns and PNTES,ad Ns to P Ns, so that the matrix dimensions are consistent with the matrix P(1) s,y,d,t, and I8 is the 8-dimensional unit matrix.

[0091] Furthermore, the specific method of step S3 is as follows:

[0092] S3-1. In the long-term planning of power structure, various equipment technologies are constantly developing. With the rapid increase in the proportion of photovoltaic and wind power installed, the emission reduction effect is significant. However, wind and solar power curtailment occurs frequently, resulting in an urgent need to improve the flexible adjustment capacity of the power system. In the long-term planning, many technologies will change, but energy storage technology has been one of the technologies that has developed rapidly in recent years. Therefore, compared with several other technologies, this application focuses on the development of energy storage technology. The uncertainty of energy storage technology development includes market size and technology maturity, which is expressed as follows:

[0093] Where c0 is the initial investment cost, λ S is the annual cumulative installed capacity level of the market, σ is the learning rate, λ T is the annual technology maturity, the number of patents divided by the patent saturation value, which is less than 1, b is the saturation value of investment cost, κ is the slope of the curve, and a is the position parameter, indicating the year when the technology matures and begins to decline; S3-2, market size and technology maturity are uncertain at all stages of long-term planning, and λ is used to represent the uncertainty. S and ξ S Indicates the annual cumulative installed capacity level of the market and its fluctuation, with λ T and ξ T represents the annual technology maturity and its fluctuation; the energy storage price cNTES,ad p of the random variable ξ can be described as a function of market size and technology maturity as follows: ; S3-3, Bayesian random optimization method is a relatively effective global optimization method, which can well solve the problem of uncertainty in energy storage technology cost caused by the uncertainty of two parameters, namely the cumulative installed capacity of the market and the maturity of the technology, in long-term planning. This method first assumes that the prior distribution of random variables is an unknown parameter, then calculates the posterior distribution of the parameter and corrects the prior distribution, and finally solves the random problem of parameter distribution to solve the model uncertainty problem brought by the assumed parameters, thereby avoiding the influence of the uncertainty of the distribution properties of random variables. In order to facilitate the derivation, step S2-1 in Considering the random vector ξ, it can be written as , the general form of the stochastic optimization problem is:

[0094]

[0095] Where χ is a set of variables, and the random vector ξ=[ξ1,ξ2,…,ξ N ] has a value range of Ξ∈R d , G is a random vector ξ in Ξ∈R d The probability distribution on , f (x, ξ) is the target function, and the distribution of ξ is related to the parameter vector θ∈Θ, the parameter set Θ∈R k Therefore, transforming and expanding the above equation yields the following equation:

[0096]

[0097] Where g(ξ|θ) is the probability density function of the distribution ξ conditioned on θ, and has the characteristics of independent and identical distribution. When the estimation of the parameter vector θ is difficult to determine, a Bayesian stochastic optimization model is proposed to reduce the error of the parameter vector θ. The specific expression is as follows:

[0098]

[0099] θ N is a random parameter group whose probability distribution is the prior probability ρ(θ), where E ξ|θ is the expectation of the distribution ξ conditional on θ. And

[0100]

[0101] For random variables Z:Θ→R, the posterior distribution ρ(θ|ξ (N) ), combining the above two equations, we can get:

[0102]

[0103] Where, ρ(θ|ξ (N) ) is the posterior probability, and the expansion is determined by the Bayesian expansion rule:

[0104]

[0105]

[0106] Therefore, the Bayesian stochastic optimization problem is expressed as follows:

[0107] In ESC-ETEM, energy storage technology, especially new energy storage technology, is not mature enough at the current stage. There is not much data on its measurement parameters, namely market size and technology maturity, and it has the characteristics of parameter uncertainty. Therefore, the Bayesian stochastic optimization method is used to solve the problem of uncertain development of energy storage technology in ESC-ETEM. f(x,ξ) can be written as the following expression:

[0108] That is, the ETEM model of electricity-energy storage-low-carbon long-term planning is optimized as follows:

[0109]

[0110] In the formula, due to ξ S and ξ T The probability distributions of are independent and identically distributed, so the probability density function g(ξ S |θ S ) and g(ξ T |θ T ) Considering the joint probability distribution, the respective conditional probability q(ξ S |θ S ) and q(ξ T |θ T ) and the prior distribution ρ(θ S ) and ρ(θ T ) also consider the joint probability distribution.

[0111] The present invention has the following beneficial effects:

[0112] 1. This invention proposes an energy-technology-environment model for the long-term planning of power supply structures in new power systems. It combines factors such as power planning and the randomness of energy storage technology development. It not only constructs a long-term planning model, but also considers phased planning goals and constraints, which is conducive to planning implementation and path optimization, and jointly guarantees the growing power supply demand in the planning area.

[0113] 2. To address the uncertainty inherent in large-scale and long-term planning problems, this paper selected the stochastic nature of the development of energy storage, a key technology, as random variables. This includes the market size and technological maturity of new energy storage and hydrolysis technologies, including new energy storage and hydrogenation. The proposed Bayesian stochastic optimization method is effective in addressing the unknown distribution of random variables and reducing the fluctuation errors of random factors.

[0114] 3. This invention is the first in the planning field to combine the ETEM and Bayesian stochastic optimization methods of long-term planning models with stochastic problem solving. It has good theoretical development and application prospects in meeting energy needs, ensuring environmental constraints, adapting to new technological developments, and even long-term planning in other industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] Figure 1 It is a schematic diagram of the planning process of the present invention;

[0116] Figure 2 A schematic diagram of the framework of the long-term planning model with phased goals in the present invention;

[0117] Figure 3 This is the photovoltaic, wind power and load clustering diagram in the present invention;

[0118] Figure 4 It is a fitting curve diagram of market size and patents in the present invention;

[0119] Figure 5 A comparison chart of the energy storage costs of the new technologies in the three solutions of the present invention;

[0120] Figure 6 This is the cumulative installed capacity and proportion of equipment in solution ④ of the present invention;

[0121] Figure 7 A diagram showing the accumulated power and proportion of each power source of the present invention;

[0122] Figure 8 This is a schematic diagram of equivalent carbon emissions of the present invention;

[0123] Figure 9 Typical daily operation of the present invention Figure 1 ;

[0124] Figure 10 Typical daily operation of the present invention Figure 2 . DETAILED DESCRIPTION

[0125] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0126] Example 1:

[0127] This embodiment provides a method for long-term planning of power supply structure in a power system, the method comprising:

[0128] S1. Establishment of a long-term planning model for a low-carbon power structure with phased targets. S1-1. Establish an investment configuration model with the goal of minimizing unit configuration costs. The objective function is as follows:

[0129] Where F represents the total cost of the entire planning period, s is the s-th planning period, r represents the annual discount rate, y is the y-th planning year in a planning stage, Finv s is the investment cost, Fmain s is the operation and maintenance cost, Fop s is the energy purchase cost, FC s is the carbon emission cost, and Fgen s is the power generation cost;

[0130] The investment cost Finv s is calculated as follows:

[0131]

[0132] Where, The newly installed capacity of equipment i in the planning stage s, including photovoltaic capacity , wind power , existing technology energy storage , new technology energy storage , thermal power and hydropower , the set I of device type i is {i PV ,i WPG ,i ETES ,i NTES ,i TP ,i HP}, ci,ad s is the unit capacity cost of equipment i in stage s;

[0133] The operation and maintenance cost Fmain s is calculated as follows:

[0134]

[0135] Where, in the sth planning stage, yth year, dth typical day scenario, and time t, Pi s,y,d,t is the power output value of device i, ci,main s,y is the operation and maintenance unit price of device i, and γ d is the proportion of the duration of a typical day in a year;

[0136] The purchased energy cost Fop s is calculated as follows:

[0137]

[0138] Where Ptran,in s,y,d,t is the power imported from outside the district, Ptran,out s,y,d,t is the power transmitted to outside the district, cele,in s,y,d,t and cele,out s,y,d,t are the electricity prices corresponding to the power imported from outside the district and the power transmitted to outside the district, respectively.

[0139] The carbon emission cost FCs is calculated as follows:

[0140]

[0141] Where c C is the unit price of carbon emissions, CP s,y is the carbon emissions, Cab s,y is the equivalent carbon absorption, and the difference between the two is the equivalent carbon emissions;

[0142] The power generation cost Fgen s is calculated as follows:

[0143]

[0144] Where Fgen,TP s and Fgen,HP s are the total cost of thermal power generation and hydropower generation respectively. cgen,TP s,y and cgen,HP s,y are the unit power thermal power generation cost and unit power hydropower generation cost respectively. PTP s,y,d,t and PHPs,y,d,t are the output power of the thermal power unit and the hydropower unit respectively.

[0145] S1-2. Constraints are based on the goal of minimizing operation and dispatch costs, including power balance constraints for typical daily scenarios, electricity balance constraints, energy storage constraints including new technology energy storage, thermal power generation constraints, hydropower generation constraints, photovoltaic power generation constraints, wind power generation constraints, power transmission constraints, peak load regulation constraints, carbon emission constraints, CCER equivalent carbon absorption constraints, and phased carbon constraints.

[0146] The power balance constraint for a typical day scenario is as follows:

[0147]

[0148] Where, is the total power generation, is the total power demand, PPV s,y,d,t is the power generation power of photovoltaic equipment, PWPG s,y,d,t is the power generation power of wind power equipment, PTP s,y,d,t is the power generation power of thermal power equipment, PHP s,y,d,t is the power generation power of hydropower equipment, Ptran,in s,y,d,t is the power from outside the area, PES,dis s,y,d,t is the energy storage discharge power, which is composed of the existing energy storage discharge power PETES,dis s,y,d,t and the new technology energy storage discharge power PNTES,diss,y,d,t, PES,ch s,y,d,t is the energy storage charging power, which is composed of the existing energy storage charging power PNTES,ch s,y,d,t and the new technology energy storage charging power PLTES,ch s,y,d,t, Ptran,out s,y,d,t is the power transmitted to outside the area, and Pload s,y,d,t is the load demand power;

[0149] The power balance constraint is as follows:

[0150]

[0151] Where Ei s,y,d is the amount of electricity absorbed or emitted by device i in a day, including photovoltaic power generation EPV s,y,d, wind power generation EWPG s,y,d, thermal power generation ETP s,y,d, the amount of electricity flowing into the regional system through the tie line Etran,in s,y,d, the amount of electricity flowing out of the regional system through the tie line Etran,out s,y,d, and the load demand Eload s,y,d. The relationship between electricity and power is shown in the second row of the equation.

[0152] The energy storage constraints for new technology energy storage are shown below:

[0153]

[0154]

[0155]

[0156]

[0157]

[0158]

[0159]

[0160] Where, in stage s, EES s,y,d,t, EETES s,y,d,t and ENTES s,y,d,t are the total energy storage capacity, existing technology energy storage capacity and new technology energy storage capacity at time t, respectively. is the current capacity of energy storage using existing technologies, The added capacity of existing technology energy storage, Current capacity for new technology storage, Increased capacity for new technology energy storage, is the total energy storage capacity, e ETES,self and e NTES,self is the self-discharge coefficient of existing energy storage technology and new energy storage technology, e ETES,ch 、e ETES,dis are the energy storage charging and discharging efficiencies of existing technologies, e NTES,ch 、e NTES,disare the new technology energy storage charging and discharging efficiencies, PETES,ch s,y,d,t and PETES,dis s,y,d,t are the existing technology energy storage charging and discharging powers, PNTES,ch s,y,d,t and PNTES,dis s,y,d,t are the new technology energy storage charging and discharging powers, PES,ch s,y,d,t and PES,dis s,y,d,t are the total energy storage charging and discharging powers, XES s is a 0-1 variable of electrical energy storage, and β ES are the upper and lower limits of the energy storage power capacity ratio, EES s,y,d,24 is the energy storage end power after 24 hours a day, and EES s,y,d,0 is the energy storage initial power before the start of the day;

[0161] The thermal power generation constraints are as follows:

[0162]

[0163]

[0164] Where, PTP s,y,d,t is the thermal power output, is the existing thermal power installed capacity in stage s, The installed capacity of thermal power generated during the s phase;

[0165] The hydropower generation constraint is expressed as follows:

[0166]

[0167]

[0168] Where PHP s,y,d,ts is the hydropower output, η HP is the upper limit of hydropower output ratio in dry season, dk and df are typical days in dry season and wet season respectively, γ dk is the proportion of dry season in the whole year, γ df is the proportion of the flood season to the number of days in the whole year, η E is the ratio of dry season power generation to wet season power generation, is the existing hydropower installed capacity in stage s, The hydropower installed capacity added in the s phase;

[0169] The photovoltaic power generation constraint is as follows:

[0170]

[0171]

[0172] Where ηPV is the photovoltaic available capacity factor, is the existing photovoltaic installed capacity in stage s, is the photovoltaic installed capacity added in stage s, and PV is the actual power consumption during this period, which cannot exceed its predicted maximum output;

[0173] The wind power generation constraints are as follows:

[0174]

[0175]

[0176] Where η WPG is the wind power available capacity factor, is the existing wind power installed capacity in stage s, The wind power installed capacity added in stage s;

[0177] The power transfer constraint is expressed as follows:

[0178]

[0179]

[0180] Where Ptran,in,max s and Ptran,out,max s are the maximum values ​​of the power flowing into and out of the system through the tie line in stage s, respectively;

[0181] The peak shaving constraint is as follows:

[0182]

[0183] Where η s Predict the disturbance of output for new energy;

[0184] The carbon emission constraint is as follows:

[0185]

[0186] Where EP s,y is the carbon emissions in a planning year, ω ep is the grid emission factor, ω TPp is the carbon emission coefficient of thermal power;

[0187] The CCER equivalent carbon absorption constraint is as follows:

[0188]

[0189] Where Eab s,y is the carbon absorption amount of renewable energy output replacing thermal power output in a planning year, ω ab is the carbon equivalent absorption factor;

[0190] The phased carbon constraint is as follows:

[0191]

[0192]

[0193] S2. Establishment of an ETEM model for long-term power-energy storage-low-carbon planning. S2-1. Divide the total planning year into S phases, each phase lasting Y years. Each year is clustered into D typical daily scenarios, each containing T moments. The ETEM model with phased targets is expressed as follows:

[0194] Where S, Y, D, and T are the sets of planning stage s, planning year y, typical planning day d, and time t within a typical planning day, respectively; A1, A2, B1, and B2 are the coefficient matrices of equality constraints and inequality constraints, respectively;

[0195] S2-2. The energy, technology, and environment in the ETEM model are refined to correspond to electricity and energy storage, respectively, to construct an ETEM model of electricity-energy storage-low carbon, namely ESC-ETEM, which can be expressed as follows:

[0196]

[0197]

[0198] Where, PNs is the column vector of newly installed equipment capacity, P s,y,d,t is the output power column vector of the equipment during operation; S2-3, based on the expression in step S2-2, construct an ESC-ETEM model with phased targets, the expression is as follows:

[0199]

[0200]

[0201]

[0202]

[0203] Where, P s,y,d,t=[P(1) s,y,d,t;P(2) s,y,d,t], P(1) Ns is obtained by adding PETES,ad Ns and PNTES,ad Ns to P Ns, so that the matrix dimensions are consistent with those of the matrix P(1) s,y,d,t, and I8 is an 8-dimensional unit matrix. S3. Characterization of uncertainty in the development of energy storage technology and its Bayesian stochastic optimization S3-1. The uncertainty in the development of energy storage technology includes market size and technology maturity, which can be expressed as follows: Where c0 is the initial investment cost, λ S is the annual cumulative installed capacity level of the market, σ is the learning rate, λ T is the annual technology maturity, the number of patents divided by the patent saturation value, which is less than 1, b is the saturation value of investment cost, κ is the slope of the curve, and a is the position parameter, indicating the year when the technology matures and begins to decline; S3-2, market size and technology maturity are uncertain at all stages of long-term planning, and λ is used to represent the uncertainty. S and ξ S Indicates the annual cumulative installed capacity level of the market and its fluctuation, with λ T and ξ T represents the annual technology maturity and its fluctuation; the development of science and technology will lead to a decrease in energy storage costs, so the energy storage price cNTES,ad p of the random variable ξ can be described as a function of market size and technology maturity as follows: ; S3-3, Bayesian stochastic optimization problem is expressed as follows:

[0204] Where f(x,ξ) is expressed as follows:

[0205] That is, the ETEM model of electricity-energy storage-low-carbon long-term planning is optimized as follows:

[0206]

[0207] In the formula, due to ξ S and ξ T The probability distributions of are independent and identically distributed, so the probability density function g(ξ S |θ S ) and g(ξ T |θ T ) Considering the joint probability distribution, the respective conditional probability q(ξ S |θ S ) and q(ξ T |θ T ) and the prior distribution ρ(θ S ) and ρ(θ T ) also consider the joint probability distribution.

[0208] Example 2:

[0209] This example uses the planning method provided in Example 1 to perform long-term planning for a power system. The power system planning period is 40 years, consisting of eight phases S. Each phase s consists of five years Y. Each year y has three typical scenarios D (summer, winter, and transition). Each typical scenario d has 24 hours t, and an annual load growth rate of 7% is assumed. To obtain the typical daily scenario D, K-means is used to cluster the per-unit values ​​of photovoltaic power output, wind power output, and electrical load for 8760 hours per year, resulting in the following: Figure 3 The photovoltaic power output per unit value, wind power output per unit value and power load conditions under the three typical days D shown represent typical days in summer, typical days in winter and typical days in the transition season, respectively.

[0210] Table 1 shows the initial installed capacity and cost of the equipment. Since thermal and hydropower technologies are mature and their construction costs fluctuate more slowly than those of renewable energy and energy storage, the parameters here are set at constant prices, with an annual discount rate of 8%. Currently, lithium is the primary electrochemical energy storage technology and is in high demand. Sodium ore is relatively abundant and offers advantages such as good thermal stability and low-temperature fast charging. Therefore, lithium batteries were selected as the existing energy storage technology for research, while sodium batteries were selected as the new energy storage technology for research. All storage options were for four hours.

[0211] Table 1 Initial installed capacity and installation cost of equipment

[0212] Equipment Type Installed capacity / 10,000 kW <![CDATA[Cost of installed capacity / yuan·kW -1 > <![CDATA[Operation and maintenance cost / yuan·kW -1 > Photovoltaics 1600 See attached Figure A1 0.58 fan 1350 See attached Figure A1 0.625 thermal power 14000 40000 0.08 hydropower 12600 70000 0.02 Existing technology energy storage 350 See Appendix A1 0.08 New technology energy storage 0 — 0.05

[0213] Energy storage cost and market size λ S and patent lambda in TRL T The change curve of market size and patents is obtained by fitting the report data of Zheshang Securities and the public data of the International Energy Agency (IEA), such as Figure 4 shown.

[0214] Any technology will go through five stages: emergence, rise, maturity, decline, and extinction, which together constitute the technology maturity curve. The lithium battery data in the report data of Zheshang Securities shows that lithium batteries have also experienced the technology maturity curve and are 20 years earlier than sodium batteries. Figure 3 It can be seen that there is a certain error between the real data and the fitted curve, which is the market installed capacity fluctuation ξ mentioned above. S and TRL fluctuation ξ T ,pass Figure 4 The probability density functions of the samples obtained by sampling the errors between the real data and the fitted curve are shown in Table 2.

[0215] Table 2 ξ S and ξ T Sample probability distribution table

[0216] Serial number <![CDATA[ξ S ]]> <![CDATA[ξ T <!-- 14 -->]]> 1 <![CDATA[N(5.56, 25.30 2 )]]> <![CDATA[N(-3.19, 25.61 2 )]]> 2 <![CDATA[N(5.69, 25.30 2 )]]> <![CDATA[N(-3.01, 25.45 2 )]]> 3 <![CDATA[N(6.44, 25.13 2 )]]> <![CDATA[N(-2.39, 24.78 2 )]]> 4 <![CDATA[N(5.98, 25.20 2 )]]> <![CDATA[N(-2.22, 24.83 2 )]]> 5 <![CDATA[N(6.11, 25.19 2 )]]> <![CDATA[N(-3.11, 26.25 2 )]]>

[0217] Since the final optimization integral calculation of the ETEM model of power-energy storage-low-carbon long-term planning is complex, the sample mean method is used to estimate it. In order to obtain an intuitive impression of the updated prior distribution of the probability density function and facilitate calculation, we only need to pay attention to the expected sample average approximation. The conjugate distribution of the probability density function can be directly taken as the prior distribution. The conjugate prior distribution of the normal distribution is a normal distribution. Therefore, according to the sample mean and variance, the prior distribution ρ(θ S ) obeys N~(5.75,0.44), ρ(θ T ) obeys N~(-2.75,0.43). Assuming the planning start year is 2021, this embodiment establishes a mixed integer linear model and calls a commercial solver in MATLAB for direct solution.

[0218] First, the results of the schemes with and without ESC-ETEM are compared and are shown in Table 3.

[0219] Table 3 Results of whether the ESC-ETEM scheme was adopted

[0220] Because the ESC-ETEM multi-stage planning approach was not adopted, the planning and operation results were relatively crude, resulting in higher costs. The equivalent carbon emissions and equivalent carbon absorption values ​​indicate that this solution has a low proportion of installed renewable energy capacity. However, the total cost of the ESC-ETEM approach was significantly lower than that without the ESC-ETEM approach, with a particularly significant reduction in operation and maintenance costs. This is because the inclusion of the ETEM model, through the multi-stage planning configuration results, enables more precise control and meeting of load demand.

[0221] Next, based on the scheme using the ESC-ETEM method, this embodiment considers multiple factors and sets four schemes for example simulation. The scheme configuration considerations are shown in Table 4, and the example results are compared in Table 5.

[0222] Table 4 Factors considered for the four options

[0223] Each plan and each factor Plan ① Plan ② Plan ③ Plan ④ ESC-ETEM use use use use New energy storage technologies Not considered consider consider consider Random variables of new energy storage technologies Not considered Not considered consider consider Random variable distribution Not considered Not considered Known unknown Bayesian stochastic optimization method Not considered Not considered Not considered use

[0224] Table 5 Simulation results of four schemes

[0225]

[0226] Scheme ① in Tables 4 and 5 represents the case using the ESC-ETEM method in Table 3. Comparing the simulation results for Schemes ①, ②, ③, and ④ reveals that all costs are reduced after accounting for the new energy storage technology. Further considering the impact of randomness, while configuration and operation and maintenance costs increase slightly after accounting for the randomness of the new energy storage technology, energy purchase costs, carbon emissions costs, and power generation costs all decrease, and equivalent carbon emissions are also lower, indicating less reliance on external energy sources and a smaller impact on the environment. The final cost calculated using the Bayesian stochastic optimization method can reduce the impact of random variables and lower the total cost.

[0227] Further comparison of the cost function of the new technology energy storage of the above scheme shows that the cost curve optimized by the Bayesian stochastic optimization method has a smaller error and is more consistent with the actual cost data. The specific results are as follows: Figure 5 and as shown in Table 6.

[0228] Table 6. Error values ​​of energy storage costs for new technologies

[0229] years Unit price cost error in Plan ② Unit price cost error in plan ③ Unit price cost error in Plan ④ 2020 55.48 33.02 29.58 2021 53.85 45.36 20.64 2022 32.53 37.62 32.08 2023 29.63 47.83 25.71

[0230] Depend on Figure 5 As shown in Table 6, the energy storage cost of the new technology in Scheme ④ is slightly higher than that of the other two schemes in the early planning stage. However, in the later planning stage, the energy storage cost is significantly lower than that of the other two schemes. In addition, the error between the cost curve and the actual value in Scheme ④ is the smallest, indicating that the energy storage cost after Bayesian stochastic optimization is more in line with the actual situation, has high practical application value, and provides good data support for long-term planning.

[0231] To verify the sensitivity of the constructed model, the discount rate and the unit price of existing energy storage technology were adjusted. Scheme 5 sets the annual discount rate to 5%, while other parameters remain unchanged. Scheme 6 sets the existing energy storage cost to 102%, while other parameters remain unchanged. The simulation results are shown in Table 7.

[0232] Table 7 Parameter sensitivity analysis results

[0233] Example

[0234] This embodiment is based on the analysis of the installed capacity configuration results of Example 2. Long-term planning is carried out under the conditions of the new technology energy storage cost optimized by the Bayesian stochastic optimization method, and the cumulative installed capacity and proportion of each equipment in each stage are obtained as shown in the figure below. Figure 6 shown.

[0235] According to the calculation results of Scheme ④ in Table 5 and Figure 6 It can be seen that:

[0236] (1) The scale and proportion of clean energy installed capacity increased significantly. The installed capacity of photovoltaic and wind turbines increased significantly. The installed capacity of the two increased sharply from 29.5 million kW in the early planning period to about 1.5 billion kW in the late planning period. The proportion of installed capacity increased from about 10% to about 70%. The installed capacity growth rate from the second phase to the third phase was the largest, at 73.72%. In the early planning period, hydropower capacity increased significantly, from 126 million kW in the early stage to 13,000 kW. Analysis shows that hydropower's better peak-shaving capacity and pollution-free nature replaced part of thermal power and promoted low-carbon effects. In the late planning period, the proportion of hydropower installed capacity decreased from 42% to 6%. This result reflects that the installed capacity of hydropower is limited by the exploitable capacity of water resources.

[0237] (2) Thermal power installed capacity shows a trend of first increasing and then decreasing. Analysis shows that in the early planning period, in order to meet the load growth and when the cost of installing new energy is still at a relatively high level, thermal power is the main power supply mode. In the later planning period, as the target approaches, with the increase in installed capacity of new energy and energy storage, they can basically meet the growing load demand together with hydropower. Thermal power mainly plays a peak-shaving role. Therefore, some idle capacity is retired, and the proportion of thermal power installed capacity also decreases.

[0238] (3) The installed capacity of energy storage increased rapidly in the later stages of the planning. In the later stages of the planning, as energy storage technology matured, the installed capacity of energy storage increased rapidly. The installed capacity of the two forms of energy storage increased from 3.5 million kW to 368 million kW, and the installed capacity ratio increased from 1.17% to 17.41%. This also led to an increase in the peak-shaving capacity of energy storage, which in turn replaced part of the peak-shaving capacity of thermal power, improved the utilization rate of new energy, increased the charging and discharging power, reduced the curtailment of wind and solar power, and improved the regulation capacity of the system.

[0239] Embodiment 4:

[0240] This embodiment is based on the analysis of the power supply results of Example 3. Figure 6 The installed configuration scheme can be used to determine the power supply of each unit under the configuration scheme, such as Figure 7 As shown, it can be seen that:

[0241] (1) Clean energy power generation has grown steadily. In 2030, clean energy power generation will be 6.11 trillion kWh, accounting for 57.33%. In 2060, clean energy power generation will increase to 22.41 trillion kWh, accounting for 79.99%. Among them, the power generation of new energy will increase significantly, from 3.32 trillion kWh in 2030 to 17.74 trillion kWh in 2060, and the proportion will increase from 31.10% to 63.32%. Analysis of the reasons shows that in the early planning stage, the installed capacity of new energy is small and the utilization hours are low, resulting in less power generation. In the later stage, the installed capacity increases, the utilization hours gradually increase, and the proportion of power generation also gradually increases. At the same time, hydropower generation is greatly affected by seasonality. Hydropower maintains full power generation in the flood season and plays a peak-shaving role in the dry season. The proportion of power generation decreases as the proportion of installed capacity gradually decreases.

[0242] (2) The proportion of electricity from outside the region shows a trend of first rising and then falling, with a peak proportion of 15.39%, and then falling to 12.51% in 2060. At the same time, the thermal power generation shows a trend of first rising and then falling, and this trend is similar to the thermal power installed capacity. In 2060, with the increase of energy storage capacity, the peak regulation capacity will gradually increase, and the thermal power generation will decrease as the thermal power capacity decreases, and the proportion of overall power generation will also decrease. This shows that in the later stage of planning, new energy has become the main source of electricity.

[0243] Example 5:

[0244] This example is based on the results of Example 4. Figure 7 The installed capacity configuration of each machine is obtained under this configuration scheme. Figure 8 and as shown.

[0245] Carbon emissions peak in 2030 and remain constant thereafter. As load increases annually, clean energy fills the supply gap, leading to a corresponding increase in equivalent carbon absorption. This also offsets some of the power supply from thermal power and external electricity, resulting in a corresponding decrease in carbon emissions in the final stage.

[0246] Example 6:

[0247] This embodiment is based on the analysis of the operating conditions under the installed capacity configuration of embodiment 3. Figure 9 and Figure 10 In the figure, negative numbers represent the supply side, positive numbers represent the demand side, 0-24h represents a typical summer day, which is the high-flow period with full hydropower generation, 25-48h represents a typical winter day, and 49-72h represents a typical transition season day, which is the dry season, when hydropower plays a peak-shaving role.

[0248] according to Figure 6 Compare the installed capacity configuration of Figure 9 and Figure 10 Comprehensive analysis shows that:

[0249] (1) Clean energy output is significantly affected by seasonality. On a typical day in the transition season, renewable energy output is the highest, while hydropower output is relatively low. Analysis shows that renewable energy utilization hours on this typical day are large, making it the main source of power supply. It can transfer surplus electricity to energy storage to reduce wind and solar power curtailment. However, hydropower cannot be fully generated during the dry season, so hydropower, energy storage, thermal power, and external power jointly bear the role of peak regulation, which also leads to a low utilization hour of thermal power and external power. On a typical day in summer, renewable energy output is relatively low, hydropower output is high, and energy storage participates in partial peak regulation. Analysis shows that this typical day is a flood season, hydropower maintains full generation and high utilization hours. Although renewable energy utilization hours are relatively low compared to a typical day in summer, there is still surplus electricity to charge energy storage. On this typical day, energy storage, thermal power, and external power participate in peak regulation, thereby reducing the utilization hours of thermal power and external power. On a typical winter day, the output of hydropower and renewable energy is relatively low, while the output of thermal power and external power is generally high, and energy storage does not participate in peak regulation. Analysis shows that this typical day falls into the dry season, hydropower cannot be fully generated, and the output of renewable energy is relatively small. The utilization hours of hydropower and renewable energy are relatively low. At the same time, there is a problem of large load demand, so thermal power and external power are required to become the main sources of electricity, resulting in the highest utilization hours of thermal power and external power.

[0250] (2) Vigorously developing energy storage can improve peak-shaving capacity. Energy storage has a stronger peak-shaving capacity during operation, which reduces the output of thermal power and external power. This shows that with the maturity of new energy storage technologies, the installed capacity of new energy storage technologies has gradually increased, the peak-shaving methods have become more diverse, and the peak-shaving capacity reserves have become more abundant. It can replace the peak-shaving capacity of thermal power and external power, thereby playing a role in low-carbon peak-shaving.

[0251] In combination with Examples 1 to 6, this application proposes an energy-technology-environment model for the long-term planning of power supply structures in new power systems. It combines factors such as power planning and the randomness of energy storage technology development. It not only constructs a long-term planning model, but also takes into account phased planning goals and constraints, which is conducive to planning implementation and path optimization. The results show that while vigorously installing new energy, it is also necessary to consider the development of new energy storage technologies, coordinate the planning and configuration of existing energy storage technologies and new energy storage technologies, and cannot ignore the supply status of thermal power and external power, so as to jointly guarantee the growing power supply demand in the planning area.

[0252] Considering the uncertainty inherent in large-scale and long-term planning problems, the stochastic nature of the development of energy storage, a key technology, was selected as a random variable. This included the market size and technological maturity of new energy storage and hydrolysis technologies, including new energy storage and hydrogen. The proposed Bayesian stochastic optimization method was effective in addressing the unknown distribution of random variables and reducing the fluctuation error of random factors.

[0253] Long-term power generation planning will require more comprehensive and precise consideration of primary energy composition, new technology development, and environmental indicators, leading to an increasing number of stochastic issues in long-term planning. The ETEM and Bayesian stochastic optimization methods, which combine long-term planning models with stochastic problem solving, are pioneering in the planning field. While they fail to fully consider the influencing factors and random variables involved in planning, they hold great promise for theoretical development and application in meeting energy demand, safeguarding environmental constraints, adapting to new technology development, and even in long-term planning for other industries.

Claims

1. A method for long-term planning of power supply structure in a power system, characterized by: The method includes: S1. Establishing a long-term planning model for a low-carbon power structure with phased targets; S2. Establishing an ETEM model for long-term planning of electricity-energy storage-low-carbon; S3. Characterizing the uncertainty of energy storage technology development and its Bayesian stochastic optimization; Characterizing the random variables of the uncertainty of energy storage technology development and performing Bayesian stochastic optimization on the random variables.

2. The method for long-term planning of power supply structure in a power system according to claim 1, characterized in that: The specific method of step S1 is as follows: S1-1, establishing an investment configuration model with the goal of minimizing the unit configuration cost, and the objective function is as follows: ; Where F represents the total cost of the entire planning period, s is the s-th planning period, r represents the annual discount rate, y is the y-th planning year in a planning stage, Finv s is the investment cost, Fmain s is the operation and maintenance cost, Fop s is the energy purchase cost, FCs is the carbon emission cost, and Fgen s is the power generation cost; S1-2. The goal of minimizing operation and dispatching costs is used as a constraint, including power balance constraints including typical daily scenarios, electricity balance constraints, energy storage constraints including new technology energy storage, thermal power generation constraints, hydropower generation constraints, photovoltaic power generation constraints, wind power generation constraints, power transmission constraints, peak regulation constraints, carbon emission constraints, CCER equivalent carbon absorption constraints, and phased carbon constraints.

3. The method for long-term planning of power supply structure in a power system according to claim 2, characterized in that: The investment cost Finvs in step S1-1 is calculated as follows: ； Where, The newly installed capacity of equipment i in the planning stage s, including photovoltaic capacity , wind power , existing technology energy storage , new technology energy storage , thermal power and hydropower , the set I of device type i is {i PV ,i WPG ,i ETES ,i NTES ,i TP ,i HP }, ci,ad s is the unit capacity cost of equipment i in stage s; The operation and maintenance cost Fmain s is calculated as follows: ； Where, in the sth planning stage, yth year, dth typical day scenario, and time t, Pi s,y,d,t is the power output value of device i, ci,main s,y is the operation and maintenance unit price of device i, and γ d is the proportion of the duration of a typical day in a year; The purchased energy cost Fop s is calculated as follows: ； Where Ptran,in s,y,d,t is the power imported from outside the district, Ptran,out s,y,d,t is the power transmitted to outside the district, cele,in s,y,d,t and cele,out s,y,d,t are the electricity prices corresponding to the power imported from outside the district and the power transmitted to outside the district, respectively. The carbon emission cost FCs is calculated as follows: ； Where c C is the unit price of carbon emissions, CP s,y is the carbon emissions, Cab s,y is the equivalent carbon absorption, and the difference between the two is the equivalent carbon emissions; The power generation cost Fgen s is calculated as follows: ； Where Fgen,TPs and Fgen,HPs are the total cost of thermal power generation and hydropower generation, cgen,TPs,y and cgen,HPs,y are the unit power cost of thermal power generation and unit power cost of hydropower generation, respectively; PTPs,y,d,t and PHPs,y,d,t are the output power of the thermal power unit and the hydropower unit, respectively.

4. The method for long-term planning of power supply structure in a power system according to claim 3, characterized in that: The power balance constraint for the typical day scenario in step S1-2 is as follows: ； Where, is the total power generation, is the total power demand, PPV s,y,d,t is the power generation power of photovoltaic equipment, PWPG s,y,d,t is the power generation power of wind power equipment, PTP s,y,d,t is the power generation power of thermal power equipment, PHP s,y,d,t is the power generation power of hydropower equipment, Ptran,in s,y,d,t is the power from outside the area, PES,dis s,y,d,t is the energy storage discharge power, which consists of the existing energy storage discharge power PETES,dis s,y,d,t and the new technology energy storage discharge power PNTES,dis s,y,d,t, PES,ch s,y,d,t is the energy storage charging power, which consists of the existing energy storage charging power PNTES,ch s,y,d,t and the new technology energy storage charging power PLTES,ch s,y,d,t, Ptran,out s,y,d,t is the power transmitted to outside the area, and Pload s,y,d,t is the load demand power; The power balance constraint is as follows: ； Where Ei s,y,d is the amount of electricity absorbed or emitted by device i in a day, including photovoltaic power generation EPV s,y,d, wind power generation EWPG s,y,d, thermal power generation ETP s,y,d, the amount of electricity flowing into the regional system through the tie line Etran,in s,y,d, the amount of electricity flowing out of the regional system through the tie line Etran,out s,y,d, and the load demand Eload s,y,d. The relationship between electricity and power is shown in the second row of the equation. The energy storage constraints for new technology energy storage are shown below: ； ； ； ； ； ； ； Where, in stage s, EES s,y,d,t, EETES s,y,d,t and ENTES s,y,d,t are the total energy storage capacity, existing technology energy storage capacity and new technology energy storage capacity at time t, respectively. is the current capacity of energy storage using existing technologies, The added capacity of existing technology energy storage, Current capacity for new technology storage, Increased capacity for new technology energy storage, is the total energy storage capacity, e ETES,self and e NTES,self is the self-discharge coefficient of existing energy storage technology and new energy storage technology, e ETES,ch 、e ETES,dis are the energy storage charging and discharging efficiencies of existing technologies, e NTES,ch 、e NTES,dis are the new technology energy storage charging and discharging efficiencies, PETES,ch s,y,d,t and PETES,dis s,y,d,t are the existing technology energy storage charging and discharging powers, PNTES,ch s,y,d,t and PNTES,dis s,y,d,t are the new technology energy storage charging and discharging powers, PES,ch s,y,d,t and PES,dis s,y,d,t are the total energy storage charging and discharging powers, XES s is a 0-1 variable of electrical energy storage, and β ES are the upper and lower limits of the energy storage power capacity ratio, EES s,y,d,24 is the energy storage end power after 24 hours a day, and EES s,y,d,0 is the energy storage initial power before the start of the day; The thermal power generation constraints are as follows: ； ； Where, PTP s,y,d,t is the thermal power output, is the existing thermal power installed capacity in stage s, The installed capacity of thermal power generated during the s phase; The hydropower generation constraint is as follows: ； ； Where PHP s,y,d,ts is the hydropower output, η HP is the upper limit of hydropower output ratio in dry season, dk and df are typical days in dry season and wet season respectively, γ dk is the proportion of dry season in the whole year, γ df is the proportion of the flood season to the number of days in the whole year, η E is the ratio of dry season power generation to wet season power generation, is the existing hydropower installed capacity in stage s, The hydropower installed capacity added in the s phase; The photovoltaic power generation constraint is as follows: ； ； Where η PV is the photovoltaic available capacity factor, is the existing photovoltaic installed capacity in stage s, is the photovoltaic installed capacity added in stage s, and PV is the actual power consumption during this period, which cannot exceed its predicted maximum output; The wind power generation constraints are as follows: ； ； Where η WPG is the wind power available capacity factor, is the existing wind power installed capacity in stage s, The wind power installed capacity added in stage s; The power transfer constraint is expressed as follows: ； ； Where Ptran,in,max s and Ptran,out,max s are the maximum values ​​of the power flowing into and out of the system through the tie line in stage s, respectively; The peak shaving constraint is as follows: ； Where η s Predict the disturbance of output for new energy; The carbon emission constraint is as follows: ； Where EP s,y is the carbon emissions in a planning year, ω ep is the grid emission factor, ω TPp is the carbon emission coefficient of thermal power; The CCER equivalent carbon absorption constraint is as follows: ； Where Eab s,y is the carbon absorption amount of renewable energy output replacing thermal power output in a planning year, ω ab is the carbon equivalent absorption factor; The phased carbon constraint is as follows: ； 。 5. The method for long-term planning of power supply structure in a power system according to claim 1, characterized in that: The specific method of step S2 is as follows: S2-1. The overall planning year is divided into S phases, each phase is Y years, and each year is clustered into D typical day scenarios. Each typical day scenario contains T moments. The ETEM model with phased targets is expressed as follows: ； ； Where S, Y, D, and T are the sets of planning stage s, planning year y, typical planning day d, and time t within a typical planning day, respectively; A1, A2, B1, and B2 are the coefficient matrices of equality constraints and inequality constraints, respectively; S2-2. Refine the energy, technology, and environment in the ETEM model separately to construct the ETEM model of electricity-energy storage-low carbon, namely ESC-ETEM, which can be expressed as follows: ； ； Where, PNs is the column vector of newly installed equipment capacity, P s,y,d,t is the output power column vector of the equipment during operation; S2-3, based on the expression in step S2-2, construct an ESC-ETEM model with phased targets, the expression is as follows: ； ； ； ； Where, P s,y,d,t =[P(1) s,y,d,t;P(2) s,y,d,t], P(1) Ns is obtained by adding PETES,ad Ns and PNTES,ad Ns to P Ns, so that the matrix dimensions are consistent with the matrix P(1) s,y,d,t, and I8 is the 8-dimensional unit matrix.

6. The method for long-term planning of power supply structure in a power system according to claim 1, characterized in that: The specific method of step S3 is as follows: S3-1. The uncertainty of energy storage technology development includes market size and technology maturity, which can be expressed as follows: ; In the formula, c0 is the initial investment cost, λ S is the annual cumulative installed capacity level of the market, σ is the learning rate, λ T is the annual technology maturity, which is the number of patents divided by the patent saturation value, and the value is less than 1. b is the saturation value of investment cost, κ is the slope of the curve, and a is the position parameter, indicating the year when the technology matures and begins to decline. S3-2. Market size and technology maturity are uncertain at all stages of long-term planning. S and ξ S Indicates the annual cumulative installed capacity level of the market and its fluctuation, with λ T and ξ T represents the annual technology maturity and its fluctuation; the energy storage price cNTES,adp of the random variable ξ can be described as a function of market size and technology maturity as follows: ; S3-3, Bayesian stochastic optimization problem is expressed as follows: ; Where f(x,ξ) is expressed as follows: ; That is, the ETEM model of electricity-energy storage-low-carbon long-term planning is optimized as follows: ； In the formula, due to ξ S and ξ T The probability distributions of are independent and identically distributed, so the probability density function g(ξ S |θ S ) and g(ξ T |θ T ) Considering the joint probability distribution, the respective conditional probability q(ξ S |θ S ) and q(ξ T |θ T ) and the prior distribution ρ(θ S ) and ρ(θ T ) also consider the joint probability distribution.

Images