Power evaluation method and system based on multi-element flexible resource participation

Abstract

The invention belongs to the technical field of power systems and energy management, and discloses a power evaluation method and system based on multi-element flexible resource participation, and the method comprises the steps: obtaining power evaluation heterogeneous data; preprocessing the obtained power evaluation heterogeneous data to obtain a heterogeneous feature data set; a multi-element flexibility resource model is constructed through the heterogeneous feature data set, and power generation output of different resources is obtained through prediction of the multi-element flexibility resource model; according to the generated output of different resources, a resource scheduling strategy is formulated and executed; through comprehensive scheduling of multi-element flexible resources, investment of resources can be adjusted according to real-time changes of demands and power production, so that a power system has better adaptability and stability; and the utilization efficiency of various generator sets can be improved, and resource waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power system and energy management, and more specifically, to a power evaluation method and system based on the participation of multiple flexibility resources. Background Art

[0002] The patent with the patent announcement number CN117543713A discloses a method and system for evaluating the flexibility of a power system. The present invention introduces the investment portfolio high-order moment analysis theory to characterize the flexibility of the power system in order to reveal the regulatory potential and risk of system flexibility. A flexibility unit probability model considering spatial correlation is constructed through a multivariate Copula function. A flexibility evaluation index based on high-order moments is established, and an index calculation method based on kernel density estimation and Monte Carlo simulation is given. The proposed index can not only reflect the average level and stability of the system's flexibility regulation capability, but also its potential and risk; it can quantitatively evaluate the impact of the types of flexibility resources and investment and construction areas on the system's flexibility. Moreover, the proposed index calculation only requires the regulation capacity distribution and installed capacity of each flexibility unit, without the need for long-term time series simulation calculations, which improves the calculation speed.

[0003] The existing power assessment methods and systems have the following main problems:

[0004] Traditional outlier detection methods often use fixed thresholds or static models to identify abnormal data, but this method cannot flexibly adapt to different scenarios according to changes in data distribution; if the data distribution changes, the fixed threshold may not be able to accurately distinguish between normal data and abnormal data, resulting in incorrect recognition results;

[0005] Relying on manually setting the number of Gaussian distributions may result in the selected number of distributions not being suitable for the actual characteristics of the data; different data sets may require different numbers of Gaussian distributions for accurate modeling, and manual settings may not accurately grasp the complexity of the data, resulting in the model being unable to fully capture the distribution characteristics of the data or oversimplifying the data; failure to use the Bayesian Information Criterion to optimize the complexity of the model may result in inadequate model fit;

[0006] Lack of diversified resource integration may cause the power system to rely on a single or a few energy sources, resulting in inefficient dispatching of the system during peak or trough loads; without the integration of diversified flexible resources, the power system may not be able to fully utilize the advantages of various types of power generation resources;

[0007] Without the support of demand response and load regulation models, the power system may lack the ability to respond flexibly to demand fluctuations, resulting in increased load fluctuations; the power grid will face greater stability challenges, and there may be problems with load peaks being too high or load valleys being unable to meet demand, increasing the complexity of power grid operation; due to the lack of modeling of the power characteristics and operating characteristics of DC interconnection lines, cross-regional power dispatch will be restricted; in the case of uneven power demand, power cannot be allocated quickly and flexibly, resulting in an imbalance in power supply between different regions in the power grid, increasing losses in long-distance power transmission, and affecting the timely supply of electricity; due to the lack of regulation support for energy storage and flexibility resources, it may be necessary to rely on more traditional power equipment for system regulation, increasing the maintenance cost and complexity of the power grid.

[0008] In view of this, the present invention proposes an electric power evaluation method and system based on the participation of multiple flexible resources to solve the above problems. Summary of the invention

[0009] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a power evaluation method based on the participation of multiple flexibility resources, comprising:

[0010] S1. Obtaining heterogeneous data for power assessment;

[0011] S2, preprocessing the acquired power assessment heterogeneous data to obtain a heterogeneous feature data set;

[0012] S3. Construct a multivariate flexibility resource model through heterogeneous feature data sets, and predict the power generation output of different resources through the multivariate flexibility resource model;

[0013] S4. Develop and implement resource scheduling strategies based on the power generation output of different resources.

[0014] Furthermore, the power evaluation heterogeneous data includes power generation side data, energy storage side data, demand side data and grid side data; the power generation side data includes photovoltaic power generation data, wind power generation data, thermal power unit data and hydropower data;

[0015] Photovoltaic power generation data include solar radiation intensity, photovoltaic panel temperature and installation angle; wind power generation data include wind turbine power generation, wind turbine cut-in and cut-out wind speed; thermal power unit data include unit power upper and lower limits and unit start and stop status; hydropower data include water flow rate through the unit and power generation head;

[0016] The data on the energy storage side include the charging and discharging power, efficiency, state of charge of battery energy storage and the upper and lower reservoir capacities of pumped storage; the data on the demand side include the user's electricity load and electricity price data; the data on the grid side include the real-time load of the grid, peak and valley load, frequency fluctuations of the grid, voltage level of the grid and load of transmission lines.

[0017] Furthermore, the method of preprocessing the acquired power evaluation heterogeneous data to obtain a heterogeneous feature data set includes:

[0018] By improving the Gaussian mixture model, outliers existing in the power assessment heterogeneous data are identified, the identified outliers are eliminated, the missing values ​​in the power assessment heterogeneous data are processed, and the duplicate data in the power assessment heterogeneous data are removed; the non-numeric data in the power assessment heterogeneous data is encoded and converted into numerical data to obtain the processed heterogeneous feature data set; the processed heterogeneous feature data set is normalized by standard deviation and converted into a standard normal distribution to obtain a normalized heterogeneous feature data set.

[0019] Furthermore, the method of identifying outliers in the power evaluation heterogeneous data by improving the Gaussian mixture model and removing the identified outliers includes:

[0020] S41, combining the power generation side data, energy storage side data, demand side data and power generation side data in the power evaluation heterogeneous data into a multidimensional feature matrix X = [X1, X2, ..., X m′ ]∈R m′×n′ ; Where m' is the number of samples, that is, the number of data points in the multidimensional feature matrix; n' is the number of feature dimensions, that is, the number of features of each sample;

[0021] S42, presetting the distribution of the power generation side data, the energy storage side data, the demand side data and the power generation side data as a mixture of M′ Gaussian distributions, and expressing them by a Gaussian mixture model, the Gaussian mixture model is: Where p(X) is the probability density value of the data point (X) in all M′ Gaussian mixture models; π j is the weight of the jth Gaussian distribution; η(X|μ j ,∑j) is the probability density function of the jth Gaussian distribution; μ j is the mean vector of the jth Gaussian distribution; ∑j is the covariance matrix of the jth Gaussian distribution; M′ is the number of Gaussian distributions; j is the index of each individual Gaussian distribution component in the Gaussian mixture model;

[0022] S43. Dynamically determine the optimal number of Gaussian distributions by Bayesian information criterion; estimate the parameters of the Gaussian mixture model by expectation maximization algorithm, and randomly initialize the mean vector, covariance matrix and weight of each Gaussian distribution; use the posterior probability calculation formula to calculate the posterior probability that each data point in the multidimensional feature matrix belongs to the jth Gaussian distribution, that is, the probability generated by the jth Gaussian distribution; the posterior probability calculation formula is: Among them, γ ij For data point X iThe probability of belonging to the jth Gaussian distribution; X i is the i-th data point in the multidimensional feature matrix; i is the index of the data point; π k is the weight of the kth Gaussian distribution; k is the index of traversing all Gaussian distributions;

[0023] S44. For each data point X in the multidimensional feature data i , calculate its probability density p(X i ), the preset probability density threshold is δ, and the preset probability density threshold is dynamically adjusted through the probability density threshold adjustment formula. The probability density threshold adjustment formula is: Among them, δ′ is the probability density threshold after dynamic adjustment; k′ is the proportional coefficient; is the average probability density of the data; r is the target outlier;

[0024] S45. If p(X i )<δ′, then determine the data point X i is an outlier; if p(X i )≥δ′, then the data point X i is a normal value, which will satisfy p(X i )<δ′ are eliminated from the heterogeneous data for power evaluation.

[0025] Furthermore, the method for dynamically determining the optimal number of Gaussian distributions by using the Bayesian Information Criterion includes:

[0026] The candidate range of the Gaussian distribution number M' is preset to M'∈[M' min ,M′ max ]; where M′ min is the minimum value of the Gaussian distribution; M′ max is the maximum value of the Gaussian distribution number; for each Gaussian distribution number M′ within the candidate range of the preset Gaussian distribution number M′, the Gaussian mixture model is trained, and the corresponding maximum likelihood value is calculated by the maximum likelihood estimation formula; the maximum likelihood estimation formula is: Where L is the maximum likelihood value corresponding to each Gaussian distribution number M′;

[0027] According to each Gaussian distribution number M′, the Bayesian information criterion value is calculated by the Bayesian information criterion calculation formula; the Bayesian information criterion calculation formula is: BIC=-2ln(L)+w·ln(m′); wherein BIC is the Bayesian information criterion value; ln(L) is the log-likelihood of the Gaussian mixture model; w is the number of Gaussian mixture model parameters, including the mean vector, covariance matrix and weight;

[0028] By comparing the Bayesian Information Criterion values ​​BIC corresponding to different Gaussian distribution numbers M′, the Gaussian distribution number M′ that minimizes the Bayesian Information Criterion value BIC is selected, which is the optimal Gaussian distribution number.

[0029] Further, the method of constructing a multivariate flexibility resource model through heterogeneous feature data sets includes:

[0030] The multivariate flexibility resource model includes photovoltaic power generation output model, wind power generation output model, thermal power unit fuel cost model, hydropower unit output model, pumped storage unit model, IBDR cost calculation model, PSDR cost calculation model, electrochemical energy storage model, DC interconnection line power characteristic model and DC interconnection line operation characteristic model;

[0031] Constructing a photovoltaic power generation power assessment model:

[0032]

[0033] in, is the output power of the photovoltaic system at time t every day; is the rated power of photovoltaic; α t is the actual light intensity at time t; α rated is the rated light intensity of the photovoltaic system; t is a certain time of the day;

[0034] The temperature linear function is used to describe its influence on the photovoltaic power generation output; the photovoltaic panel temperature is preset to be T, the photovoltaic power generation output is P, and the temperature linear function is:

[0035] P=P ref -h×(TT ref );

[0036] Among them, P ref At the reference temperature T ref Photovoltaic power generation output under T ref is the reference temperature; h is a constant, which indicates the rate at which temperature affects photovoltaic power generation output;

[0037] The cosine law is used to calculate the incident angle of sunlight, and the photoelectric conversion efficiency of the photovoltaic panel is considered to estimate the power generation. The output model is:

[0038] P out =P STC ×η cell ×cosθ;

[0039] Among them, P out The contribution of the installation angle and orientation of the solar panel to the photovoltaic power generation output; P STC is the rated output power of the photovoltaic panel; η cellis the conversion efficiency of the photovoltaic cell; θ is the angle between the sunlight and the normal of the photovoltaic panel;

[0040] Taking into account the solar radiation intensity, photovoltaic panel temperature, and the installation angle and orientation of the solar panel, the photovoltaic power generation output model is obtained by using the weighted average method. The photovoltaic power generation output model is:

[0041]

[0042] Among them, f1 is the contribution weight of solar radiation intensity to photovoltaic power generation output; f2 is the contribution weight of photovoltaic panel temperature to photovoltaic power generation output; f3 is the contribution weight of solar panel installation angle and orientation to photovoltaic power generation output.

[0043] Furthermore, the wind power generation output model includes: Among them, P WT is the wind turbine power generation; v is the actual wind speed; v u is the cut-in wind speed; v o is the cut-out wind speed; v N is the rated wind speed; is the rated output power of the fan;

[0044] The output power of thermal power units is subject to upper and lower limits, and its constraint expression is:

[0045] U g,t P g,min ≤P g,t ≤U g,t P g,max ;

[0046] Among them, P g,t is the output power of thermal power unit g at time t in the thermal power unit model; P g,max and P g,mion Respectively represent the upper and lower limits of thermal power unit output g; U g,t is the binary variable of the operating status of thermal power unit g at time t, U g,t =0 means shutdown, U g,t =1 means running;

[0047] The fuel cost model of thermal power units is formed by coal consumption cost and start-up and shutdown cost. The fuel cost model of thermal power units is: Among them, C f is the total fuel cost of thermal power units; C H is the coal consumption cost of thermal power units; is the start-up and shutdown cost of thermal power units;

[0048] The coal consumption cost expression of thermal power units is:

[0049]

[0050] Where T is the total operating time of thermal power unit g; G is the total number of thermal power units; a g , b g and c g are the coal consumption cost coefficients of thermal power unit g; P g,t is the active power of thermal power unit g at time t;

[0051] The start-up and shutdown cost expression of thermal power units is: Among them, S g is the start-up and shutdown cost of thermal power unit g; is the binary variable of the operating status of thermal power unit g at time t-1;

[0052] The hydropower units include run-of-river hydropower stations and pumped storage units; the output model of the hydropower units is:

[0053] P g,l,t =ρg′η g,l Q R,l,t H g,l,t ;

[0054] Among them, P g,l,t is the output power of the lth run-of-river hydropower station at time t; ρ is the density of water; g′ is the acceleration due to gravity; η g,l is the power generation efficiency; Q R,l,t is the power generation flow; H g,l,t The water head for power generation;

[0055] The flow constraint formula is used to constrain the power generation flow of the run-of-river hydropower station. The flow constraint formula is:

[0056] Q R,l,min ≤Q R,l,t ≤Q R,l,max ,;

[0057] Among them, Q R,l,min and Q R,l,max are the maximum and minimum power generation flow of the lth run-of-river hydropower station, l=1,2,…,N R ; N R is the number of run-of-river hydropower stations;

[0058] The pumped storage unit is composed of units in turbine working condition and units in pump working condition. The actual output power of the unit in power generation condition is measured by the turbine working condition output formula. The turbine working condition output formula is:

[0059] U gen,t P gen,min ≤P gen,t ≤U gen,t P gen,max ;

[0060] Among them, U gen,t is the number of units in power generation condition at time t; P gen,t P is the output of the pumped storage unit turbine under working condition at time t; gen,min and P gen,max It is the upper and lower limits of the output of the pumped storage unit under power generation conditions;

[0061] The actual output power of the unit under pumping conditions is measured by the water pump working condition output formula; the water pump working condition output formula is:

[0062] U pump,t P pump,min ≤P pump,t ≤U pump,t P pump,max ;

[0063] Among them, U pump,t is the number of units in pumping condition at time t; P pump,t P is the output of the pumped storage unit under water pump working condition at time t; pump,min and P pump,max It is the upper and lower limits of the output of the pumped storage unit under the working condition of the water pump;

[0064] An IBDR cost calculation model is established based on the changes in the power company's revenue from demand-side users with IBDR capabilities;

[0065] Before users participate in IBDR, the power company's income is:

[0066]

[0067] in, The income of the power company before the user participates in IBDR; t is the fixed electricity price on the retail side at time t; Q IB,n,t is the load of the nth IBDR at time t before the implementation of the demand response project;

[0068] The cost of IBDR is the power outage loss caused by load reduction. When users reduce Q I,n,t When the load is on, the power company's compensation is:

[0069]

[0070] in, Reduce Q for users through IBDR I,n,t When the load is on, the power company compensates it; K 1n and K 2n are the quadratic term and linear term coefficient of the nth IBDR compensation amount; Q I,n,t is the load reduction of the nth IBDR at time t;

[0071] After users participate in IBDR, the power company's income is:

[0072]

[0073] in, The income of the power company after the users participate in IBDR;

[0074] The IBDR cost calculation model is:

[0075]

[0076] Where: a I,n,t =K 1,n ; b I,n,t =K 2,n +p t ;

[0077] Before users participate in PSDR, the power company's revenue is:

[0078]

[0079] Among them, Q PB,m,t is the load of the mth PSDR at time t before the implementation of the demand response project; after the user participates in PSDR, the income of the power company is:

[0080]

[0081] Among them, R is the preferential electricity price ratio given by the power company to users in order to increase their enthusiasm for participation; Δp t Q is the increase in electricity price at time t; P,m,t is the load reduction of the mth PSDR at time t;

[0082] According to the demand price elasticity coefficient ε P,m available:

[0083]

[0084] The PSDR cost calculation model is:

[0085]

[0086] Among them, a P,m,t =Rp t / ε P,m Q PB,m,t ; b P,m,t =Rp t (1-1 / ε P,m );c P,m,t =(1-R)p tQ PB,m,t ;

[0087] The electrochemical energy storage model is:

[0088]

[0089] Among them, θ dis is the discharge efficiency of the energy storage system; θ ch Charging efficiency for energy storage systems; The charging power of the energy storage device at time t; is the discharge power of the energy storage device at time t; is the amount of electricity stored in the energy storage device at time t; is the amount of electricity released by the energy storage device at time t;

[0090] The charging and discharging power constraint formula is used to constrain the charging and discharging power of the energy storage at each moment; the charging and discharging power constraint formula is:

[0091]

[0092] Among them, ρ is a 0-1 variable, ρ is 1 when the energy storage is charged, and ρ is 0 when the energy storage is discharged;

[0093] Each stepped power of the DC tie line is regarded as a feasible state of the DC tie line power. If the DC tie line is in operation, the power of the DC tie line is and can only be in the feasible state set, that is, p d,t ∈{P d,1 ,P d,2 ,...,P d,m ,...,P d,M};

[0094] Among them, p d,t is the active power of DC tie line d at time t; P d,m is the power value of the DC tie line d in feasible state m; M is the number of feasible states of the DC tie line d;

[0095] The power characteristic model of DC tie line is:

[0096]

[0097] ; Among them, p d,t is the power of the DC interconnection line; I d,m,t is a 0-1 variable, indicating whether the DC tie line d is in a feasible state m at time t;

[0098] The operating characteristic model of DC tie line is:

[0099] P d,min u d,t ≤pd,t ≤P d,max u d,t ;

[0100] Among them, P d,min With P d,max are the minimum power and maximum power that can be transmitted by the DC tie line d; u d,t is the operating power of the DC tie line d at time t;

[0101] Based on the data from the power generation side, energy storage side, grid side and demand side, simulation technology analysis is carried out to quantitatively evaluate the impact of different flexibility resources on the total cost by comparing the cost effects of different flexibility resources when participating in power assessment.

[0102] Furthermore, the method of running simulation technology analysis based on power generation side data, energy storage side data, grid side data and demand side data, and quantitatively evaluating the impact of different flexibility resources on total cost by comparing the cost effects of different flexibility resources when participating in power evaluation, includes:

[0103] Taking the minimization of the total cost of power system operation with the participation of multiple flexible resources as the goal, the optimal objective function of economic benefits of the system joint operation model is constructed. The optimal objective function of economic benefits is:

[0104] f=min(C FD +C XQ +C CN +C HJ );

[0105] Among them, C FD , C XQ , C CN and C HJ They represent the power generation cost, demand side cost, energy storage cost and environmental cost respectively;

[0106] The power generation cost is expressed as:

[0107]

[0108] Among them, C HD is the total operating cost of thermal power units; C SD is the total operating cost of the hydropower unit; C CX is the total operating cost of the pumped storage unit; I is the number of thermal power units that have undergone flexibility transformation;

[0109] The demand side cost is expressed as:

[0110]

[0111] The energy storage cost is expressed as:

[0112]

[0113] The environmental cost refers to the carbon emission cost. Carbon emission comes from thermal power generating units. The carbon emission is expressed as:

[0114]

[0115] ; Among them, a1, b1, c1 are the carbon emission calculation parameters of thermal power units;

[0116] Carbon emission quota is expressed as:

[0117]

[0118] Among them, χ e is the carbon emission quota per unit of power generation of thermal power units; the carbon emission cost is expressed as:

[0119] C HJ =λ(CE1-CE2);

[0120] Among them, λ is the price per unit of carbon emissions. By comparing the cost effects of different flexibility resources when participating in electricity assessment, the impact of different flexibility resources on the total cost is quantitatively evaluated.

[0121] Furthermore, the method for formulating and executing resource scheduling strategies according to the power generation output of different resources includes collecting the power generation output data and operating status of different resources in real time through the power assessment intelligent management terminal, and feeding back to the scheduling center; the scheduling center specifies the corresponding resource scheduling strategies according to the collected power generation output data and operating status of different resources, and sends resource adjustment instructions to each generator set to control the operating status of each generator set.

[0122] An electric power assessment system based on the participation of multiple flexibility resources, comprising:

[0123] A data acquisition module, used to acquire heterogeneous data for power evaluation;

[0124] A data processing module is used to pre-process the acquired power evaluation heterogeneous data to obtain a heterogeneous feature data set;

[0125] The multivariate flexibility resource modeling module is used to construct a multivariate flexibility resource model through heterogeneous feature data sets, and predict the power generation output of different resources through the multivariate flexibility resource model;

[0126] The scheduling decision execution module is used to formulate and execute resource scheduling strategies according to the power generation output of different resources; each module is connected by wired and / or wireless means.

[0127] The technical effects and advantages of the power evaluation method and system based on the participation of multiple flexible resources of the present invention are as follows:

[0128] The Gaussian mixture model can efficiently identify data points in the power assessment data that do not conform to the normal distribution. These outliers may be caused by data collection errors, equipment failures, or other external factors. Accurately identifying and eliminating these outliers helps improve data quality and ensure the accuracy of subsequent analysis and decision-making; dynamically determine the optimal number of Gaussian distributions through the Bayesian information criterion, and dynamically adjust the threshold for outlier detection through the probability density threshold adjustment formula. In this way, it can flexibly adapt to the needs of outlier identification in different scenarios according to changes in data distribution, thereby avoiding overly strict or overly loose threshold settings and improving the robustness of the method;

[0129] The Bayesian Information Criterion is used to dynamically select the optimal number of Gaussian distributions without manual setting. The Bayesian Information Criterion takes into account the degree of fit of the model and the complexity of the model, thereby selecting a model that can fit the data well but is not too complex. This makes the entire process more automated and adaptable to different data characteristics;

[0130] By integrating multiple flexible resources such as photovoltaic, wind power, thermal power, hydropower, and energy storage, a comprehensive power evaluation model has been constructed; diversified resource utilization can comprehensively consider the advantages and disadvantages and characteristics of different types of power generation, and provide diversified solutions for the power system, thereby achieving optimal allocation and scheduling of resources; by integrating multiple flexible resources such as wind power, thermal power, hydropower, energy storage, and DC interconnection lines, it is possible to achieve optimal scheduling of power production, ensuring that the power system can meet demand while minimizing the operating cost of the system;

[0131] Through IBDR (demand response) and PSDR (load regulation) models, it is possible to adjust power demand, reduce system load fluctuations, and optimize the operation of the power grid. At the same time, the electrochemical energy storage model helps to regulate power flow between peaks and valleys, provide backup power, and improve the stability and anti-interference ability of the system; by modeling the power characteristics and operating characteristics of DC interconnection lines, it is possible to achieve rapid cross-regional power dispatch, improve the flexibility of power supply, and reduce losses during long-distance transmission; by achieving efficient dispatch and cost optimization of flexible resources, it can help power companies better adapt to market changes, support the government's carbon reduction goals and sustainable energy policies, and quantify the cost effect of flexible resources to provide a basis for policy formulation and market design. BRIEF DESCRIPTION OF THE DRAWINGS

[0132] Figure 1 A schematic diagram of a flow chart of a power assessment method based on the participation of multiple flexibility resources according to the present invention;

[0133] Figure 2 It is a schematic diagram of the structure of an electric power assessment system based on the participation of multiple flexibility resources according to the present invention. DETAILED DESCRIPTION

[0134] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0135] Example 1

[0136] See also Figure 1 As shown, this embodiment provides a power assessment method based on the participation of multiple flexibility resources, including:

[0137] S1. Obtaining heterogeneous data for power assessment;

[0138] S2, preprocessing the acquired power assessment heterogeneous data to obtain a heterogeneous feature data set;

[0139] S3. Construct a multivariate flexibility resource model through heterogeneous feature data sets, and predict the power generation output of different resources through the multivariate flexibility resource model;

[0140] S4. Develop and implement resource scheduling strategies based on the power generation output of different resources.

[0141] Heterogeneous data for power evaluation include data on the power generation side, data on the energy storage side, data on the demand side, and data on the grid side; data on the power generation side include data on photovoltaic power generation, wind power generation, thermal power unit data, and hydropower data;

[0142] Photovoltaic power generation data include solar radiation intensity, photovoltaic panel temperature and installation angle; wind power generation data include wind turbine power generation, wind turbine cut-in and cut-out wind speed; thermal power unit data include unit power upper and lower limits and unit start and stop status; hydropower data include water flow rate through the unit and power generation head;

[0143] The data on the energy storage side include the charging and discharging power, efficiency, state of charge of battery energy storage and the upper and lower reservoir capacities of pumped storage; the data on the demand side include the user's electricity load and electricity price data; the data on the grid side include the real-time load of the grid, peak and valley load, frequency fluctuations of the grid, voltage level of the grid and load of transmission lines.

[0144] The method of preprocessing the acquired power evaluation heterogeneous data to obtain a heterogeneous feature data set includes:

[0145] By improving the Gaussian mixture model, outliers existing in the power assessment heterogeneous data are identified, the identified outliers are eliminated, the missing values ​​in the power assessment heterogeneous data are processed, and the duplicate data in the power assessment heterogeneous data are removed; the non-numeric data in the power assessment heterogeneous data is encoded and converted into numerical data to obtain the processed heterogeneous feature data set; the processed heterogeneous feature data set is normalized by standard deviation and converted into a standard normal distribution to obtain a normalized heterogeneous feature data set.

[0146] The method of identifying outliers in the heterogeneous data of power evaluation by improving the Gaussian mixture model and removing the identified outliers includes:

[0147] S41, combining the power generation side data, energy storage side data, demand side data and power generation side data in the power evaluation heterogeneous data into a multidimensional feature matrix X = [X1, X2, ..., X m′ ]∈R m′×n′ ; Where m' is the number of samples, that is, the number of data points in the multidimensional feature matrix; n' is the number of feature dimensions, that is, the number of features of each sample;

[0148] S42, presetting the distribution of the power generation side data, the energy storage side data, the demand side data and the power generation side data as a mixture of M′ Gaussian distributions, and expressing them by a Gaussian mixture model, the Gaussian mixture model is: Where p(X) is the probability density value of the data point (X) in all M′ Gaussian mixture models; π j is the weight of the jth Gaussian distribution; η(X|μ j ,∑j) is the probability density function of the jth Gaussian distribution; μ j is the mean vector of the jth Gaussian distribution; ∑j is the covariance matrix of the jth Gaussian distribution; M′ is the number of Gaussian distributions; j is the index of each individual Gaussian distribution component in the Gaussian mixture model;

[0149] S43. Dynamically determine the optimal number of Gaussian distributions by Bayesian information criterion; estimate the parameters of the Gaussian mixture model by expectation maximization algorithm, and randomly initialize the mean vector, covariance matrix and weight of each Gaussian distribution; use the posterior probability calculation formula to calculate the posterior probability that each data point in the multidimensional feature matrix belongs to the jth Gaussian distribution, that is, the probability generated by the jth Gaussian distribution; the posterior probability calculation formula is: Among them, γ ij For data point X i The probability of belonging to the jth Gaussian distribution; X i is the i-th data point in the multidimensional feature matrix; i is the index of the data point; π k is the weight of the kth Gaussian distribution; k is the index of traversing all Gaussian distributions;

[0150] S44. For each data point X in the multidimensional feature data i , calculate its probability density p(X i ), the preset probability density threshold is δ, and the preset probability density threshold is dynamically adjusted through the probability density threshold adjustment formula. The probability density threshold adjustment formula is: Among them, δ′ is the probability density threshold after dynamic adjustment; k′ is the proportional coefficient; is the average probability density of the data; r is the target outlier;

[0151] For example, the proportional coefficient is 2, the average probability density of the data is 0.2, the number of feature dimensions is 5, the number of Gaussian distributions is 10, the number of data points in the multidimensional feature matrix is ​​50, the preset probability density threshold is 0.8, and the target anomaly value is 0.01, then the dynamically adjusted probability density threshold is 0.32.

[0152] S45. If p(X i )<δ′, then determine the data point X i is an outlier; if p(X i )≥δ′, then the data point X i is a normal value, which will satisfy p(X i )<δ′ are eliminated from the heterogeneous data for power evaluation.

[0153] Methods for dynamically determining the optimal number of Gaussian distributions using the Bayesian Information Criterion include:

[0154] The candidate range of the Gaussian distribution number M' is preset to M'∈[M' min ,M′ max ]; where M′ min is the minimum value of the Gaussian distribution; M′ max is the maximum value of the Gaussian distribution number; for each Gaussian distribution number M′ within the candidate range of the preset Gaussian distribution number M′, the Gaussian mixture model is trained, and the corresponding maximum likelihood value is calculated by the maximum likelihood estimation formula; the maximum likelihood estimation formula is: Where L is the maximum likelihood value corresponding to each Gaussian distribution number M′;

[0155] According to each Gaussian distribution number M′, the Bayesian information criterion value is calculated by the Bayesian information criterion calculation formula; the Bayesian information criterion calculation formula is: BIC=-2ln(L)+w·ln(m′); wherein BIC is the Bayesian information criterion value; ln(L) is the log-likelihood of the Gaussian mixture model; w is the number of Gaussian mixture model parameters, including the mean vector, covariance matrix and weight;

[0156] By comparing the Bayesian Information Criterion values ​​BIC corresponding to different Gaussian distribution numbers M′, the Gaussian distribution number M′ that minimizes the Bayesian Information Criterion value BIC is selected, which is the optimal Gaussian distribution number.

[0157] Methods for constructing multivariate flexibility resource models through heterogeneous feature datasets include;

[0158] The multivariate flexibility resource model includes photovoltaic power generation output model, wind power generation output model, thermal power unit fuel cost model, hydropower unit output model, pumped storage unit model, IBDR cost calculation model, PSDR cost calculation model, electrochemical energy storage model, DC interconnection line power characteristic model and DC interconnection line operation characteristic model;

[0159] Photovoltaic power generation, as a clean and renewable energy technology, is increasingly popular in application and development around the world. However, its power generation efficiency is highly dependent on solar irradiance and is therefore significantly affected by meteorological conditions. When the sunshine is good, photovoltaic panels can effectively absorb sunlight and improve power generation efficiency; on the contrary, if there is adverse weather such as rainy or cloudy weather, the power generation may be greatly reduced or even lead to a sharp drop in output. This brings great uncertainty to the output of photovoltaic power generation systems.

[0160] Constructing a photovoltaic power generation power assessment model:

[0161]

[0162] in, is the output power of the photovoltaic system at time t every day; is the rated power of photovoltaic; α t is the actual light intensity at time t; α rated is the rated light intensity of the photovoltaic system; t is a certain time of the day;

[0163] The influence of photovoltaic panel temperature on photovoltaic power generation output is studied, and the temperature linear function is used to describe its influence on photovoltaic power generation output; the photovoltaic panel temperature is preset to be T, the photovoltaic power generation output is P, and the temperature linear function is:

[0164] P=P ref -h×(TT ref );

[0165] Among them, P ref At the reference temperature T ref Photovoltaic power generation output under T ref is the reference temperature; h is a constant, which indicates the rate at which temperature affects photovoltaic power generation output;

[0166] Photovoltaic power generation is affected by the installation angle and orientation of the solar panel. The cosine law is used to calculate the incident angle of the sun's rays, and the photoelectric conversion efficiency of the photovoltaic panel is considered to estimate the power generation. The output model is:

[0167] P out =P STC ×η cell ×cosθ;

[0168] Among them, P out The contribution of the installation angle and orientation of the solar panel to the photovoltaic power generation output; P STC is the rated output power of the photovoltaic panel; η cell is the conversion efficiency of the photovoltaic cell; θ is the angle between the sunlight and the normal of the photovoltaic panel, which can be calculated based on the solar altitude angle, solar azimuth angle and photovoltaic panel inclination angle;

[0169] Taking into account the solar radiation intensity, photovoltaic panel temperature, and the installation angle and orientation of the solar panel, the photovoltaic power generation output model is obtained by using the weighted average method. The photovoltaic power generation output model is:

[0170]

[0171] Among them, f1 is the contribution weight of solar radiation intensity to photovoltaic power generation output; f2 is the contribution weight of photovoltaic panel temperature to photovoltaic power generation output; f3 is the contribution weight of solar panel installation angle and orientation to photovoltaic power generation output.

[0172] Research shows that the output power of a wind turbine is mainly related to the rated wind speed, cut-in and cut-out wind speeds, and can be divided into the following three situations: If the actual wind speed is smaller than the cut-in wind speed, or the actual wind speed is larger than the cut-out wind speed, the wind turbine will not work and the output power will be 0; If the actual wind speed is between the cut-in wind speed and the rated wind speed, the output power of the wind turbine is proportional to the actual wind speed; If the actual wind speed is between the rated wind speed and the cut-out wind speed, the output power of the wind turbine is the rated power.

[0173] The wind power generation output model includes: Among them, P WT is the wind turbine power generation; v is the actual wind speed; v u is the cut-in wind speed; v o is the cut-out wind speed; v N is the rated wind speed; is the rated output power of the fan;

[0174] The output power of thermal power units is subject to upper and lower limits, and its constraint expression is:

[0175] U g,t P g,min ≤P g,t≤U g,t P g,max ;

[0176] Among them, P g,t is the output power of thermal power unit g at time t in the thermal power unit model; P g,max and P g,min Respectively represent the upper and lower limits of thermal power unit output g; U g,t is the binary variable of the operating status of thermal power unit g at time t, U g,t =0 means shutdown, U g,t =1 means running;

[0177] The fuel cost model of thermal power units is formed by coal consumption cost and start-up and shutdown cost. The fuel cost model of thermal power units is: Among them, C f is the total fuel cost of thermal power units; C H is the coal consumption cost of thermal power units; is the start-up and shutdown cost of thermal power units;

[0178] The coal consumption cost expression of thermal power units is:

[0179]

[0180] Where T is the total operating time of thermal power unit g; G is the total number of thermal power units; a g , b g and c g are the coal consumption cost coefficients of thermal power unit g; P g,t is the active power of thermal power unit g at time t;

[0181] The start-up and shutdown cost expression of thermal power units is: Among them, S g is the start-up and shutdown cost of thermal power unit g; is the binary variable of the operating status of thermal power unit g at time t-1;

[0182] Hydropower units have the advantages of short start-up and stop time, fast load lifting and lowering speed, large load adjustment range, and strong ability to cope with rapid load changes. They are ideal flexible power sources. They can be started up to full power in a few minutes, and the load adjustment rate reaches more than 30% of the rated capacity, reaching 100% peak adjustment range; run-of-river hydropower stations refer to hydropower stations that do not have regulating reservoirs and directly use the natural flow of rivers to generate electricity; their output is mainly affected by factors such as water flow, unit performance, and working head.

[0183] The hydropower units include run-of-river hydropower stations and pumped storage units; the output model of the hydropower units is:

[0184] P g,l,t =ρg′ηg,l Q R,l,t H g,l,t ;

[0185] Among them, P g,l,t is the output power of the lth run-of-river hydropower station at time t; ρ is the density of water; g′ is the acceleration due to gravity; η g,l is the power generation efficiency; Q R,l,t is the power generation flow; H g,l,t The water head for power generation;

[0186] A pumped storage unit is a special type of hydropower station with operating characteristics similar to those of a conventional hydropower station, and it has two operating modes: turbine operating mode and pump operating mode.

[0187] Physical characteristic model: According to the operating characteristics and modes of the pumped storage unit, the main consideration is to establish its turbine operating condition and pump operating condition operation model, upper and lower reservoir capacity model and operating condition selection constraints.

[0188] The flow constraint formula is used to constrain the power generation flow of the run-of-river hydropower station. The flow constraint formula is:

[0189] Q R,l,min ≤Q r,l,t ≤Q R,l,max ,;

[0190] Among them, Q R,l,min and Q R,l,max are the maximum and minimum power generation flow of the lth run-of-river hydropower station, l=1,2,…,N R ; N R is the number of run-of-river hydropower stations;

[0191] The pumped storage unit is composed of units in turbine working condition and units in pump working condition. The actual output power of the unit in power generation condition is measured by the turbine working condition output formula. The turbine working condition output formula is:

[0192] U gen,t P gen,min ≤P gen,t ≤U gen,t P gen,max ;

[0193] Among them, U gen,t is the number of units in power generation condition at time t; P gen,t P is the output of the pumped storage unit turbine under working condition at time t; gen,min and P gen,max It is the upper and lower limits of the output of the pumped storage unit under power generation conditions;

[0194] The actual output power of the unit under pumping conditions is measured by the water pump working condition output formula; the water pump working condition output formula is:

[0195] U pump,t P pump,min ≤P pump,t ≤U pump,t P pump,max ;

[0196] Among them, U pump,t is the number of units in pumping condition at time t; P pump,t P is the output of the pumped storage unit under water pump working condition at time t; pump,min and P pump,max It is the upper and lower limits of the output of the pumped storage unit under the working condition of the water pump;

[0197] Demand response refers to the response of power users to power prices or incentive signals, which optimizes their own power consumption and the comprehensive optimization of power system resources by changing their normal power consumption patterns. According to a research report by the U.S. Department of Energy, demand response in the competitive power market can be divided into price-based demand response (PSDR) and incentive-based demand response (IBDR) according to different user response methods;

[0198] Based on the change in the power company's revenue from demand-side users with IBDR capabilities, an IBDR cost calculation model is established; the power company can obtain electricity revenue by supplying electricity to demand-side users with IBDR capabilities. When the load is reduced, they will be financially compensated, and at the same time, the power supply load is reduced, and the electricity revenue is also reduced accordingly. Therefore, the cost of IBDR is the change in the power company's revenue before and after the load is reduced.

[0199] Before users participate in IBDR, the power company's income is:

[0200]

[0201] in, The income of the power company before the user participates in IBDR; t is the fixed electricity price on the retail side at time t; Q IB,n,t is the load of the nth IBDR at time t before the implementation of the demand response project;

[0202] The cost of IBDR is the power outage loss caused by load reduction. When users reduce Q I,n,t When the load is on, the power company's compensation is:

[0203]

[0204] in, Reduce Q for users through IBDRI,n,t When the load is on, the power company compensates it; K 1n and K 2n are the quadratic term and linear term coefficient of the nth IBDR compensation amount; Q I,n,t is the load reduction of the nth IBDR at time t;

[0205] After users participate in IBDR, the power company's income is:

[0206]

[0207] in, The income of the power company after the users participate in IBDR;

[0208] The IBDR cost calculation model is:

[0209]

[0210] Where: a I,n,t =K 1,n ; b I,n,t =K 2,n +p t ;

[0211] PSDR means that users can respond to changes in retail electricity prices and adjust their electricity demand accordingly, which can be described by the demand price elasticity coefficient. The load reduction can be obtained in the dispatch optimization, and the actual retail electricity price issued to users is the retail electricity price converted according to the elasticity coefficient. The cost of PSDR is the change in the power company's electricity bill revenue before and after the electricity price changes, which is also the benefit obtained by users participating in the demand response project.

[0212] Before users participate in PSDR, the power company's revenue is:

[0213]

[0214] Among them, Q PB,m,t is the load of the mth PSDR at time t before the implementation of the demand response project; after the user participates in PSDR, the income of the power company is:

[0215]

[0216] Among them, R is the preferential electricity price ratio given by the power company to users in order to increase their enthusiasm for participation; Δp t Q is the increase in electricity price at time t; P,m,t is the load reduction of the mth PSDR at time t; since PSDR users can not only reduce load demand when electricity prices increase, but also increase load demand when electricity prices decrease, Δp t and Q P,m,tMay be negative.

[0217] According to the demand price elasticity coefficient ε P,m available:

[0218]

[0219] The PSDR cost calculation model is:

[0220]

[0221] Among them, a P,m,t =Rp t / ε P,m Q PB,m,t ; b P,m,t =Rp t (1-1 / ε P,m );c P,m,t =(1-R)p t Q PB,m,t ;

[0222] The electrochemical energy storage model is:

[0223]

[0224] Among them, θ dis is the discharge efficiency of the energy storage system; θ ch Charging efficiency for energy storage systems; The charging power of the energy storage device at time t; is the discharge power of the energy storage device at time t; is the amount of electricity stored in the energy storage device at time t; is the amount of electricity released by the energy storage device at time t;

[0225] The charging and discharging power constraint formula is used to constrain the charging and discharging power of the energy storage at each moment; the charging and discharging power constraint formula is:

[0226]

[0227] Among them, ρ is a 0-1 variable, ρ is 1 when the energy storage is charged, and ρ is 0 when the energy storage is discharged;

[0228] Constraints on the number of charge and discharge cycles: When the energy storage battery is working, if it is overcharged and discharged, the energy storage life will drop sharply. This working mode is not economical. Similarly, if it is charged and discharged frequently, the service life of the energy storage will also be reduced. In order to prevent excessive charge and discharge times from causing damage to the battery, it is necessary to limit the number of battery charge and discharge cycles, that is, the daily discharge power is less than the rated capacity of the energy storage.

[0229] High-voltage DC converter equipment is not suitable for frequent adjustment, and it is difficult to coordinate power with the receiving power grid. Therefore, the current operation mode of high-voltage DC interconnection lines mostly adopts a stepped segmented operation mode, that is, the transmission power is relatively stable in each section. The fluctuation of the new energy output of the sending power grid is smoothed by the thermal power units bundled with it for external transmission to ensure the relative stability of the DC transmission power. The difficulty of the DC interconnection line model lies in the stepped operation of its exchange power. In order to model this operation characteristic more accurately, each stepped power of the DC interconnection line is regarded as a feasible state of the DC interconnection line power. If the DC interconnection line is in operation, the power of the DC interconnection line is and can only be concentrated in this feasible state.

[0230] Each stepped power of the DC tie line is regarded as a feasible state of the DC tie line power. If the DC tie line is in operation, the power of the DC tie line is and can only be in the feasible state set, that is, p d,t ∈{P d,1 ,P d,2 ,...,P d,m ,...,P d,M};

[0231] Among them, p d,t is the active power of DC tie line d at time t; P d,m is the power value of the DC tie line d in feasible state m; M is the number of feasible states of the DC tie line d;

[0232] The power characteristic model of DC tie line is:

[0233]

[0234] ; Among them, p d,t is the power of the DC interconnection line; I d,m,t is a 0-1 variable, indicating whether the DC tie line d is in a feasible state m at time t;

[0235] The operating characteristic model of DC tie line is:

[0236] P d,min u d,t ≤p d,t ≤P d,max u d,t ;

[0237] Among them, P d,min With P d,max are the minimum power and maximum power that can be transmitted by the DC tie line d; u d,t is the operating power of the DC tie line d at time t;

[0238] The feasible step power of the DC tie line during operation is related to its operating characteristics and dispatching mode, and generally does not change for a long period of time. The upper and lower limits of the DC tie line power are affected by factors such as system load conditions, the status of the sending-end power grid units, the receiving-end power grid acceptance capacity, and the operation of the DC tie line. They will change frequently and may even change within a day. After introducing the start and stop status of the DC tie line, only the upper and lower limits of the DC power can be modified without changing the feasible step power, which improves the flexibility and engineering adaptability of the model.

[0239] Based on the data from the power generation side, energy storage side, grid side and demand side, simulation technology analysis is carried out to quantitatively evaluate the impact of different flexibility resources on the total cost by comparing the cost effects of different flexibility resources when participating in power assessment.

[0240] Based on the data from the power generation side, energy storage side, grid side and demand side, simulation technology analysis is performed to compare the cost effects of different flexibility resources when participating in power assessment. The methods for quantitatively evaluating the impact of different flexibility resources on total cost include:

[0241] Given that the unit cost of renewable energy generation is relatively low, its large-scale access can reduce dependence on traditional energy resources by increasing the proportion of renewable energy consumption, thereby reducing overall fuel consumption and energy costs. However, when a variety of flexible resources are included in the scope of power system dispatch, although this enhances the flexibility of the system, it may also lead to an increase in the complexity of power system operation and additional costs such as adaptive transformation of units. The objective function described above is used as an analysis benchmark. By comparing and analyzing the cost effects of different flexible resources participating in the operation of the power system, the impact of these resources on the total cost of the system is quantitatively evaluated, thereby scientifically measuring the economic performance of the power system under the condition of large-scale access to renewable energy.

[0242] Taking the minimization of the total cost of power system operation with the participation of multiple flexible resources as the goal, the optimal objective function of economic benefits of the system joint operation model is constructed. The optimal objective function of economic benefits is:

[0243] f=min(C FD +C XQ +C CN +C HJ );

[0244] Among them, C FD , C XQ , C CN and C HJ They represent the power generation cost, demand side cost, energy storage cost and environmental cost respectively;

[0245] The power generation cost is expressed as:

[0246]

[0247] Among them, C HD is the total operating cost of thermal power units; C SD is the total operating cost of the hydropower unit; C CX is the total operating cost of the pumped storage unit; I is the number of thermal power units that have undergone flexibility transformation;

[0248] The demand side cost is expressed as:

[0249]

[0250] The energy storage cost is expressed as:

[0251]

[0252] The environmental cost refers to the carbon emission cost. Carbon emission comes from thermal power generating units. The carbon emission is expressed as:

[0253]

[0254] ; Among them, a1, b1, c1 are the carbon emission calculation parameters of thermal power units;

[0255] Carbon emission quota is expressed as:

[0256]

[0257] Among them, χ e is the carbon emission quota per unit of power generation of thermal power units; the carbon emission cost is expressed as:

[0258] C HJ =λ(CE1-CE2);

[0259] Among them, λ is the price per unit of carbon emissions. By comparing the cost effects of different flexibility resources when participating in electricity assessment, the impact of different flexibility resources on the total cost is quantitatively evaluated.

[0260] The carbon trading mechanism is a market mechanism that uses CO2 emission rights as commodities for free trading, thereby achieving low-carbon environmental protection goals. It uses economic means to promote the power industry towards a low-carbon economy, and completes the buyer's greenhouse gas reduction task by purchasing the specified carbon emission quota from the seller through the purchase of carbon emission rights. On the premise of increasing the economic benefits obtained by the seller, the buyer and the seller can achieve a win-win situation. As a quota trading mechanism, the carbon trading mechanism is carried out through the purchase and sale of carbon emission rights. The state has a certain limit on the total amount of carbon emission rights, and issues a certain amount of carbon emission rights to certain power generation companies based on analysis. Power generation companies that exceed the limit will be fined according to the degree of excess to achieve energy conservation and emission reduction. The carbon trading market born on this basis gives power generation companies that exceed the carbon emission limit a certain opportunity to purchase carbon emission rights to minimize fines and minimize operating costs.

[0261] The method of formulating and executing resource dispatching strategies according to the power generation output of different resources includes collecting the power generation output data and operating status of different resources in real time through the power assessment intelligent management terminal, and feeding back to the dispatching center; the dispatching center specifies the corresponding resource dispatching strategies according to the collected power generation output data and operating status of different resources, and sends the resource adjustment instructions to each generator set to control the operating status of each generator set.

[0262] In this embodiment, the Gaussian mixture model can efficiently identify data points in the power assessment data that do not conform to the normal distribution. These outliers may be caused by data collection errors, equipment failures, or other external factors. Accurately identifying and eliminating these outliers helps to improve data quality and ensure the accuracy of subsequent analysis and decision-making; dynamically determine the optimal number of Gaussian distributions through the Bayesian information criterion, and dynamically adjust the threshold for outlier detection through the probability density threshold adjustment formula. In this way, it can flexibly adapt to the needs of outlier identification in different scenarios according to changes in data distribution, thereby avoiding overly strict or overly loose threshold settings and improving the robustness of the method;

[0263] The Bayesian Information Criterion is used to dynamically select the optimal number of Gaussian distributions without manual setting. The Bayesian Information Criterion takes into account the degree of fit of the model and the complexity of the model, thereby selecting a model that can fit the data well but is not too complex. This makes the entire process more automated and adaptable to different data characteristics;

[0264] By integrating multiple flexible resources such as photovoltaic, wind power, thermal power, hydropower, energy storage, etc., a comprehensive power evaluation model has been constructed. This diversified resource utilization can comprehensively consider the advantages and disadvantages and characteristics of different types of power generation, and provide a variety of solutions for the power system, so as to achieve the optimal allocation and scheduling of resources; by integrating multiple flexible resources such as wind power, thermal power, hydropower, energy storage, DC interconnection lines, etc., it is possible to achieve optimal scheduling of power production, ensuring that the power system can meet demand while minimizing the operating cost of the system;

[0265] Through IBDR (demand response) and PSDR (load regulation) models, it is possible to adjust power demand, reduce system load fluctuations, and optimize the operation of the power grid. At the same time, the electrochemical energy storage model helps to regulate power flow between peaks and valleys, provide backup power, and improve the stability and anti-interference ability of the system; by modeling the power characteristics and operating characteristics of DC interconnection lines, it is possible to achieve rapid cross-regional power dispatch, improve the flexibility of power supply, and reduce losses during long-distance transmission; by achieving efficient dispatch and cost optimization of flexible resources, it can help power companies better adapt to market changes, support the government's carbon reduction goals and sustainable energy policies, and quantify the cost effect of flexible resources to provide a basis for policy formulation and market design.

[0266] Example 2

[0267] See also Figure 2 As shown, this embodiment provides a power evaluation system based on the participation of multiple flexibility resources, including:

[0268] A data acquisition module, used to acquire heterogeneous data for power evaluation;

[0269] A data processing module is used to pre-process the acquired power evaluation heterogeneous data to obtain a heterogeneous feature data set;

[0270] The multivariate flexibility resource modeling module is used to construct a multivariate flexibility resource model through heterogeneous feature data sets, and predict the power generation output of different resources through the multivariate flexibility resource model;

[0271] The scheduling decision execution module is used to formulate and execute resource scheduling strategies according to the power generation output of different resources; each module is connected by wired and / or wireless means.

[0272] Since the electronic device introduced in this embodiment is an electronic device used to implement a method and system for electric power assessment based on the participation of multiple flexible resources in the embodiment of this application, based on the method and system for electric power assessment based on the participation of multiple flexible resources introduced in the embodiment of this application, the technical personnel in this field can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application will not be introduced in detail here. As long as the technical personnel in this field implement the electronic device used in the method and system for electric power assessment based on the participation of multiple flexible resources in the embodiment of this application, it belongs to the scope of protection of this application.

[0273] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0274] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A power assessment method based on the participation of multiple flexibility resources, characterized in that: include: S1. Obtaining heterogeneous data for power assessment; S2, preprocessing the acquired power assessment heterogeneous data to obtain a heterogeneous feature data set; S3. Construct a multivariate flexibility resource model through heterogeneous feature data sets, and predict the power generation output of different resources through the multivariate flexibility resource model; S4. Develop and implement resource scheduling strategies based on the power generation output of different resources.

2. According to claim 1, a power assessment method based on the participation of multiple flexibility resources is characterized in that: The heterogeneous data for power evaluation include power generation side data, energy storage side data, demand side data and grid side data; the power generation side data include photovoltaic power generation data, wind power generation data, thermal power unit data and hydropower data; Photovoltaic power generation data include solar radiation intensity, photovoltaic panel temperature and installation angle; wind power generation data include wind turbine power generation, wind turbine cut-in and cut-out wind speed; thermal power unit data include unit power upper and lower limits and unit start and stop status; hydropower data include water flow rate through the unit and power generation head; The data on the energy storage side include the charging and discharging power, efficiency, state of charge of battery energy storage and the upper and lower reservoir capacities of pumped storage; the data on the demand side include the user's electricity load and electricity price data; the data on the grid side include the real-time load of the grid, peak and valley load, frequency fluctuations of the grid, voltage level of the grid and load of transmission lines.

3. The power evaluation method based on the participation of multiple flexibility resources according to claim 2 is characterized in that: The method of preprocessing the acquired power evaluation heterogeneous data to obtain a heterogeneous feature data set includes: By improving the Gaussian mixture model, outliers existing in the power assessment heterogeneous data are identified, the identified outliers are eliminated, the missing values ​​in the power assessment heterogeneous data are processed, and the duplicate data in the power assessment heterogeneous data are removed; the non-numeric data in the power assessment heterogeneous data is encoded and converted into numerical data to obtain the processed heterogeneous feature data set; the processed heterogeneous feature data set is normalized by standard deviation and converted into a standard normal distribution to obtain a normalized heterogeneous feature data set.

4. The power evaluation method based on the participation of multiple flexibility resources according to claim 3 is characterized in that: The method of identifying outliers in the power evaluation heterogeneous data by improving the Gaussian mixture model and removing the identified outliers includes: S41, combining the power generation side data, energy storage side data, demand side data and power generation side data in the power evaluation heterogeneous data into a multidimensional feature matrix X = [X1, X2, ..., X m′ ]∈R m′×n′ ; Where m' is the number of samples, that is, the number of data points in the multidimensional feature matrix; n' is the number of feature dimensions, that is, the number of features of each sample; S42, presetting the distribution of the power generation side data, the energy storage side data, the demand side data and the power generation side data as a mixture of M′ Gaussian distributions, and expressing them by a Gaussian mixture model, the Gaussian mixture model is: Where p(X) is the probability density value of the data point (X) in all M′ Gaussian mixture models; π j is the weight of the jth Gaussian distribution; η(X|μ j ,∑j) is the probability density function of the jth Gaussian distribution; μ j is the mean vector of the jth Gaussian distribution; ∑j is the covariance matrix of the jth Gaussian distribution; M′ is the number of Gaussian distributions; j is the index of each individual Gaussian distribution component in the Gaussian mixture model; S43. Dynamically determine the optimal number of Gaussian distributions by Bayesian information criterion; estimate the parameters of the Gaussian mixture model by expectation maximization algorithm, and randomly initialize the mean vector, covariance matrix and weight of each Gaussian distribution; use the posterior probability calculation formula to calculate the posterior probability that each data point in the multidimensional feature matrix belongs to the jth Gaussian distribution, that is, the probability generated by the jth Gaussian distribution; the posterior probability calculation formula is: Among them, γ ij For data point X i The probability of belonging to the jth Gaussian distribution; X i is the i-th data point in the multidimensional feature matrix; i is the index of the data point; π k is the weight of the kth Gaussian distribution; k is the index of traversing all Gaussian distributions; S44. For each data point X in the multidimensional feature data i , calculate its probability density p(X i ), the preset probability density threshold is δ, and the preset probability density threshold is dynamically adjusted through the probability density threshold adjustment formula. The probability density threshold adjustment formula is: Among them, δ′ is the probability density threshold after dynamic adjustment; k′ is the proportional coefficient; is the average probability density of the data; r is the target outlier; S45. If p(X i )<δ′, then the data point X i is an outlier; if p(X i )≥δ′, then the data point X i is a normal value, which will satisfy p(X i )<δ′ are eliminated from the heterogeneous data for power evaluation.

5. The power evaluation method based on the participation of multiple flexibility resources according to claim 4 is characterized in that: The method for dynamically determining the optimal number of Gaussian distributions by Bayesian information criterion includes: The candidate range of the Gaussian distribution number M' is preset to M'∈[M' min ,M′ max ]; where M′ min is the minimum value of the Gaussian distribution; M′ max is the maximum value of the Gaussian distribution number; for each Gaussian distribution number M′ within the candidate range of the preset Gaussian distribution number M′, the Gaussian mixture model is trained, and the corresponding maximum likelihood value is calculated by the maximum likelihood estimation formula; the maximum likelihood estimation formula is: Where L is the maximum likelihood value corresponding to each Gaussian distribution number M′; According to each Gaussian distribution number M′, the Bayesian information criterion value is calculated by the Bayesian information criterion calculation formula; the Bayesian information criterion calculation formula is: BIC=-2ln(L)+w·ln(m′); wherein BIC is the Bayesian information criterion value; ln(L) is the log-likelihood of the Gaussian mixture model; w is the number of Gaussian mixture model parameters, including the mean vector, covariance matrix and weight; By comparing the Bayesian Information Criterion values ​​BIC corresponding to different Gaussian distribution numbers M′, the Gaussian distribution number M′ that minimizes the Bayesian Information Criterion value BIC is selected, which is the optimal Gaussian distribution number.

6. The power evaluation method based on the participation of multiple flexibility resources according to claim 5 is characterized in that: The method for constructing a multivariate flexibility resource model through heterogeneous feature data sets includes: The multivariate flexibility resource model includes photovoltaic power generation output model, wind power generation output model, thermal power unit fuel cost model, hydropower unit output model, pumped storage unit model, IBDR cost calculation model, PSDR cost calculation model, electrochemical energy storage model, DC interconnection line power characteristic model and DC interconnection line operation characteristic model; Constructing photovoltaic power generation power assessment model: Among them, P t PV is the output power of the photovoltaic system at time t every day; is the rated power of photovoltaic; α t is the actual light intensity at time t; α rated is the rated light intensity of the photovoltaic system; t is a certain time of the day; The temperature linear function is used to describe its influence on the photovoltaic power generation output; the photovoltaic panel temperature is preset to be T, the photovoltaic power generation output is P, and the temperature linear function is: P=P ref -h×(T-T ref ); Among them, P ref At the reference temperature T ref Photovoltaic power generation output under T ref is the reference temperature; h is a constant, which indicates the rate at which temperature affects photovoltaic power generation output; The cosine law is used to calculate the incident angle of sunlight, and the photoelectric conversion efficiency of the photovoltaic panel is considered to estimate the power generation. The output model is: P out =P STC ×η cell ×cosθ; Among them, P out The contribution of the installation angle and orientation of the solar panel to the photovoltaic power generation output; P STC is the rated output power of the photovoltaic panel; η cell is the conversion efficiency of the photovoltaic cell; θ is the angle between the sunlight and the normal of the photovoltaic panel; Taking into account the solar radiation intensity, photovoltaic panel temperature, and the installation angle and orientation of the solar panel, the photovoltaic power generation output model is obtained by using the weighted average method. The photovoltaic power generation output model is: Among them, f1 is the contribution weight of solar radiation intensity to photovoltaic power generation output; f2 is the contribution weight of photovoltaic panel temperature to photovoltaic power generation output; f3 is the contribution weight of solar panel installation angle and orientation to photovoltaic power generation output.

7. The power evaluation method based on the participation of multiple flexibility resources according to claim 6 is characterized in that: The wind power generation output model includes: Among them, P WT is the wind turbine power generation; v is the actual wind speed; v u is the cut-in wind speed; v o is the cut-out wind speed; v N is the rated wind speed; is the rated output power of the fan; The output power of thermal power units is subject to upper and lower limits, and its constraint expression is: U g,t Q g,min ≤P g,t ≤U g,t Q g,max ; Among them, P g,t is the output power of thermal power unit g at time t in the thermal power unit model; P g,max and P g,min Respectively represent the upper and lower limits of thermal power unit output g; U g,t is the binary variable of the operating status of thermal power unit g at time t, U g,t =0 means shutdown, U g,t =1 means running; The fuel cost model of thermal power units is formed by coal consumption cost and start-up and shutdown cost. The fuel cost model of thermal power units is: Among them, C f is the total fuel cost of thermal power units; C H is the coal consumption cost of thermal power units; is the start-up and shutdown cost of thermal power units; The coal consumption cost expression of thermal power units is: Where T is the total operating time of thermal power unit g; G is the total number of thermal power units; a g 、b g and c g are the coal consumption cost coefficients of thermal power unit g; P g,t is the active power of thermal power unit g at time t; The start-up and shutdown cost expression of thermal power units is: Among them, S g is the start-up and shutdown cost of thermal power unit g; is the binary variable of the operating status of thermal power unit g at time t-1; The hydropower units include run-of-river hydropower stations and pumped storage units; the output model of the hydropower units is: P g,l,t =ρg′η g,l Q R,l,t H g,l,t ; Among them, P g,l,t is the output power of the lth run-of-river hydropower station at time t; ρ is the density of water; g′ is the acceleration due to gravity; η g,l is the power generation efficiency; Q R,l,t is the power generation flow; H g,l,t The water head for power generation; The flow constraint formula is used to constrain the power generation flow of the run-of-river hydropower station. The flow constraint formula is: Q R,l,min ≤Q R,l,t ≤Q R,l,max ,; Among them, Q R,l,min and Q R,l,max are the maximum and minimum power generation flow of the lth run-of-river hydropower station, l=1,2,…,N R ; N R is the number of run-of-river hydropower stations; The pumped storage unit is composed of units in turbine working condition and units in pump working condition. The actual output power of the unit in power generation condition is measured by the turbine working condition output formula. The turbine working condition output formula is: U gen,t Q gen,min ≤P gen,t ≤U gen,t Q gen,max ; Among them, U gen,t is the number of units in power generation condition at time t; P gen,t P is the output of the pumped storage unit turbine under working condition at time t; gen,min and P gen,max It is the upper and lower limits of the output of the pumped storage unit under power generation conditions; The actual output power of the unit under pumping conditions is measured by the water pump working condition output formula; the water pump working condition output formula is: U pump,t Q pump,min ≤P pump,t ≤U pump,t Q pump,max ; Among them, U pump,t is the number of units in pumping condition at time t; P pump,t P is the output of the pumped storage unit under water pump working condition at time t; pump,min and P pump,max It is the upper and lower limits of the output of the pumped storage unit under the working condition of the water pump; An IBDR cost calculation model is established based on the changes in the power company's revenue from demand-side users with IBDR capabilities; Before users participate in IBDR, the power company's income is: in, The income of the power company before the user participates in IBDR; t is the fixed electricity price at the retail side at time t; Q IB,n,t is the load of the nth IBDR at time t before the implementation of the demand response project; The cost of IBDR is the power outage loss caused by load reduction. When users reduce Q I,n,t When the load is on, the power company's compensation is: in, Reduce Q for users through IBDR I,n,t When the load is on, the power company compensates it; K 1n and K 2n are the quadratic term and linear term coefficient of the nth IBDR compensation amount; Q I,n,t is the load reduction of the nth IBDR at time t; After users participate in IBDR, the power company's income is: in, The income of the power company after the users participate in IBDR; The IBDR cost calculation model is: Where: a I,n,t =K 1,n ; b I,n,t =K 2,n +p t ; Before users participate in PSDR, the power company's revenue is: Among them, Q PB,m,t is the load of the mth PSDR at time t before the implementation of the demand response project; after the user participates in PSDR, the income of the power company is: Among them, R is the preferential electricity price ratio given by the power company to users in order to increase their enthusiasm for participation; Δp t Q is the increase in electricity price at time t; P,m,t is the load reduction of the mth PSDR at time t; According to the demand price elasticity coefficient ε P,m available: The PSDR cost calculation model is: where a P,m,t = Rp t / ε P,m Q PB,m,t ; b P,m,t = Rp t (1 - 1 / ε P,m ); c P,m,t = (1 - R)p t Q PB,m,t ; The electrochemical energy storage model is: Among them, θ dis is the discharge efficiency of the energy storage system; θ ch Charging efficiency for energy storage systems; The charging power of the energy storage device at time t; is the discharge power of the energy storage device at time t; is the amount of electricity stored in the energy storage device at time t; is the amount of electricity released by the energy storage device at time t; The charging and discharging power constraint formula is used to constrain the charging and discharging power of the energy storage at each moment; the charging and discharging power constraint formula is: Among them, ρ is a 0-1 variable, ρ is 1 when the energy storage is charged, and ρ is 0 when the energy storage is discharged; Each stepped power of the DC tie line is regarded as a feasible state of the DC tie line power. If the DC tie line is in operation, the power of the DC tie line is and can only be in the feasible state set, that is, Among them, p d,t is the active power of DC tie line d at time t; P d,m is the power value of the DC tie line d in feasible state m; M is the number of feasible states of the DC tie line d; The power characteristic model of DC tie line is: ； Among them, p d,t is the power of the DC interconnection line; I d,m,t is a 0-1 variable, indicating whether the DC tie line d is in a feasible state m at time t; The operating characteristic model of DC tie line is: P d,min in d,t ≤p d,t ≤P d,max in d,t ; Among them, P d,min With P d,max are the minimum power and maximum power that can be transmitted by the DC tie line d; u d,t is the operating power of the DC tie line d at time t; Based on the data from the power generation side, energy storage side, grid side and demand side, simulation technology analysis is carried out to quantitatively evaluate the impact of different flexibility resources on the total cost by comparing the cost effects of different flexibility resources when participating in power assessment.

8. The power evaluation method based on the participation of multiple flexibility resources according to claim 7 is characterized in that: The method of running simulation technology analysis based on power generation side data, energy storage side data, grid side data and demand side data, and quantitatively evaluating the impact of different flexibility resources on total cost by comparing the cost effects of different flexibility resources when participating in power evaluation, includes: Taking the minimization of the total cost of power system operation with the participation of multiple flexible resources as the goal, the optimal objective function of economic benefits of the system joint operation model is constructed. The optimal objective function of economic benefits is: f=min(C FD +C XQ +C CN +C HJ ); Among them, C FD , C XQ , C CN and C HJ They represent the power generation cost, demand side cost, energy storage cost and environmental cost respectively; The power generation cost is expressed as: Among them, C HD is the total operating cost of thermal power units; C SD is the total operating cost of the hydropower unit; C CX is the total operating cost of the pumped storage unit; I is the number of thermal power units that have undergone flexibility transformation; The demand side cost is expressed as: The energy storage cost is expressed as: The environmental cost refers to the carbon emission cost. Carbon emission comes from thermal power generating units. The carbon emission is expressed as: Among them, a1, b1, and c1 are the carbon emission calculation parameters of thermal power units; Carbon emission quota is expressed as: Among them, χ e is the carbon emission quota per unit of power generation of thermal power units; the carbon emission cost is expressed as: C HJ =λ(CE1-CE2); Among them, λ is the price per unit of carbon emissions. By comparing the cost effects of different flexibility resources when participating in electricity assessment, the impact of different flexibility resources on the total cost is quantitatively evaluated.

9. The power evaluation method based on the participation of multiple flexibility resources according to claim 8 is characterized in that: The method for formulating and executing resource scheduling strategies according to the power generation output of different resources includes collecting the power generation output data and operating status of different resources in real time through the power evaluation intelligent management terminal, and feeding back to the scheduling center; the scheduling center specifies the corresponding resource scheduling strategies according to the collected power generation output data and operating status of different resources, and sends resource adjustment instructions to each generator set to control the operating status of each generator set.

10. An electric power evaluation system based on the participation of multiple flexible resources, used to implement an electric power evaluation method based on the participation of multiple flexible resources as claimed in any one of claims 1 to 9, characterized in that: include: A data acquisition module, used to acquire heterogeneous data for power evaluation; A data processing module is used to pre-process the acquired power evaluation heterogeneous data to obtain a heterogeneous feature data set; The multivariate flexibility resource modeling module is used to construct a multivariate flexibility resource model through heterogeneous feature data sets, and predict the power generation output of different resources through the multivariate flexibility resource model; The scheduling decision execution module is used to formulate and execute resource scheduling strategies according to the power generation output of different resources; each module is connected by wired and / or wireless means.

Images