Wind-light and load typical scene optimization method

Abstract

The invention relates to the technical field of power systems, in particular to a wind-light and load typical scene optimization method. The method comprises the following steps: establishing a mathematical model of core equipment in the multi-energy coupling system; on the basis of the mathematical model, an improved Gaussian mixture model and an improved expectation maximization algorithm are adopted, and a typical scene set with probability distribution fitting reality is generated in combination with space-time correlation characteristics of wind and light output and loads; the method comprises the following steps of: constructing an optimal scheduling model containing multi-energy supply and demand balance constraints by taking system total cost minimization as a target and combining an operation boundary condition and a typical scene set of core equipment, and solving the optimal scheduling model after performing linearization processing on the optimal scheduling model to obtain a flexibility adjustment scheme of the multi-energy coupling system. According to the method, the influence of multiple uncertain factors on the system regulation capability can be comprehensively reflected, the modeling precision of the probability distribution of uncertain variables such as wind, light, load and the like is improved, and the accuracy of the flexibility evaluation of the power system and the rationality of resource allocation are improved.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to an optimization method for typical wind, solar and load scenarios. Background Technology

[0002] Driven by the "dual carbon" goals, large-scale grid connection of highly volatile renewable energy sources such as wind power and photovoltaics has significantly increased the uncertainty and volatility faced by the power system, placing higher demands on the system's flexible operation capabilities. Ensuring the safe and stable operation of the power system requires not only achieving real-time power balance across dispatching periods but also possessing sufficient cross-period power regulation capabilities to cope with random fluctuations in load and renewable energy output. The scientific quantification and assessment of operational flexibility has become a core issue.

[0003] Currently, power system dispatch mainly adopts traditional models, relying heavily on deterministic power balance and fixed reserve configuration strategies, and responding to fluctuations by pre-setting fixed reserve capacity. In terms of demand response, it mainly relies on the independent application of single price-based or incentive-based mechanisms, and has not yet formed a multi-mechanism coordinated load control model, and there is insufficient overall integration of various flexibility resources.

[0004] Existing technologies have significant limitations: they struggle to characterize flexibility risks in a refined and probabilistic manner, lack sufficient description of the spatiotemporal correlation characteristics of new energy output and load forecasting errors, and fail to fully reflect the impact of multiple uncertainties on system regulation capabilities. This leads to biases in flexibility assessment results, easily causing operational risks due to insufficient allocation of flexibility resources, or increased operating costs due to over-allocation, making it difficult to adapt to the system's operational needs after a high proportion of renewable energy is integrated. Summary of the Invention

[0005] This invention provides an optimization method for typical wind and solar load scenarios to address the problems in existing technologies where the inability to fully reflect the impact of multiple uncertainties on system regulation capabilities leads to distorted flexibility assessments and inappropriate resource allocation.

[0006] In a first aspect, embodiments of the present invention provide an optimization method for typical scenarios of wind, solar, and load, including: Establish mathematical models of the core equipment in the multi-energy coupling system to characterize the output characteristics, energy conversion efficiency and operating boundary conditions of each equipment; Based on the mathematical model, an improved Gaussian mixture model and an improved expectation-maximization algorithm are used, combined with the spatiotemporal correlation characteristics of wind and solar power output and load, to generate a set of typical scenarios with probability distributions that fit reality. With the goal of minimizing the total system cost, and in conjunction with the operating boundary conditions of the core equipment and the typical scenario set, an optimized scheduling model containing multi-energy supply and demand balance constraints is constructed. After linearizing the optimized scheduling model, the solution is obtained to obtain a flexible adjustment scheme for the multi-energy coupled system.

[0007] In a second aspect, embodiments of the present invention provide a terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the wind and solar load typical scenario optimization method as described in the first aspect or any possible implementation of the first aspect.

[0008] This invention provides a method for optimizing typical scenarios of wind and solar power and load. It establishes a mathematical model of the core equipment in a multi-energy coupled system, characterizing the output characteristics, energy conversion efficiency, and operational boundary conditions of each device. Based on this mathematical model, an improved Gaussian mixture model and an improved expectation-maximization algorithm are used, combined with the spatiotemporal correlation characteristics of wind and solar power output and load, to generate a set of typical scenarios with a probability distribution that closely matches reality. With the goal of minimizing the total system cost, and combining the operational boundary conditions of the core equipment and the set of typical scenarios, an optimized scheduling model containing multi-energy supply and demand balance constraints is constructed. After linearizing the optimized scheduling model, it is solved to obtain a flexible adjustment scheme for the multi-energy coupled system. This invention breaks through the assumption of mutual independence of latent variables in traditional GMM. By introducing the concept of "mutual assistance of latent variables," it cleverly characterizes the temporal correlation between data points at adjacent time points. In the E-step of the improved EM algorithm, the calculation method of responsibility degree is reconstructed using conditional probability, enabling it to integrate historical latent variable information. This allows it to capture the dynamic evolution of uncertain data, comprehensively reflect the impact of multiple uncertain factors on the system's regulation capability, improve the modeling accuracy of the probability distribution of uncertain variables such as wind, solar, and load, enhance the accuracy of power system flexibility assessment, and improve the rationality of resource allocation. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating the implementation of the optimization method for typical scenarios of wind, solar and load provided in this embodiment of the invention. Figure 2 This is a flowchart illustrating the implementation of the method for obtaining the core parameters of a latent variable mutually supportive Gaussian mixture model provided in this embodiment of the invention. Figure 3 This is a flowchart illustrating the implementation of a method for obtaining the core parameters of a latent variable mutually supportive Gaussian mixture model, provided in another embodiment of the present invention. Figure 4 This is a schematic diagram of the iterative process of the improved EM algorithm provided in the embodiments of the present invention; Figure 5 This is a schematic diagram of the terminal provided in an embodiment of the present invention. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0012] To make the objectives, technical solutions, and advantages of the present invention clearer, specific embodiments will be described below in conjunction with the accompanying drawings.

[0013] Figure 1 The implementation flowchart of a typical scenario optimization method for wind and solar load provided by an embodiment of the present invention is described in detail below: Step 101: Establish mathematical models of the core equipment in the multi-energy coupling system to characterize the output characteristics, energy conversion efficiency and operating boundary conditions of each equipment.

[0014] In this embodiment, the core equipment may include power generation equipment, heating and cooling equipment, combined energy supply equipment, energy storage equipment, and flexible load and demand response equipment.

[0015] The mathematical models of power generation equipment include photovoltaic models and wind power models. The photovoltaic model considers the influence of solar radiation intensity and ambient temperature on photovoltaic output power, while the wind power model considers the influence of cut-in wind speed, rated wind speed and cut-out wind speed on wind power output power. Photovoltaic power generation is a technology that utilizes the photovoltaic effect of semiconductor materials to directly convert solar radiation energy into electrical energy, and it has become an important pillar of the clean energy system. Against the backdrop of continuously deepening dual-carbon goals, the power supply structure is undergoing a transformation from being dominated by thermal power to being dominated by new energy sources. Among these, photovoltaic power generation, with its clean and renewable characteristics, is gradually becoming a key component of the future power supply structure. However, photovoltaic output is affected by various factors such as solar irradiance and ambient temperature, exhibiting strong volatility. A photovoltaic model can be expressed as: ; in, Indicates time period The output power of the photovoltaic system within the system is expressed in kW. This indicates the degradation factor of photovoltaic power generation efficiency. Indicates the total number of photovoltaic modules. This indicates the rated power capacity of a single module, in kW. Indicates time period The intensity of light radiation within, in units of , This represents the reference radiation intensity under standard test conditions, in units of... , This represents the temperature-power correction factor. Indicates time period The corresponding ambient temperature, in °C. Represents the temperature-radiation coupling coefficient. This indicates the standard operating condition reference temperature, in °C.

[0016] Wind power generation is a technology that converts natural wind energy into electrical energy. Wind energy development has relatively little impact on the ecological environment, thus holding significant strategic importance in the low-carbon transition of the energy structure. The operating characteristics of wind turbine units are mainly determined by three key wind speed parameters: cut-in wind speed, rated wind speed, and cut-out wind speed. Furthermore, due to the random fluctuations in wind speed, wind power output also exhibits significant uncertainty. The wind power model is as follows: ; in, Indicates time period The output power of the wind power generation system within the system is measured in kW. Indicates time period Internal airflow intensity, in m / s. This indicates the fan's cut-in wind speed, expressed in m / s. This represents the reference airflow intensity under standard operating conditions, expressed in m / s. This indicates the nominal power capacity of the wind turbine unit, in kW. This indicates the cut-off wind speed of the fan, in m / s.

[0017] The mathematical models of heating and cooling energy supply equipment include gas boiler models, electric boiler models, heat pump models, and electric refrigeration unit models, which respectively characterize the conversion relationship between heat / cold energy output and energy input of each equipment.

[0018] Gas-fired boilers use natural gas as fuel to produce steam or high-temperature hot water through combustion, meeting various heat load demands within the industrial park, including industrial processing and heating. While they offer excellent operational economics, the initial investment cost is relatively high due to the need for numerous auxiliary equipment. The gas-fired boiler model is as follows: ; in, Indicates time period The thermal power output of the gas-fired boiler, in kW. Indicates time period The volume of natural gas consumed within, in units of , Indicates the thermal conversion efficiency coefficient of the unit. This indicates the parameter for converting the calorific value of natural gas.

[0019] Electric boilers, based on resistance heating or electromagnetic induction principles, convert electrical energy into heat energy, providing various forms of heat output such as steam, high-temperature hot water, or organic heat-conducting fluids to meet different heating needs within the industrial park. Their integrated design gives them an advantage in initial purchase cost. However, due to limitations in energy conversion efficiency, their long-term operating energy costs often exceed the initial investment. The operating characteristics of an electric boiler model can be characterized by the following mathematical model: ; in, Indicates time period The unit's thermal power output value, in kW. Indicates time period The unit's electrical energy input parameters are in kW. This represents the unit's thermal conversion efficiency coefficient.

[0020] A heat pump is a heat energy transfer device based on the reverse Carnot cycle principle. It uses electricity to upgrade low-grade heat energy from soil, water, or air into usable high-grade heat energy. This device requires only a small amount of electricity to achieve a several-fold increase in heat energy conversion, exhibiting significant energy-saving and environmentally friendly characteristics. The heat pump operates through heat spatial redistribution: in summer, it discharges indoor heat to the external environment, and in winter, it draws heat from the outside and sends it indoors, thus fulfilling the seasonal temperature control needs of the building environment. The output characteristics of a heat pump model can be characterized by the following model: ; in, Indicates time period The internal heat pump's heating and cooling output power, measured in kW. Indicates time period The internal heat pump cooling output power, measured in kW. Indicates time period The internal heat pump output power, measured in kW. Indicates the coefficient of performance (COP) of a heat pump. This indicates the coefficient of performance (COP) of the heat pump for cooling.

[0021] The basic working principle of an electric chiller is similar to that of a household air conditioner and refrigerator. The device uses electricity to drive a compressor, which causes the refrigerant to circulate within the system. The refrigerant evaporates into a gas under low pressure, absorbing heat from the surrounding environment in this process. After being pressurized by the compressor, the gaseous refrigerant releases the absorbed heat under high pressure. Essentially, this process achieves spatial heat transfer through the refrigerant, delivering indoor heat to the external environment, rather than directly generating cooling. Therefore, electric chillers have high cooling efficiency and relatively low operating power consumption. The output model of the electric chiller unit is as follows: ; in, Indicates time period The unit's cooling power output, in kW. Indicates time period The unit's electrical input power is expressed in kW. This represents the energy efficiency coefficient of an electric refrigeration unit.

[0022] The mathematical models of multi-energy supply equipment include the combined heat and power model and the combined cooling-heating-power (CCHP) model, which reflect the logic of energy cascade utilization.

[0023] Combined heat and power (CHP) is a highly efficient energy utilization technology. Traditional power generation methods generate significant waste heat during fuel combustion, which is typically discharged directly, leading to energy waste. CHP systems, however, can recover this waste heat for secondary power generation or direct heating, achieving cascaded energy utilization and combining resource conservation with economic advantages. The core equipment of this system includes power generation units (such as gas turbines) and heating units (such as waste heat boilers), which work together to provide combined electricity and heat. The energy output characteristics of this CHP system can be described by the following model: Gas turbines use natural gas as fuel. The high-temperature, high-pressure gas generated by combustion drives a turbine to rotate, which in turn drives a generator, achieving a step-by-step conversion of chemical energy into thermal energy, mechanical energy, and electrical energy. The high-temperature flue gas emitted during operation can be converted into useful heat energy through a waste heat recovery system. This cascaded energy utilization significantly improves overall energy efficiency. The output characteristics of a gas turbine unit are as follows: ; in, Indicates time period The unit's electrical power output, in kW. Indicates time period The volume of natural gas consumed within, in units of , This represents the unit's electrical conversion efficiency coefficient.

[0024] The waste heat power model of a gas turbine during power generation is as follows: ; in, Indicates time period The waste heat generated during power generation, measured in kW. This indicates the unit's output heat loss efficiency.

[0025] The waste heat boiler uses the waste heat emitted by the gas turbine during power generation to provide heating, and its output is as follows: ; in, Indicates time period The effective thermal power output of the waste heat boiler within the unit is expressed in kW. Indicates time period The waste heat input power of the waste heat boiler within the facility, measured in kW. This indicates the thermal conversion efficiency of the waste heat boiler.

[0026] The combined cooling, heating, and power (CCHP) system introduces absorption refrigeration technology into the traditional cogeneration system, constructing a multi-grade energy cascade utilization system. During system operation, high-grade heat energy generated from fuel is first used for gas turbine power generation; waste heat in the mid-temperature section drives the absorption refrigeration unit for cooling; and low-temperature waste heat at the end is used to meet heating or hot water needs through a heat exchange network. When heat supply is insufficient, temperature levels can be increased through supplementary combustion. This graded utilization method significantly improves overall energy efficiency, reduces pollutant emissions, and effectively balances seasonal peak-valley differences in electricity load through CCHP. The gas turbine and waste heat boiler in the system operate in CCHP mode, and the output characteristic model of the refrigeration unit is as follows: ; in, Indicates time period The unit's cooling capacity output is in kW. Indicates time period The unit's waste heat input power, in kW. This represents the unit's energy efficiency coefficient for converting cooling capacity into energy.

[0027] Since the heat source for both devices comes from the waste heat generated during the gas turbine power generation process, their heat source coupling relationship can be described as follows: .

[0028] The mathematical models of energy storage devices include electrical energy storage models, thermal energy storage models, and cold energy storage models, which constrain the charging and discharging power and energy storage boundaries of each energy storage device.

[0029] Energy storage devices, which store and release energy through specific media, are key facilities supporting the operation of multi-energy coupled systems. Their main functions include smoothing supply and demand fluctuations, improving the overall efficiency of energy utilization, and ensuring the reliability of energy supply, thus contributing to the flexibility and stability of the energy system. Based on the form of stored energy, energy storage technologies can be divided into electrical energy storage, thermal energy storage, and cryogenic energy storage. The synergistic application of multiple energy storage technologies can meet the coordination and reserve needs of large-scale renewable energy distribution in time and space, support the reliable operation of the power grid under a high proportion of new energy access, and provide an effective means of regulating the long-term supply and demand balance of the power system, thereby improving the overall operating efficiency and coordination capability of the energy system.

[0030] Energy storage technology is mainly divided into two types: chemical energy storage and physical energy storage. Based on the installation location and functional positioning of the equipment, its application scenarios can be divided into three categories: power source side, grid side, and user side. Power source side energy storage is usually built in conjunction with renewable energy generation to smooth out output fluctuations from intermittent power sources such as wind and solar power, and improve the stability and absorption level of clean energy grid connection. Grid side energy storage mainly includes independent energy storage power stations and grid alternative facilities, aiming to enhance the power supply reliability and operational flexibility of the power system. User side energy storage focuses on improving power quality and reliability, alleviating peak grid pressure through peak-valley regulation, and reducing users' energy costs. Through the coordinated optimization of energy storage resources on the power source side, grid side, and user side, stable support can be provided for high-proportion renewable energy power systems, alleviating the short-term regulation and long-term balance pressure caused by the mismatch between power output and load demand. The operating characteristics of the energy storage model can be characterized by the following model: ; in, Indicates time period Internal storage energy level, in kWh Indicates time period Internal storage energy level, in kWh Indicates the energy loss coefficient; Indicates time period Energy conversion efficiency during charging, measured in kW. Indicates time period Energy conversion efficiency during internal discharge, expressed in kW. Indicates charging efficiency. An identifier representing the charging state; 1 indicates that the device is in that state. An identifier representing the discharge state; 1 indicates that the state is in that state. Indicates discharge efficiency. This indicates the maximum storage capacity, expressed in kWh.

[0031] Thermal energy storage technology utilizes media such as water and molten salt to store energy from solar thermal, geothermal, and industrial waste heat, and then releases the heat directly or converts it into other energy forms as needed. Its applications are mainly concentrated on the power supply side and the load side: on the power supply side, molten salt thermal storage systems are often used to temporarily store solar heat, releasing it during fluctuations in renewable energy output to stabilize the power grid; on the load side, hot water is commonly used as the storage medium to meet building heating and domestic hot water needs, and load regulation is used to achieve peak shaving and valley filling. Benefiting from its long-term regulation characteristics, thermal energy storage has significant value in power grid balance and load management. The operating characteristics of thermal energy storage systems can be characterized by the following model: ; in, Indicates time period The remaining heat storage in the container is expressed in kWh. Indicates time period The remaining heat storage in the container is expressed in kWh. This represents the heat loss coefficient, expressed as a percentage. Indicates time period Internal thermal energy storage capacity, measured in kW. Indicates time period The internal thermal energy storage heat release power, measured in kW. Indicates the efficiency of thermal energy conversion. This represents the thermal storage status identifier; 1 indicates that the system is in that state. This indicates the exothermic state identifier; 1 indicates that the state is in that state. Indicates the efficiency of exothermic energy conversion. This indicates the rated thermal storage capacity, expressed in kWh.

[0032] Cold energy storage systems primarily utilize media such as water and ice to store cold energy through a refrigeration cycle driven by electricity. This technology is typically applied on the load side, enabling the transfer of electricity load in time and space by regulating peak and off-peak periods of cooling demand, effectively alleviating the power supply pressure on the grid during peak hours. The cold energy storage model can be characterized as follows: ; in, Indicates time period The remaining cooling capacity within, in kWh. Indicates time period The remaining cooling capacity within, in kWh. This represents the cooling loss coefficient, expressed as a percentage. Indicates time period Internal cold storage capacity, measured in kW. Indicates time period Internal cooling capacity, measured in kW. Indicates the efficiency of cold energy storage conversion. Indicates the efficiency of energy conversion from cooling. This indicates the cold storage state identifier; 1 indicates that the state is in that state. This indicates the cooling state identifier; 1 indicates that the state is in that state. This indicates the rated cold storage capacity, expressed in kWh.

[0033] Mathematical models for flexible loads and demand response equipment include flexible load utility function models, price-based demand response models, and incentive-based demand response models, which quantify the economics of flexible loads and the load regulation characteristics of demand response.

[0034] Flexible loads possess the ability to respond to electricity price signals and dynamically adjust their electricity consumption patterns based on market incentives. Within the multi-energy coupled system framework constructed in this embodiment, flexible loads can be modeled as dispatchable units within the system. The operational economy of flexible loads is assessed through a utility function. Quantitative representation, its modeling is usually in the form of a quadratic function: ; in, Indicates time period The power consumption adjustment of flexible loads within the system, in kW. This represents the coefficient of the quadratic term, which is usually taken as a negative value to reflect the property of diminishing marginal returns. This represents the linear term coefficient, the basic energy efficiency value of flexible loads. This represents a constant term, or fixed cost.

[0035] In addition, when the active power output of the flexible load When the flexible load exhibits piecewise linear variation characteristics, the energy consumption of the flexible load is... It can be calculated using the following formula: .

[0036] Demand response uses price signals to guide users to change their electricity consumption behavior, thereby reducing load fluctuations and offering good flexibility. Based on its mechanism of action, demand response can be divided into price-based demand response and incentive-based demand response.

[0037] Price-based demand response models guide users to adjust their electricity consumption behavior based on diversified electricity price signals, thereby achieving optimal load allocation in a time-series manner. Based on microeconomic market equilibrium theory, this paper uses elasticity coefficients to characterize the dynamic relationship between electricity price changes and load response. Demand elasticity mainly includes two types: self-elasticity coefficients and cross-elasticity coefficients. Self-elasticity coefficients reflect the degree of impact of electricity price changes on load within the same time period; cross-elasticity coefficients characterize the effect of electricity price changes in other time periods on the load in the current time period. By introducing these elasticity coefficients, the interaction between electricity price and load can be effectively quantified. Price-based demand response actively guides users' electricity consumption patterns through electricity price leverage, and is an important load management tool for promoting the spatiotemporal balance of electricity supply and demand at the system level. The price-based demand response model can be characterized as follows: ; in, This represents the electricity price elasticity coefficient over a time period. The extent to which changes in domestic electricity prices affect changes in load demand. Indicates time period Real-time frontload demand values ​​within the demand response, in kW. Indicates time period Real-time load demand values ​​after demand response, in kW. Indicates time period The real-time difference between the load demand before and after the load, expressed in kW. Indicates time period Pre-demand response electricity price level Indicates time period Domestic demand response electricity price level Indicates time period The difference in electricity prices before and after the demand response.

[0038] Incentive-based demand response differs from price-based mechanisms in that it operates based on a pre-existing load control agreement between virtual power plants (VPS) and users. When the system needs to regulate load, the VPS can directly initiate contract execution, and users participate in the regulation as agreed, receiving corresponding economic compensation upon completion. However, since incentive expenditures, as a cost item, affect the VPS's revenue level, this mechanism typically needs to incorporate market electricity price fluctuations and employ a time-of-use dynamic incentive strategy to balance load regulation demand with operational economics. The incentive-based demand response model can be characterized as follows: ; in, Indicates the time period The first In each scheduling step, the actual load adjustment provided by the load user is expressed in kW. Indicates the time period The first In each scheduling step, the lower limit of the load adjustment that the user can provide is... Indicates the time period The first The upper limit of load adjustment that a user can provide in each scheduling step. Indicates the time period The first The upper limit of load adjustment that a user can provide in each scheduling step. Indicates the time period The internal virtual power plant passed before The total load regulation achieved by each scheduling step, in kW.

[0039] Based on the above analysis, after users within a virtual power plant access price-oriented and agreement-incentive-based demand response mechanisms, the total load reduction can be expressed as: ; in, Indicates the time period The total electrical load after demand response from users within the internal virtual power plant, expressed in kW. Indicates the time period Electricity demand value of users participating in price signal-driven demand response, in kW.

[0040] Step 102: Based on the mathematical model, an improved Gaussian mixture model and an improved expectation-maximization algorithm are used, combined with the spatiotemporal correlation characteristics of wind and solar power output and load, to generate a set of typical scenarios with probability distributions that fit reality.

[0041] A Gaussian Mixture Model (GMM) is a generative model based on probability statistics. Its core idea is to fit the overall distribution of a complex dataset using a linear combination of the probability density functions of several Gaussian distributions. Suppose there exists a Gaussian mixture model... Let be a historical dataset of independent and identically distributed samples, denoted as . ,in Representing the Data points, The probability distribution can be approximated by the GMM as follows: ; in, for A set of mixed weights for Gaussian components. Indicates the first The weights of each Gaussian component in the mixture model; for A set of mean vectors of Gaussian components. Indicates the first The mean vector of Gaussian components; for A set of covariance matrices of Gaussian components Indicates the first The covariance matrix of Gaussian components; For dataset Total amount of data; The total number of Gaussian components; Indicates to The probability of each data point is multiplied together. Since the data points are independent, the joint probability is obtained by multiplying the probabilities of the individual data points. Indicates to The probabilities of the Gaussian components are weighted and summed. Let be the probability density function of a Gaussian distribution, describing the th Data points In the Gaussian components (mean is 1) Covariance is The probability of generating ( ). If , indicating the first The data point belongs to the th data point A Gaussian component, if , indicating the first The data point does not belong to the first... One Gaussian component, Latent variables With weight The relationship satisfies the following conditions: .

[0042] In order to determine To determine the probability distribution, we first need to obtain the GMM's probability distribution. , , and These are the core parameters. The EM algorithm is typically used to optimize these core parameters through iterative E-step and M-step calculations. The EM algorithm optimizes the core parameters of the Gaussian distribution (GMM) iteratively. In the E-step, the component assignment probability of data points is calculated based on the current Gaussian distribution; in the M-step, the Gaussian distribution parameters are updated based on these probabilities. This process is repeated until convergence.

[0043] The expression for step E is: ; in, Denotes responsibility (posterior probability), and represents the first... The data point belongs to the th data point The probabilities of the Gaussian components are calculated using the E-step: , It is a data point, representing the first... The specific numerical value of each data point (which can be a scalar or a vector).

[0044] The expression for M steps is ; in, Indicating the first in the prior art The expected number of samples for the _th Gaussian component, representing the total number of data points for the _th _th Gaussian component. The sum of the responsibilities of each component.

[0045] The traditional GMM mentioned above is based on the assumption of independence of latent variables, and its responsibility calculation depends only on the mixed weight parameters. and by mean Covariance The Gaussian distribution term is formed. However, this method requires specific input data. The inherent time correlations are not adequately characterized. In actual power system operation, renewable energy output and load data exhibit significant time-series correlations. Ignoring such dynamic dependencies will limit the accuracy of uncertainty modeling.

[0046] To address this, this embodiment proposes the concept of "latent variable mutual assistance" to describe the conditional dependency between latent variables at adjacent time points, meaning that the state of a latent variable at the current time point is influenced by the value of the latent variable at the previous time point. Based on this, a Latent Variable Mutual Assistance-Gaussian Mixture Model (LMA-GMM) is constructed. By introducing a mutual assistance mechanism between latent variables, it effectively captures the input data. By understanding the dynamic evolution patterns over time, we can improve the accuracy of uncertainty representation and the ability to reconstruct models.

[0047] See Figure 2 As shown, based on historical time-series data, an improved expectation-maximization algorithm is used to train a latent variable mutual-aid Gaussian mixture model. The core parameters of the latent variable mutual-aid Gaussian mixture model can include: Step 201: Obtain historical time-series data of the core equipment.

[0048] The historical time series data here can include... The historical dataset of independent and identically distributed samples is... .

[0049] Step 202: Based on historical time series data and the mutual assistance among latent variables, construct a latent variable mutual assistance Gaussian mixture model.

[0050] Conditional probability can be used to describe the mutual assistance between latent variables, and its expression can be: ; Based on the latent variable interaction information of adjacent data, define The expression is: ; in, Indicates the first The data point at the th th The values ​​of the Gaussian components, Indicates the first The data point belongs to the th data point One Gaussian component, Indicates the first The data point at the th th The values ​​of the Gaussian components, Indicates the first The data point belongs to the th data point One Gaussian component, Indicates belonging to the first Data points with Gaussian components Depends on the first Data points with Gaussian components The probability, Indicates belonging to the first Data points with Gaussian components Dependent on data points The probability, Indicates the first The data point belongs to the th data point The probability of a Gaussian component, Indicates the first The data point belongs to the th data point The probability of a Gaussian component.

[0051] in, yes The explicit expression form helps to extract the historical time-series input data more directly. The algorithm incorporates correlation information. To more effectively solve the parameter estimation problem in a latent variable mutual aid Gaussian mixture model, this embodiment designs an improved expectation-maximization algorithm. This algorithm introduces the mutual aid relationship between latent variables in the E-step and adjusts the parameter update process in the M-step accordingly, thereby improving the accuracy of model parameter estimation and accelerating the overall convergence speed of the algorithm. See step 203 for details.

[0052] Step 203: Based on historical time series data, train the latent variable mutual aid Gaussian mixture model using the improved expectation-maximization algorithm to obtain the core parameters of the latent variable mutual aid Gaussian mixture model. Substitute the core parameters into the latent variable mutual aid Gaussian mixture model to obtain the target latent variable mutual aid Gaussian mixture model.

[0053] In one embodiment, see Figure 3As shown, based on historical time-series data, an improved expectation-maximization algorithm is used to train a latent variable mutual-aid Gaussian mixture model to obtain the core parameters of the latent variable mutual-aid Gaussian mixture model, which may include the following steps.

[0054] Step 301: Input historical time series data and initialize the core parameters in the latent variable mutual Gaussian mixture model.

[0055] Optional, initialize kernel parameters , and Calculate the initial log-likelihood function .

[0056] Step 302: Based on the dependency relationship between latent variables at adjacent time points and the Gaussian mixture model of latent variable mutual assistance, calculate the responsibility degree of each data point in the historical time series data.

[0057] For each data point, when recalculating its responsibility, the parameters at the current moment are estimated based on information from the previous moment, thus corresponding to the time correlation in the modeling.

[0058] Based on the dependencies between latent variables at adjacent time points, the degree of responsibility for each data point in the historical time series data can be calculated, which may include: according to Calculate the degree of responsibility for each data point in the historical time series data; in, This indicates the degree of responsibility considering the mutual assistance of latent variables. Indicates the first The weights of each Gaussian component in a latent variable mutual Gaussian mixture model. Indicates the first Data points, Indicates the first The mean vector of Gaussian components, Indicates the first The covariance matrix of Gaussian components, This represents a Gaussian mixture model with latent variable mutual assistance. This indicates the status identifier of the data point.

[0059] Step 303: Update the core parameters according to the degree of responsibility to obtain the updated core parameters.

[0060] Based on the above consideration of the responsibility degree regarding the mutual assistance of latent variables, the core parameters are revised, namely according to... Update the core parameters separately; in, Indicates the first The expected number of samples for each Gaussian component. This represents the total amount of historical time-series data.

[0061] Step 304: Based on the updated core parameters, a new latent variable mutual-aid Gaussian mixture model is obtained.

[0062] The updated core parameters replace the original core parameters, resulting in a new latent variable mutual Gaussian mixture model, which forms a new Gaussian distribution.

[0063] Step 305: Check whether the absolute value of the difference between the updated core parameters and the corresponding core parameters before the update meets the preset conditions. If the absolute value of the difference does not meet the preset conditions, the new latent variable mutual aid Gaussian mixture model is used as the current latent variable mutual aid Gaussian mixture model. Then, proceed to "Calculate the responsibility of each data point in the historical time series data based on the dependency relationship between latent variables at adjacent time points and the latent variable mutual aid Gaussian mixture model" and subsequent steps until the absolute value of the difference meets the preset conditions, and then output the current core parameters.

[0064] Optionally, in this embodiment, with , Two core parameters are used to determine the convergence of the detection target. When , If both conditions are met, the preset condition is satisfied, meaning the improved expectation-maximization algorithm has converged, and the process ends. Otherwise, the process jumps to step 302 to recalculate the responsibility degree based on the new Gaussian distribution until... , Established at the same time.

[0065] , The threshold used to measure the magnitude of change directly affects the accuracy of parameter estimation results and the convergence speed of the EM algorithm. , Let represent the mean vectors before and after the update, respectively. , Let represent the covariance matrices before and after the update, respectively.

[0066] In fact, Defined as the lower bound of LMA-GMM, see [link / reference] Figure 4 The diagram illustrates the iterative process of the improved EM algorithm, where the lower bound is obtained through step 302 (E-step). In step 303 (M-step), after determining the core parameters, the log-likelihood function is calculated. During the iteration of the improved EM algorithm, the lower bound and the log-likelihood function value increase alternately until the lower bound reaches its maximum value, at which point the improved EM algorithm is considered to have converged.

[0067] from Figure 4The relationship between the convergence curve and the lower bound shows that the modified E-step yields a more compact variational lower bound. This compact lower bound implies a larger likelihood function gain in each iteration, theoretically guaranteeing that the improved EM algorithm has better convergence performance. Essentially, by introducing the mutual correlation between latent variables, it improves the representation learning of the latent structure, enhances the efficiency of using implicit information, and makes the lower bound model closer to the true posterior distribution. Theoretical analysis and experimental verification together demonstrate that the LMA-GMM method achieves a better balance between generalization performance and computational complexity in uncertainty modeling.

[0068] Step 204: Using the target latent variable mutual Gaussian mixture model as the probability distribution template, Markov chain Monte Carlo is used to generate wind-solar-load time series trajectories that conform to the probability distribution template.

[0069] The target latent variable mutual Gaussian mixture model has accurately characterized the spatiotemporal correlation probability distribution of wind and solar power output and load (including the correlation of data at different times and the overall fluctuation pattern), providing a target stationary distribution for Markov chain Monte Carlo simulation. Finally, MCMC generates wind and solar load data samples that both conform to the actual fluctuation pattern and retain the temporal correlation, providing high-quality original samples for subsequent K-means clustering to obtain a "representative typical scenario set".

[0070] Step 205: The K-means clustering algorithm is used to perform dimensionality reduction clustering on the wind-solar-load time series trajectory to obtain a set of typical scenarios with probability distributions that fit the actual situation.

[0071] The large number of generated wind-solar-load time-series trajectories would lead to a significant increase in computational load if directly used to optimize the scheduling model. Therefore, the K-means clustering algorithm is used to perform "dimensionality reduction clustering" on the trajectories. Each time-series trajectory is treated as a high-dimensional vector (dimension is the number of time steps × the number of wind / solar / load types). Similar trajectories are clustered into one class using the K-means algorithm, and the center trajectory of each cluster represents a typical scenario. The resulting typical scenarios cover the main patterns of wind-solar and load fluctuations (such as "high wind-solar output - low load" and "low wind-solar output - high load" scenarios), which are used for subsequent multi-objective optimization scheduling.

[0072] Step 103: With the goal of minimizing the total system cost, and in combination with the operating boundary conditions of the core equipment and typical scenario sets, construct an optimized scheduling model that includes multi-energy supply and demand balance constraints. After linearizing the optimized scheduling model, solve it to obtain a flexible adjustment scheme for the multi-energy coupled system.

[0073] In one embodiment, the objective function included in the optimized scheduling model is constructed with the goal of minimizing the sum of energy market interaction costs, equipment asset depreciation costs, facility maintenance expenses, energy consumption costs during unit start-up and shutdown, electricity purchase and sale costs, and load response control costs.

[0074] To minimize the total system cost, the objective function for constructing multi-energy cooperative scheduling can be expressed as: ; in, This represents the total system cost, in yuan. This represents the energy market interaction cost, expressed in yuan. This indicates equipment depreciation costs, expressed in yuan. This indicates facility maintenance expenses, expressed in yuan. This indicates the energy consumption cost during the start-up and shutdown process of the generating unit, expressed in yuan. This represents the cost of purchasing and selling electricity from the power grid, in yuan. This represents the cost of load response control, expressed in yuan.

[0075] The various costs will be analyzed separately below.

[0076] Energy market interaction costs are ; in, Indicates the power grid during the time period Electricity purchase volume, in kWh. Indicates time period The unit price of electricity in the power grid, expressed in yuan / kWh. Indicates the time period of the hot network The amount of heating resources input, in kWh. Indicates time period The unit price of heat energy in the heating network is yuan / kWh. This indicates that the combined cooling, heating and power (CCHP) unit operates within a certain time period. Energy output, measured in kW. This indicates the time period of the combined cooling, heating, and power (CCHP) system. Energy conversion efficiency This represents the conversion parameter for the lower heating value of natural gas, in units of... , Indicates time period The gas market purchase price, in units of .

[0077] Equipment asset depreciation expense is ; in, Indicates time period Inner Estimated net asset value of each generating unit, in yuan. Indicates the first The total initial purchase cost of each generating unit, in yuan. Indicates the first The residual depreciation ratio of each unit Indicates the first The expected remaining runtime of each unit, in hours. This indicates the total number of schedulable units within the system.

[0078] Facility maintenance expenses are ; in, Indicates time period No. Real-time power output of each generating unit, in kW. Indicates the first The unit power operation and maintenance cost parameters for each generating unit are expressed in yuan / kW.

[0079] The energy consumption cost during the start-up and shutdown of the unit is ; in, Indicates the first The comprehensive loss cost of each generating unit's start-up and shutdown process, in yuan. Indicates time period No. The logical identifier parameter for the operating status of each unit, where 0 indicates shutdown and 1 indicates operation. Indicates time period No. Logical identifier parameters for the operating status of each unit.

[0080] The electricity purchase and sale cost of the power grid is ; in, Indicates time period The power input between the virtual power plant and the power grid, expressed in kW. Indicates time period The power output between the virtual power plant and the power grid, measured in kW, is defined as follows: Electricity purchase is at time >0. Electricity is sold at 0:00. Indicates the virtual power plant during the time period Electricity purchase market rates of the power grid Indicates the virtual power plant during the time period The electricity sales settlement price.

[0081] ; in, Indicates time period The dynamic incentive compensation rate for domestic demand response, in yuan / kWh, Indicates time interval The schedulable reduction capacity of interruptible load resources, in kW.

[0082] Multi-energy supply and demand balance constraints include: dynamic balance constraints on electricity supply and demand, dynamic balance constraints on heat supply and demand, and dynamic balance constraints on cooling supply and demand; The optimized scheduling model also includes: equipment safety operation constraints, flexible load constraints, and demand response constraints. Among them, equipment safety operation constraints include unit rated capacity constraints, unit power ramp-up rate constraints, and energy storage system operating state boundary constraints. Energy storage system operating state boundary constraints include dynamic boundary constraints of state of charge, dynamic constraints of unit capacity of thermal energy storage system, and dynamic constraints of unit capacity of cold energy storage system.

[0083] The dynamic balance constraint of power supply and demand is ; in, Indicates time period Internal photovoltaic array output power, in kW. Indicates time period The output power of the internal wind turbine unit, in kW. Indicates time period Output power of internally cooled combined heat and power (CHP) units, in kW. Indicates time period Net output power of internal energy storage devices, in kW. Indicates time period Baseline value for internal electricity load demand, in kW. Indicates time period Internally adjustable load response power, in kW. Indicates time period Electrical power input of the internal heat pump unit, in kW. This represents the thermoelectric conversion efficiency coefficient of a heat pump. Indicates time period Rated power of the internal electric heating unit, in kW. This indicates the thermal efficiency parameter of the electric boiler. Indicates time period The operating power of the internal electric cooling unit is measured in kW. Indicates the energy efficiency ratio of the refrigeration unit. Indicates time period The discharge power of the internal energy storage system is expressed in kW. Indicates time period The charging power of the internal energy storage system, measured in kW.

[0084] The dynamic balance constraint of heat supply and demand is ; in, Indicates time period Heating output of internally cooled combined heat and power (CHP) units, measured in kW. Indicates time period Heating power of internal heat pump units, in kW. Indicates time period The thermal power output value of the indoor unit is expressed in kW. Indicates time period The net heat release power of the internal heat storage device is expressed in kW. Indicates time period Baseline internal heat load demand, in kW. Indicates time period Internally adjustable heat load reduction capacity, in kW. Indicates time period The heat release power of the internal heat storage device is expressed in kW. Indicates time period The thermal storage capacity of the internal thermal storage device is expressed in kW.

[0085] The dynamic balance constraint of cold supply and demand is ; in, Indicates time period Cooling output of a combined cooling, heating, and power (CCHP) unit, measured in kW. Indicates time period Cooling power of internal heat pump units, measured in kW. Indicates time period Internal heating power output value, in kW. Indicates time period The net cooling power of the internal cold storage device is expressed in kW. Indicates time period Baseline internal cooling energy demand, in kW. Indicates time period The cooling power of the internal cold storage device is measured in kW. Indicates time period The internal cold storage device's cold storage power is measured in kW.

[0086] The rational and economical operation of virtual power plants requires the optimization of scheduling of multiple energy resources to achieve energy supply and demand matching for end users. Therefore, in the process of regulation optimization, the regulation model is required to strictly follow the boundary conditions for safe operation, such as the power limits and ramp rates of generator sets, energy storage systems and transmission and distribution equipment.

[0087] The rated capacity constraint of the unit is ; in, Indicates the first The maximum rated output power of each unit, in kW.

[0088] The unit power ramp-up rate constraint is limited to ; in, Indicates the first The power downgrade rate constraint limit for each generating unit, in kW. Indicates time period Inner Real-time power output of each generating unit, in kW. Indicates the first The power ramp-up rate constraint for each generating unit is in kW.

[0089] In order to extend the cycle life of energy storage batteries, deep charging and discharging should be avoided during operation. Therefore, in addition to being limited by the rated power, the charging and discharging power of energy storage batteries also needs to meet the dynamic boundary constraints of the state of charge.

[0090] The dynamic boundary constraints of the charged state are ; in, Indicates time period State of charge of internal energy storage battery, in kWh. This indicates the lower limit of the safe state of charge of the energy storage battery, in kWh. This indicates the safe upper limit of the state of charge of the energy storage battery, in kWh. This indicates the initial state of charge (SPC) of the energy storage battery during the scheduling cycle, expressed in kWh. This indicates the state of charge at the end of the energy storage battery scheduling cycle, in kWh. This indicates the maximum charging power limit of the energy storage battery, in kW. This indicates the maximum discharge power limit of the energy storage battery, in kW. This indicates the rated capacity of the energy storage battery, in kWh. This represents the energy conversion efficiency coefficient of an energy storage battery during charging. This represents the discharge energy conversion efficiency coefficient of the energy storage battery. Indicates the scheduling time interval.

[0091] The heat storage and release power of the heat storage devices in a thermal energy storage system must comply with rated power limits and meet the dynamic capacity constraints of the system. The dynamic capacity constraints of the thermal energy storage system are as follows: ; in, This indicates the maximum thermal storage capacity limit of the thermal storage tank, in kW. This indicates the maximum heat release power limit of the thermal storage tank, in kW. This indicates the rated capacity of the thermal storage tank, in kWh. This represents the thermal energy conversion efficiency coefficient of the thermal storage tank. This represents the efficiency coefficient of heat energy conversion in the heat storage tank.

[0092] The cold storage and release power of the cold energy storage devices in a cold energy storage system must comply with rated power limits and meet the dynamic capacity constraints of the cold energy storage system. The dynamic capacity constraints of the cold energy storage system are as follows: ; in, This indicates the maximum cold storage capacity limit of the ice storage tank, in kW. This indicates the maximum cooling capacity limit of the ice storage tank, in kW. This indicates the rated capacity of the ice storage tank, in kWh. This represents the energy conversion efficiency coefficient of the ice storage tank. This indicates the energy conversion efficiency coefficient of the ice storage tank. Flexible load constraints are ; in, Indicates flexible load In time period Scene The actual operating power is expressed in kW. Indicates flexible load In the scene The minimum permissible power, in kW. Indicates flexible load In the scene The maximum permissible power, in kW. Indicates flexible load The power decay rate limit, in kW / min. Indicates flexible load The power rise rate limit, in kW / min. Indicates flexible load In time period Scene The power value is given below, in kW. Indicates flexible load In time period Scene The cumulative energy below, in kWh. Indicates flexible load The minimum required energy, in kWh. This indicates the total number of time periods in the scheduling cycle.

[0093] The effectiveness of virtual power plants in regulating internal load through demand response mechanisms is limited, especially in price-based demand response. To prevent peak-valley swapping of load curves and subsequent peak-valley inversion, constraints are needed to limit the spatiotemporal migration range of user loads.

[0094] Demand response constraints are ; in, Indicates time period Adjusted load power after implementing price-based demand response, in kW. Indicates time period The baseline load power, in kW, represents the original load demand before demand response was implemented. Indicates time period The maximum allowable load adjustment range, in kW, is determined by the user's willingness to respond and the contract terms. Indicates the first The dispatchable reduction capacity per user, in kW, represents the actual reduction in load power during incentive-driven demand response. This represents the permissible reduction percentage of interruptible load in a virtual power plant; it is dimensionless. Limited by user contracts or technical conditions Indicates the first The total interruptible load capacity of a user, expressed in kW, is the maximum adjustable load that can participate in demand response.

[0095] In one embodiment, the aforementioned optimization scheduling model is a mixed-integer nonlinear programming problem, which is linearized and then solved using the CPLEX solver. Multi-level optimization, progressive refinement of the solution method, and relaxation techniques are employed to improve computational robustness.

[0096] This invention provides an optimization method for typical scenarios of wind and solar power and load. It establishes mathematical models of core equipment within a multi-energy coupled system to characterize the output characteristics, energy conversion efficiency, and operational boundary conditions of each device. Based on these mathematical models, an improved Gaussian mixture model and an improved expectation-maximization algorithm are used, combined with the spatiotemporal correlation characteristics of wind and solar power output and load, to generate a set of typical scenarios with a probability distribution that closely matches reality. With the goal of minimizing the total system cost, and combining the operational boundary conditions of the core equipment and the set of typical scenarios, an optimized scheduling model containing multi-energy supply and demand balance constraints is constructed. The optimized scheduling model is then linearized and solved to obtain a flexible adjustment scheme for the multi-energy coupled system. This invention breaks through the assumption of mutual independence of latent variables in traditional GMMs. By introducing the concept of "mutual assistance of latent variables," it cleverly characterizes the temporal correlation between data points at adjacent time points. In the E-step of the improved EM algorithm, the calculation method of responsibility degree is reconstructed using conditional probability, enabling it to integrate historical latent variable information. This allows it to capture the dynamic evolution of uncertain data, comprehensively reflect the impact of multiple uncertain factors on the system's adjustment capability, and improve the modeling accuracy of the probability distribution of uncertain variables such as wind, solar, and load. Ultimately, this improves the model fitting accuracy and accelerates the algorithm's convergence speed, laying a more accurate probabilistic foundation for generating more representative typical scenarios of wind, solar, and load in the future.

[0097] This invention also constructs an optimized scheduling model for virtual power plants with multi-energy supply and demand balance constraints, deeply integrating multi-energy flow coordination and demand response, thus realizing the coordinated regulation of multi-energy complementarity and demand-side resources. This optimized scheduling model incorporates multi-energy flow coupling equipment (electricity, heat, cooling, and gas) and price-based and incentive-based demand response mechanisms into a unified optimization framework. By establishing a refined operation model and multi-energy balance constraints covering photovoltaic, wind power, combined heat and power, heat pumps, electric chillers, and various energy storage systems, it fully characterizes the transformation and balance relationships of multi-energy flows in the spatiotemporal dimensions. Simultaneously, it incorporates price-based demand response based on elasticity coefficients and incentive-based demand response based on contract mechanisms into the objective function and constraints, making flexible loads dynamically dispatchable virtual resources. This achieves coordinated optimization of resources on both the source and load sides, improving system operational economy while ensuring energy supply reliability, and providing an advanced technical solution for the commercial operation of virtual power plants with a high proportion of renewable energy access.

[0098] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0099] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0100] Figure 5 This is a schematic diagram of a terminal provided in an embodiment of the present invention. Figure 5 As shown, the terminal 5 in this embodiment includes a processor 50, a memory 51, and a computer program 52 stored in the memory 51 and executable on the processor 50. When the processor 50 executes the computer program 52, it implements the steps in the various embodiments of the optimization methods for typical wind and solar load scenarios described above, for example... Figure 1 Steps 101 to 103 are shown.

[0101] For example, computer program 52 can be divided into one or more modules / units, one or more of which are stored in memory 51 and executed by processor 50 to complete the present invention. One or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 52 in terminal 5.

[0102] Terminal 5 may include, but is not limited to, a processor 50 and a memory 51. Those skilled in the art will understand that... Figure 5 This is merely an example of terminal 5 and does not constitute a limitation on terminal 5. It may include more or fewer components than shown, or combine certain components, or different components. For example, the terminal may also include input / output devices, network access devices, buses, etc.

[0103] The processor 50 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0104] The memory 51 can be an internal storage unit of the terminal 5, such as the hard disk or memory of the terminal 5. The memory 51 can also be an external storage device of the terminal 5, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the terminal 5. Furthermore, the memory 51 can include both internal storage units and external storage devices of the terminal 5. The memory 51 is used to store computer programs and other programs and data required by the terminal. The memory 51 can also be used to temporarily store data that has been output or will be output.

[0105] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0106] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0107] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0108] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0109] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0110] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0111] If integrated modules / units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above embodiments of the optimization methods for typical wind and solar load scenarios. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0112] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A wind, light and load typical scene optimization method, characterized in that, The application relates to a method for generating a flexible regulation scheme of a multi-energy coupling system. The method comprises the following steps: establishing a mathematical model of core equipment in the multi-energy coupling system to represent output characteristics, energy conversion efficiency and operation boundary conditions of the equipment; based on the mathematical model, using an improved Gaussian mixture model and an improved expectation maximization algorithm, combining the temporal and spatial correlation characteristics of wind and light output and load, and generating a typical scenario set with probability distribution fitting actual conditions; 2. The wind, light and load typical scenario optimization method according to claim 1, characterized in that, minimizing the total cost of the system as a target, combining the operation boundary conditions of the core equipment and the typical scenario set, constructing an optimization scheduling model containing multi-energy supply and demand balance constraints, and linearizing and solving the optimization scheduling model to obtain a flexible regulation scheme of the multi-energy coupling system. Based on the mathematical model, an improved Gaussian mixture model and an improved expectation maximization algorithm are used to generate a typical scenario set with probability distribution fitting actual conditions, including: obtaining historical time series data of the core equipment; according to the historical time series data, based on the mutual assistance between hidden variables, a hidden variable mutual assistance Gaussian mixture model is constructed; according to the historical time series data, the hidden variable mutual assistance Gaussian mixture model is trained by using an improved expectation maximization algorithm to obtain core parameters of the hidden variable mutual assistance Gaussian mixture model, the core parameters are substituted into the hidden variable mutual assistance Gaussian mixture model to obtain a target hidden variable mutual assistance Gaussian mixture model; using the target hidden variable mutual assistance Gaussian mixture model as a probability distribution template, a Markov chain Monte Carlo is used to generate wind-light-load time series trajectories conforming to the probability distribution template distribution; 3. The wind, light and load typical scenario optimization method according to claim 2, characterized in that, a K-means clustering algorithm is used to perform dimensionality reduction clustering processing on the wind-light-load time series trajectories to obtain a typical scenario set with probability distribution fitting actual conditions. ； in, Indicates the first The data point at the th th The values ​​of the Gaussian components, Indicates the first The data point belongs to the th data point One Gaussian component, Indicates the first The data point at the th th The values ​​of the Gaussian components, Indicates the first The data point belongs to the th data point One Gaussian component, Indicates belonging to the first Data points with Gaussian components Depends on the first Data points with Gaussian components The probability, Indicates belonging to the first Data points of Gaussian components Dependent on data points The probability, Indicates the first The data point belongs to the th data point The probability of a Gaussian component, Indicates the first The data point belongs to the th data point The probability of a Gaussian component.

4. The wind, light and load typical scenario optimization method according to claim 3, characterized in that, The mutual assistance between hidden variables is expressed by conditional probability as follows: According to the historical time series data, the hidden variable mutual assistance Gaussian mixture model is trained by using an improved expectation maximization algorithm to obtain core parameters of the hidden variable mutual assistance Gaussian mixture model, including: the historical time series data is inputted, and core parameters in the hidden variable mutual assistance Gaussian mixture model are initialized; based on the dependence relationship between adjacent time hidden variables and the hidden variable mutual assistance Gaussian mixture model, the responsibility degree of each data point in the historical time series data is calculated; according to the responsibility degree, the core parameters are updated respectively to obtain updated core parameters; a new hidden variable mutual assistance Gaussian mixture model is obtained based on the updated core parameters; 5. The wind, light and load typical scenario optimization method according to claim 4, characterized in that, whether the absolute value of the difference between the updated core parameters and the corresponding core parameters before updating meets a preset condition is detected, if the absolute value of the difference does not meet the preset condition, the new hidden variable mutual assistance Gaussian mixture model is taken as a current hidden variable mutual assistance Gaussian mixture model, and the step of calculating the responsibility degree of each data point in the historical time series data based on the dependence relationship between adjacent time hidden variables and the hidden variable mutual assistance Gaussian mixture model and subsequent steps are executed until the absolute value of the difference meets the preset condition, and the current core parameters are outputted. The responsibility degree of each data point in the historical time series data is calculated based on the dependence relationship between adjacent time hidden variables, including: According to computing a responsibility degree for each data point in the historical time series data; wherein, denotes the responsibility considering the latent variable mutual assistance, denotes the weight of the th Gaussian component in the latent variable mutual assistance Gaussian mixture model, denotes the th data point, denotes the mean vector of the th Gaussian component, denotes the covariance matrix of the th Gaussian component, denotes the latent variable mutual assistance Gaussian mixture model, denotes the state identification of the data point.

6. The wind, light and load typical scenario optimization method according to claim 5, characterized in that, According to the responsibility degree, the core parameters are updated respectively to obtain updated core parameters, including: According to updating the core parameters respectively; wherein, represents the expected number of samples of the th Gaussian component, represents the total amount of data of the historical time series data.

7. The wind, light and load typical scenario optimization method according to any one of claims 1-6, characterized in that, The objective function included in the optimization scheduling model is constructed with the minimum sum of energy market interaction cost, equipment asset depreciation fee, facility maintenance expense, energy consumption cost in unit start-stop process, power grid purchase and sale electricity fee and load response regulation cost as the target; The multi-energy supply and demand balance constraint includes: power supply and demand dynamic balance constraint, heat supply and demand dynamic balance constraint and cold supply and demand dynamic balance constraint; The optimization scheduling model further includes: equipment safe operation constraint, flexible load constraint and demand response constraint, wherein the equipment safe operation constraint includes unit rated capacity constraint, unit power ramping rate constraint and energy storage system operation state boundary constraint, and the energy storage system operation state boundary constraint includes state of charge dynamic boundary constraint, unit capacity dynamic constraint of heat energy storage system and unit capacity dynamic constraint of cold energy storage system.

8. The wind, light and load typical scenario optimization method according to claim 7, characterized in that, The objective function is ; wherein, represents the total cost of the system, represents the cost of energy market interaction, represents the depreciation cost of equipment assets, represents the facility maintenance expenditure, represents the energy consumption cost of unit start-stop process, represents the cost of grid electricity purchase and sale, represents the cost of load response regulation; ； wherein, represents the electricity purchase amount of the power grid in the time period , represents the electricity price of the power grid in the time period , represents the heat resource input amount of the heat grid in the time period , represents the heat energy price of the heat grid in the time period , represents the energy output of the combined cooling heating and power (CCHP) unit in the time period , represents the energy conversion efficiency of the CCHP system in the time period , represents a conversion parameter of the low heat value of natural gas, represents the gas market purchase price in the time period ; ； in, Indicates time period Inner Estimated net asset value of each generating unit Indicates the first Total initial purchase cost of each unit Indicates the first The residual depreciation ratio of each unit Indicates the first The expected remaining runtime of each unit. Indicates the total number of schedulable generating units within the system; ； wherein, denotes a time period a first real-time power output of the first denotes a unit power operation and maintenance cost parameter of the first unit ； wherein, represents the comprehensive loss cost of the start-stop process of the th unit, represents the time period represents the operation state logical identification parameter of the th unit, represents the time period represents the operation state logical identification parameter of the th unit; ； wherein, represents the time period the virtual power plant's electricity purchase input power from the grid, represents the time period the virtual power plant's electricity purchase market rate of the grid, represents the time period the virtual power plant's electricity sale output power to the grid, represents the time period the virtual power plant's electricity sale settlement unit price; ； wherein, denotes a time period a dynamic incentive compensation rate for demand response, denotes a time interval a dispatchable reduction capacity of interruptible load resources.

9. The wind, light and load typical scenario optimization method according to claim 8, characterized in that, The power supply and demand dynamic balance constraint is ； wherein, denotes a time period internal photovoltaic array output power, denotes a time period internal wind turbine output power, denotes a time period internal combined cooling heating and power unit output power, denotes a time period internal electric energy storage device net output power, denotes a time period internal electric power load demand baseline value, denotes a time period internal adjustable load response power, denotes a time period internal heat pump unit electric power input, denotes a heat pump heat-to-power conversion efficiency coefficient, denotes a time period internal electric heating unit rated power, denotes an electric boiler thermal efficiency parameter, denotes a time period internal electric refrigeration device operating power, denotes a refrigeration unit energy efficiency ratio, denotes a time period internal energy storage system discharge power, denotes a time period internal energy storage system charge power; The heat supply and demand dynamic balance constraint is ； wherein, denotes a time period heating power of the internal cogeneration unit, denotes a time period heating power of the internal heat pump unit, denotes a time period heating power output value of the internal unit, denotes a time period net heating power of the internal heat storage device, denotes a time period baseline value of the internal thermal load demand, denotes a time period adjustable thermal load reduction capacity of the internal unit, denotes a time period heating power of the internal heat storage device, denotes a time period heat storage power of the internal heat storage device; The cold supply and demand dynamic balance constraint is ； wherein, denotes a time period the internal cooling heat and power cogeneration unit refrigeration output, denotes a time period the internal heat pump unit refrigeration power, denotes a time period the internal heat power output value, denotes a time period the internal cold storage device net cold release power, denotes a time period the internal cold energy demand baseline value, denotes a time period the internal cold storage device cold release power, denotes a time period the internal cold storage device cold storage power; The unit rated capacity constraint is ; wherein, represents the maximum rated output power of the nth unit. The unit power ramp rate constraint limits to ; wherein, denotes a power ramp rate constraint limit for the th unit, denotes the real-time supply output of the th unit over the time period denotes the real-time supply output of the th unit over the time period denotes a power ramp rate constraint limit for the th unit. The state of charge dynamic boundary constraint is ； wherein, denotes a time period denotes an energy storage battery state of charge, denotes an energy storage battery state of charge safety lower limit, denotes an energy storage battery state of charge safety upper limit, denotes an energy storage battery dispatch period initial state of charge, denotes an energy storage battery dispatch period termination state of charge, denotes an energy storage battery maximum charge power limit, denotes an energy storage battery maximum discharge power limit, denotes an energy storage battery rated capacity, denotes an energy storage battery charge energy conversion efficiency coefficient, denotes an energy storage battery discharge energy conversion efficiency coefficient, denotes a dispatch time interval; The unit capacity dynamic constraint of heat energy storage system is ； wherein, represents the maximum heat storage power limit of the heat storage tank, represents the maximum heat release power limit of the heat storage tank, represents the rated capacity of the heat storage tank, represents the heat storage energy conversion efficiency coefficient of the heat storage tank, represents the heat release energy conversion efficiency coefficient of the heat storage tank; The unit capacity dynamic constraint of cold energy storage system is ； wherein, represents the maximum ice storage tank cold charging power limit, represents the maximum ice storage tank cold discharging power limit, represents the ice storage tank rated capacity, represents the ice storage tank cold charging energy conversion efficiency coefficient, represents the ice storage tank cold discharging energy conversion efficiency coefficient; The flexible load constraint is ; wherein, denotes a flexible load actual operating power of the flexible load under a time period scenario actual operating power of the flexible load under a time period denotes a flexible load minimum allowable power of the flexible load under a time period scenario denotes a flexible load maximum allowable power of the flexible load under a time period scenario denotes a flexible load power drop rate limit of the flexible load denotes a flexible load power rise rate limit of the flexible load denotes a flexible load power value of the flexible load under a time period scenario power value of the flexible load under a time period scenario cumulative energy of the flexible load under a time period scenario cumulative energy of the flexible load under a time period scenario minimum required energy demand of the flexible load denotes a scheduling cycle total time period number; The demand response constraint is ; in, Indicates time period Adjusted load power after implementing price-based demand response. Indicates time period Baseline load power within, Indicates time period The maximum allowable load adjustment range within the range, Indicates the first Scheduled capacity reduction for individual users This indicates the permissible reduction percentage of interruptible load in a virtual power plant. Indicates the first Total interruptible load capacity for each user.

10. A terminal comprising a memory for storing a computer program and a processor for invoking and running the computer program stored in the memory, characterized in that, The processor implements the steps of the wind-solar and load typical scene optimization method according to any one of claims 1 to 9 when executing the computer program.

Images