Method and system for optimizing electricity consumption of producer and consumer based on load clustering algorithm

Abstract

The invention discloses a method and a system for optimizing power consumption of a producer and a consumer based on a load clustering algorithm, and the method comprises the steps: collecting power consumption data, load data and photovoltaic power generation data of the producer and the consumer in a preset time period, and generating load curve data according to a time sequence; performing feature extraction on the load curve data, and classifying load features of each producer based on a clustering algorithm to obtain load modes of different energy demand responses; based on the load modes obtained through classification, the user self-sufficiency rate and the power product total cost of the producer and the consumer under various scheduling models are calculated respectively, and the user self-sufficiency rate is used for representing the renewable energy source proportion in the user load; and calculating a comprehensive index value according to the self-sufficiency rate of the user and the total cost of the power product, and selecting the scheduling model with the highest comprehensive index value to optimize the power consumption of the producer and the consumer. The optimization method provided by the invention can be used for improving the utilization rate of the spontaneous energy by the disappearing people.

Description

Technical Field

[0001] The present invention relates to prosumer self-consumption modeling and energy management technology, and in particular to a prosumer electricity optimization method and system based on a load clustering algorithm. Background Art

[0002] The flow of electricity in the traditional power system is undergoing structural changes due to the increased use of RES (Renewable Energy Sources). In order to achieve efficient operation of self-consumers combined with distributed energy resources, developments on the demand side are changing the way electricity is consumed and supplied. Consumers who produce more electricity than they consume in a specific period and sell their surplus electricity to the grid have become prosumers. In addition, some prosumers may invest in energy storage systems to store their surplus electricity for subsequent self-use.

[0003] However, rooftop solar PV (Photovoltaic)-home BESS (Battery Energy Storage System) is still expensive in many countries. On the one hand, due to the capacity conditions of prosumers, they will inevitably face the risk of electricity price fluctuations when participating in the day-ahead electricity wholesale market transactions; on the other hand, given that prosumers are usually at the bottom of the system architecture, it is difficult for them to be directly and accurately controlled and dispatched by the upper-level dispatching system. The cost of installing photovoltaic home BESS is high. Another option that meets economic feasibility is to combine energy storage technology with power dispatch strategies, such as residential battery energy storage, virtual battery systems, and adjustable loads in DSM (Demand Side Management) systems. Existing studies have oversimplified the load demand of household solar cells and the power consumption diagram of photovoltaic output.

[0004] At present, the existing technology has the following deficiencies:

[0005] 1) Although the existing methods adopt prediction and optimization strategies to control PV-BESS, the improvement of SSR is still very low, which shows that the current optimization strategy is not effective in improving SSR.

[0006] 2) Without the optimization algorithm, the annual waste of excess energy generated by user generators is as high as 45.5%, which shows that the existing system has a very serious energy waste phenomenon without effective management. This is not only a waste of renewable energy, but also increases the user's dependence on the external power grid.

[0007] How to overcome the above problems is an urgent problem to be solved. Summary of the invention

[0008] In order to solve the deficiencies in the prior art, the present invention proposes a method and system for optimizing electricity consumption of prosumers based on a load clustering algorithm. Through the prosumer self-consumption optimization scheduling model and the peak shaving algorithm scheduling model, the expected self-consumption energy of the prosumer is evaluated, and it is classified into six load pattern groups through unsupervised machine learning (i.e., K-means clustering and Gaussian mixture model) to display the load demand derived from the collected data. After selecting a peak reduction algorithm with high efficiency, it is applied to the established prosumer mathematical model. The model has a benefit optimization function, can increase the self-consumption rate and self-sufficiency rate, and can improve the utilization rate of renewable energy.

[0009] The present invention adopts the following technical solution.

[0010] The first aspect of the present invention provides a method for optimizing electricity consumption of prosumers based on a load clustering algorithm, comprising the following contents:

[0011] Collect electricity consumption data, load data and photovoltaic power generation data of producers and consumers within a preset time period, and generate load curve data in time series;

[0012] Extracting features from the load curve data to obtain load features of each prosumer, the load features comprising at least one of the following: consumption pattern, peak demand time, change in power consumption, load factor, peak load position, negative peak load position, and standard deviation;

[0013] Based on the clustering algorithm, the load characteristics of each prosumer are classified to obtain the load patterns in response to different energy demands;

[0014] Based on the classified load patterns, the user self-sufficiency rate and the total cost of power products under various dispatch models of prosumers are calculated respectively. The user self-sufficiency rate is used to characterize the proportion of renewable energy in the user load.

[0015] The comprehensive index value is calculated based on the user's self-sufficiency rate and the total cost of electricity products, and the dispatching model with the highest comprehensive index value is selected to optimize the electricity consumption of producers and consumers.

[0016] Optionally, feature extraction is performed on the load curve data based on the established Gaussian mixture model, including:

[0017] Each data point on the load curve is modeled as a combination of multiple Gaussian distributions, each of which has a mean, covariance, and corresponding weight. A Gaussian mixture model is established for the data point x according to the following formula:

[0018]

[0019] Where x is a data point in the load curve, MTo model x as the number of Gaussian distributions in a combination of Gaussian distributions, μ i is the mean of the ith Gaussian distribution, C i is the covariance matrix of the i-th Gaussian distribution, π i is the weight corresponding to the i-th Gaussian distribution, p i (x; μ i ,C i ) is the mean value μ i , the covariance is C i Gaussian distribution of

[0020] The Gaussian mixture model is used to assign each data point in the load curve to the corresponding Gaussian distribution, and the characteristics of each data point are obtained as the load characteristics of each prosumer.

[0021] Optionally, the load characteristics of each prosumer are classified based on a clustering algorithm to obtain load patterns in response to different energy demands, including:

[0022] The objective function of the K-means clustering algorithm is established according to the following formula:

[0023]

[0024] in,

[0025] m is the number of data points, indicating the total number of load characteristic data points involved in clustering.

[0026] n is the number of clusters, which means that m data points are divided into n clusters.

[0027] x i is the ith data point,

[0028] ||x i -c a || 2 is the relationship between each data location and each cluster centroid c a The distance between

[0029] Based on the K-means clustering algorithm, the categories of the clusters to which the data points belong are adjusted until the objective function reaches the minimum value, and the load patterns that respond to different energy demands are obtained.

[0030] Optionally, before extracting features from the load curve to obtain the load features of each prosumer, the method further includes:

[0031] The data in the load curve is preliminarily classified according to the time scenario to obtain the load curve data corresponding to multiple time scenarios.

[0032] Optionally, the comprehensive index value is calculated based on the user's self-sufficiency rate and the total cost of power products according to the following formula:

[0033] L=w 1 ×SSR+w 2 ×LCOE 2

[0034] Among them, L is the comprehensive index value, SSR is the user self-sufficiency rate, LCOE is the total cost of power products, w 1 and w 2 are the weights of user self-sufficiency rate and total cost of electricity products respectively.

[0035] Optionally, the user self-sufficiency rate is calculated according to the following formula:

[0036]

[0037] Among them, SSR is the user self-sufficiency rate, E du E is the energy directly supplied to the load by photovoltaic power generation within the preset time period. dis is the actual discharge amount of the battery within the preset time period, E load is the total annual user-side load of the system, E PV It is the amount of electricity generated by the photovoltaic power generation system in a given period of time, DOD is the depth of discharge of the battery, and C is the rated capacity of the battery.

[0038] Optionally, the total cost of electricity products is calculated as follows:

[0039]

[0040] CRF=i(1+i n ) / ((1+i n )-1)

[0041]

[0042] Among them, LCOE is the total cost of electricity products, which is used to represent the cost of producing one kilowatt-hour of electricity, A is the cost generated by the scheduling model, and E bygrid ×γ retail Refers to the market value fuel cost, E sellsback ×γ sell To provide value for electricity sales, E load is the total annual user-side load of the system, I PV is the photovoltaic system capacity, I BES is the capacity of the battery energy storage system, OM is the proportion of annual operation and maintenance costs to the cost generated by the dispatch model, CRF is the capital recovery factor, n is the service life of the photovoltaic power generation equipment, i is the interest rate, and E sellsback E is the remaining energy of the photovoltaic system after self-use of energy. PV,DC is the total amount of DC power generated by the photovoltaic system, Eff BESis the charging efficiency of the battery energy storage system, Eff inv is the inverter efficiency, E sum is the total energy supported by the battery energy storage system, E sc,dc is the self-used energy with battery energy storage system, E sc,dc,0 is the self-consumed energy without battery energy storage system, E SC is the energy self-consumption when there is a battery energy storage system, E SC,initial It is the self-consumption of energy without battery energy storage system.

[0043] Optionally, the scheduling model includes a prosumer self-consumption optimization scheduling model, the prosumer self-consumption optimization scheduling model includes a solar rooftop photovoltaic maximum self-consumption algorithm and an optimal battery energy storage system scheduling algorithm, the scheduling of the battery energy storage system capacity is set to the prosumer maximum energy self-consumption, when the power generated by the photovoltaic system is greater than the household demand, the battery will continue to charge, otherwise, the battery will discharge until the minimum allowed level;

[0044] The maximum battery discharge power P in each time interval in the prosumer self-consumption optimization scheduling model dsc,i It is expressed as follows:

[0045]

[0046] In the prosumer self-consumption optimization scheduling model, the battery charging power is maximized P chr,i It is expressed as follows:

[0047]

[0048] The battery state of charge SOC in the prosumer self-consumption optimization scheduling model i It is expressed as follows:

[0049]

[0050] The battery discharge state SOC in the prosumer self-consumption optimization scheduling model i It is expressed as follows:

[0051]

[0052] Among them, SOC i is the battery charge and discharge state at the i-th moment, SOC i-1 is the battery charge and discharge state at the i-1th moment, Eff BES is the charging efficiency of the inverter, P BES,cap It represents the power that cannot be utilized due to the capacity limitation of the battery energy storage system, P BES is the efficiency of the battery energy storage system, δt is the time interval, p charge is the charging power of the energy storage device, pdis is the discharge power of the energy storage device, η charge is the charging efficiency of the energy storage device, η dis is the discharge efficiency of the energy storage device, E EES The amount of charge or discharge during a time interval.

[0053] Optionally, the scheduling model includes a peak shaving algorithm scheduling model, wherein the peak shaving algorithm scheduling model is a scheduling control for performing peak shaving according to the electric energy generated by the photovoltaic system and an energy threshold.

[0054] The energy threshold in the peak shaving algorithm scheduling model is expressed as follows:

[0055] E th =(∑P Th )δ t -E BES,cap

[0056]

[0057] The discharge peak shaving Φ in the peak shaving algorithm model dsc,i Calculated as follows:

[0058] Φ dsc,i =[min(+P Th,i ,P max,dsc,i )]

[0059] The battery charging power Φ in the peak shaving algorithm model chr,i Calculated as follows:

[0060] Φ chr,i =[min(-P Th,i ,P max,dsc,i )]

[0061] Among them, E th is the energy threshold, P Th represents the rate of electrical energy generation, δ t Indicates the time interval, E BES,cap represents the energy that cannot be utilized due to the capacity limitation of the battery energy storage system, P Th,i Indicates the amount of power that needs to be released from the battery in order to reduce the peak load during the time interval. It refers to the part of the electricity generated by the photovoltaic system that can meet the load demand after being inverted. PV,DC Indicates the DC power actually generated by the photovoltaic system; P max,dsc,i It is the maximum allowable power when the battery is discharged to cut the peak during the time interval.

[0062] Optionally, when the peak shaving algorithm model performs dispatching control of peak shaving according to the electric energy generated by the photovoltaic system and the energy threshold, it includes:

[0063] The electrical energy E generated by the photovoltaic system PV (t) is less than the energy threshold E th In the case of PV (t) is used to satisfy the load L(t); if E PV (t) is less than L(t), then E PV (t) After the load is applied, the insufficient part is supplemented by the battery discharge, and the discharge amount is L(t)-E PV (t), otherwise the remaining power E PV (t)-L(t) is sent to the public grid, and the peak shaving amount of battery discharge is the load L(t) and E PV The difference of (t);

[0064] The electrical energy E generated by the photovoltaic system PV (t) is greater than or equal to the energy threshold E th In the case of E PV (t) is equal to the energy threshold E th The amount of electricity is used to meet the load or to be fed into the public grid; if L(t) is less than E th , then E th -L(t) is fed into the public grid, otherwise it is discharged by the battery L(t)-E th Or purchase electricity L(t)-E th To meet the load;

[0065] E PV (t) is equal to the energy threshold E th The amount of electricity used to meet the load or sent to the public grid is the energy E PV (t)-E th Stored in the battery, the storage capacity is Among them, C is the rated capacity of the battery.

[0066] A second aspect of the present invention provides a prosumer electricity optimization system of a prosumer electricity optimization method, the system comprising:

[0067] A collection module is used to collect electricity consumption data, load data and photovoltaic power generation data of prosumers within a preset period of time, and generate load curve data in time series;

[0068] A feature extraction module, for extracting features from the load curve data to obtain the load features of each prosumer, wherein the load features include at least one of the following: consumption pattern, peak demand time, change in power consumption, load factor, peak load position and negative peak load position and standard deviation;

[0069] A classification module is used to classify the load characteristics of each prosumer based on a clustering algorithm to obtain load patterns that respond to different energy demands;

[0070] A calculation module is used to calculate the user self-sufficiency rate and the total cost of power products under various scheduling models of prosumers based on the classified load patterns, wherein the user self-sufficiency rate is used to characterize the proportion of renewable energy in the user load;

[0071] The dispatch optimization module is used to calculate the comprehensive index value based on the user's self-sufficiency rate and the total cost of power products, and select the dispatch model with the highest comprehensive index value to optimize the electricity consumption of producers and consumers.

[0072] The third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is loaded into the processor, the method for optimizing electricity consumption of producers and consumers based on a load clustering algorithm is implemented.

[0073] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for optimizing electricity consumption of producers and consumers based on a load clustering algorithm.

[0074] Compared with the prior art, the beneficial effects of the present invention include at least:

[0075] The existing technology mainly links prosumers with photovoltaic home battery energy storage systems, and only uses the average historical time interval load curves of various countries without analyzing the load pattern. The present invention identifies and analyzes the load pattern through clustering and demand curve analysis, using K-means clustering and Gaussian mixture model technology to address these limitations.

[0076] The present invention constructs a prosumer cost optimization model using a data exploration and analysis method (i.e., K-means clustering algorithm and Gaussian mixture model to analyze power load demand) for self-consumption of BESS prosumers.

[0077] This paper proposes a new strategy for the prosumer dispatch model for all load modes to increase the income of surplus energy: a prosumer self-consumption optimization dispatch model and a peak shaving algorithm dispatch model are compared. The model with the prosumer self-consumption optimization dispatch method provides moderate profits, while the model with the peak shaving algorithm provides higher profits and self-sufficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0079] Figure 1 It is a simplified diagram of a prosumer electricity optimization model provided by an embodiment of the present invention;

[0080] Figure 2 It is a flow chart of a method for clustering load patterns using GMMs and K-means algorithms provided by an embodiment of the present invention;

[0081] Figure 3 It is a schematic diagram of load patterns of different energy demand responses obtained by a K-means algorithm provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0082] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only embodiments of a part of the present invention, rather than all embodiments. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the protection scope of the present invention.

[0083] SSR (Self Sufficiency Rate) refers to the percentage of total energy consumption that is met by the prosumer's own electricity generation, representing the proportion of renewable energy in the user's load.

[0084] Home batteries are stationary energy storage devices used to support photovoltaic power generation. The size of the battery energy storage depends on the required storage energy and the relative amount of system load. Larger battery size can increase capacity and extend service life by avoiding severe deep discharge.

[0085] Figure 1This is an implementation environment diagram of the present invention, the implementation environment includes the photovoltaic power generation system and battery inside the prosumer's home, and the external environment is the power grid. By setting up a photovoltaic power generation system in the prosumer's home, the photovoltaic power generates direct current and then is transmitted to the power grid or stored in the battery through the inverter. The output current destination is determined after the model algorithm analysis. The model determines whether the output current is directly connected to the power grid to obtain economic benefits or stored in the battery by analyzing the peak and valley periods of electricity prices and the prosumer's own load conditions. The current stored in the battery can be sent to the power grid during the peak period of electricity prices to obtain higher benefits or used for self-consumption by the prosumer. The load condition data generated thereby is returned to the model as memory data.

[0086] Combination Figures 1 to 3 As shown, embodiment 1 of the present invention provides a method for optimizing electricity consumption of prosumers based on a load clustering algorithm, comprising the following steps:

[0087] Step 1: Collect electricity consumption data, load data and photovoltaic power generation data of producers and consumers within a preset time period, and generate a load curve according to the time series.

[0088] Specifically, the electricity consumption data is the daily electricity consumption of the prosumer within a preset time period, the load data is the load of the energy storage system, and the photovoltaic power generation data is the photovoltaic power generation within a preset time period.

[0089] Specifically, the preset time period may be one year, one month, or three months, and the present invention does not limit the specific value of the preset time period.

[0090] Optionally, before step 2, the method further includes: preliminarily classifying the data in the load curve according to the time scenario to obtain load curve data corresponding to multiple time scenarios.

[0091] Specifically, the electricity consumption data is marked according to the electricity demand pattern, where the electricity demand pattern refers to the characteristics and regularities of the user's demand for electricity in different time scenarios.

[0092] In this embodiment, two specific electricity demand modes, working days and non-working days, are considered. When collecting electricity consumption data, load conditions, and photovoltaic power generation data of households with photovoltaic power generation in a preset time period, the impact of these two specific electricity demand modes on the overall data is considered at the same time. For example, when recording the timestamp of electricity consumption data, it is marked whether it is a working day or a non-working day, so as to analyze the electricity consumption behavior under different modes later.

[0093] Further, Preliminary Calendar Segmentation: A preliminary segmentation is first performed based on standard calendar dates, where weekends are easily distinguishable. Anomaly Detection: Using the GMM-processed data, energy consumption anomalies that deviate from typical weekday patterns are identified. These anomalies indicate possible holidays or special non-working days. Checking against public holidays: Detected anomalies are checked against the known public holiday calendar to confirm their holiday status. Adjustment for unofficial holidays: Additional adjustments are made for locally observed holidays or unique events that are not recorded in the standard calendar but are indicated by abnormal consumption patterns. This process ensures a comprehensive and accurate classification of holidays, resulting in more precise clustering based on energy usage patterns.

[0094] Optionally, before step 2, the method further includes preprocessing the load curve data, wherein the preprocessing includes EDA (Exploratory Data Analysis) and data cleaning.

[0095] In this embodiment, EDA and data cleaning are both existing technologies, and the present invention will not elaborate on them.

[0096] Step 2: Extract features from the load curve data to obtain the load characteristics of each producer and consumer, wherein the load characteristics include at least one of the following: consumption pattern, peak demand time, change in electricity consumption, load factor, peak load position, negative peak load position, and standard deviation.

[0097] Optionally, in step 2, extracting features from the load curve data of each user based on the established GMMS (Gaussian Mixture Model) includes:

[0098] Each data point in the load curve is modeled as a combination of multiple Gaussian distributions. Each distribution has a mean, covariance, and corresponding weight. A Gaussian mixture model is established for the data point x according to the following formula:

[0099]

[0100] Where x is a data point in the load curve, M represents the number of Gaussian distributions in the combination that models x as a Gaussian distribution, μ i is the mean of the ith Gaussian distribution, C i is the covariance matrix of the i-th Gaussian distribution, π i is the weight corresponding to the i-th Gaussian distribution, p i (x; μ i ,C i ) is the mean value μ i , the covariance is C i Gaussian distribution of

[0101] The Gaussian mixture model is used to assign each data point in the load curve to the corresponding Gaussian distribution, and the characteristics of each data point are obtained as the load characteristics of each prosumer.

[0102] In this embodiment, the load characteristics and their corresponding values ​​are shown in Table 1.

[0103] Table 1

[0104]

[0105]

[0106] In Table 1, consumption patterns can be classified in more detail according to actual conditions, such as ordinary residents, high-energy-consuming residents, etc. The load factor represents the ratio of average load to peak load, reflecting the stability of the load, and the standard deviation reflects the discreteness of the load data.

[0107] Through this operation, characteristic information of each data point under different Gaussian distributions is obtained, including consumption patterns, peak demand time, usage changes, load factors, peak load positions and negative peak load positions, and standard deviations. This embodiment takes into account the disordered and ordered properties of the data, and by establishing a Gaussian mixture model, the subsequent clustering algorithm improves efficiency and reduces time complexity when processing large-scale data, and the clustering results are more stable, thereby further improving the accuracy and effectiveness of load pattern classification.

[0108] Step 3: Classify the load characteristics of each producer and consumer based on the clustering algorithm to obtain the load patterns that respond to different energy demands.

[0109] Optionally, the load mode includes at least one of the following: an industrial load mode, a commercial load mode, and a residential load mode.

[0110] Optionally, the load characteristics of each prosumer are classified based on a clustering algorithm to obtain load patterns in response to different energy demands, including:

[0111] The objective function of the K-means clustering algorithm is established according to the following formula:

[0112]

[0113] in,

[0114] m is the number of data points, indicating the total number of load characteristic data points involved in clustering.

[0115] n is the number of clusters, which means that m data points are divided into n clusters.

[0116] x iis the ith data point,

[0117] ||x i -c a || 2 is the relationship between each data location and each cluster centroid c a The distance between

[0118] Based on the K-means clustering algorithm, the categories of the clusters to which the data points belong are adjusted until the objective function reaches the minimum value, and the load patterns that respond to different energy demands are obtained.

[0119] Specifically, the K-means clustering algorithm first collects separate main load patterns, and then clusters load patterns with similar characteristics, including load factor, peak and negative peak positions, and standard deviation, which can be regarded as extracted features and are also related to different load patterns. Load factor, peak, negative peak position, and standard deviation help to characterize and distinguish different load patterns. Then, by continuously adjusting the category to which the data point belongs, the objective function is minimized. Through this operation, load patterns that respond to different energy demands are obtained. This can remove data deviations caused by different power consumption patterns during the calculation process, effectively improve the accuracy of data collation, and achieve detailed division of load patterns, providing a basis for accurately analyzing the timing behavior of self-generation, power consumption, and power sales of prosumers, thereby maximizing the benefits of self-generation.

[0120] Step 4: Based on the classified load patterns, calculate the user self-sufficiency rate and the total cost of electricity products under various scheduling models of producers and consumers, where the user self-sufficiency rate is used to characterize the proportion of renewable energy in the user load.

[0121] Optionally, the comprehensive index value is calculated based on the user's self-sufficiency rate and the total cost of power products according to the following formula:

[0122] L=w 1 ×SSR+w 2 ×LCOE 2

[0123] Among them, L is the comprehensive index value, SSR is the user self-sufficiency rate, LCOE is the total cost of power products, w 1 and w 2 are the weights of user self-sufficiency rate and total cost of electricity products respectively.

[0124] If the self-sufficiency rates are the same, the model with lower cost is better; if the costs are the same, the model with higher self-sufficiency rate is better. This results in a cost optimization solution with more accurate cost prediction and more obvious optimization effect.

[0125] It should be noted that, in the embodiment of the present invention, 1 and w 2Both are 0.5. In practical applications, the weight values ​​can be set according to actual emphasis.

[0126] Optionally, the user self-sufficiency rate is calculated according to the following formula:

[0127]

[0128] Among them, SSR is the user self-sufficiency rate, E du E is the energy directly supplied to the load by photovoltaic power generation within the preset time period. dis is the actual discharge amount of the battery within the preset time period, E load is the total annual user-side load of the system, E PV It is the amount of electricity generated by the photovoltaic power generation system in a given period of time, DOD is the depth of discharge of the battery, and C is the rated capacity of the battery.

[0129] Optionally, the total cost of electricity products is calculated as follows:

[0130]

[0131] CRF=i(1+i n ) / ((1+i n )-1)

[0132]

[0133] Among them, LCOE is the total cost of electricity products, which is used to represent the cost of producing one kilowatt-hour of electricity, A is the cost generated by the scheduling model, and E bygrid ×γ retail Refers to the market value fuel cost, E sellsback ×γ sell To sell electricity and return value, E load is the total annual user-side load of the system, I PV is the photovoltaic system capacity, I BES is the capacity of the battery energy storage system, OM is the proportion of annual operation and maintenance costs to the cost generated by the dispatch model, CRF is the capital recovery factor, n is the service life of the photovoltaic power generation equipment, i is the interest rate, and E sellsback E is the remaining energy of the photovoltaic system after self-use of energy. PV,DC is the total amount of DC power generated by the photovoltaic system, Eff BES is the charging efficiency of the battery energy storage system, Eff inv is the inverter efficiency, E sum is the total energy supported by the battery energy storage system, E sc,dc is the self-used energy with battery energy storage system, E sc,dc,0 is the self-consumed energy without battery energy storage system, E SC is the energy self-consumption when there is a battery energy storage system, E SC,initialIt is the self-consumption of energy without battery energy storage system.

[0134] Understandably, E. sellsback The amount of energy remaining from the photovoltaic system after the energy required for self-use is the total energy that can be supplied to the public grid.

[0135] Understandably, the electricity that flows from prosumers to the grid can be considered a negative cost, transforming users into revenue-generating prosumers using PV battery energy storage systems.

[0136] It can be understood that LCOE is the total cost of electricity products, which is used to represent the cost of producing one kilowatt-hour of electricity, and represents the net present value of the total cost of electricity products, that is, the operating life of a power station, including construction, operation, fuel and maintenance costs. In order to better fit the actual data, the levelized electricity cost is presented in the form of a square in the model. The square can filter out fluctuations and improve the prediction accuracy of the model.

[0137] Optionally, the scheduling model includes a scheduling model including a prosumer self-consumption optimization scheduling model, the prosumer self-consumption optimization scheduling model includes a solar roof photovoltaic maximum self-consumption algorithm and an optimal battery energy storage system scheduling algorithm, the scheduling of the battery energy storage system capacity is set to the maximum energy self-consumption of the prosumer, when the power generated by the photovoltaic system is greater than the household demand, the battery will continue to charge, otherwise, the battery will discharge until the minimum allowed level;

[0138] The maximum battery discharge power P in each time interval in the prosumer self-consumption optimization scheduling model dsc,i It is expressed as follows:

[0139]

[0140] In the prosumer self-consumption optimization scheduling model, the battery charging power is maximized P chr,i It is expressed as follows:

[0141]

[0142] The battery state of charge SOC in the prosumer self-consumption optimization scheduling model i It is expressed as follows:

[0143]

[0144] The battery discharge state SOC in the prosumer self-consumption optimization scheduling model i It is expressed as follows:

[0145]

[0146] Among them, SOC i is the battery charge and discharge state at the i-th moment, SOC i-1is the battery charge and discharge state at the i-1th moment, Eff BES is the charging efficiency of the inverter, P BES,cap It represents the power that cannot be utilized due to the capacity limitation of the battery energy storage system, P BES is the efficiency of the battery energy storage system, δt is the time interval, p charge is the charging power of the energy storage device, p dis is the discharge power of the energy storage device, η charge is the charging efficiency of the energy storage device, η dis is the discharge efficiency of the energy storage device, E EES The charge / discharge amount during the time interval.

[0147] Optionally, the scheduling model includes a peak shaving algorithm scheduling model, wherein the peak shaving algorithm scheduling model is a scheduling control for performing peak shaving according to the electric energy generated by the photovoltaic system and an energy threshold.

[0148] The energy threshold in the peak shaving algorithm scheduling model is expressed as follows:

[0149] E th =(∑P Th )δ t -E BES,cap

[0150]

[0151] The discharge peak shaving Φ in the peak shaving algorithm model dsc,i Calculated as follows:

[0152] Φ dsc,i =[min(+P Th,i ,P max,dsc,i )]

[0153] The battery charging power Φ in the peak shaving algorithm model chr,i Calculated as follows:

[0154] Φ chr,i =[min(-P Th,i ,P max,dsc,i )]

[0155] Among them, E th is the energy threshold, P Th represents the rate of electrical energy generation, δ t Indicates the time interval, E BES,cap represents the energy that cannot be utilized due to the capacity limitation of the battery energy storage system, P Th,i Indicates the amount of power that needs to be released from the battery in order to shave the peak during the time interval. It refers to the part of the electricity generated by the photovoltaic system that can meet the load demand after being inverted.PV,DC Indicates the DC power actually generated by the photovoltaic system; P max,dsc,i It is the maximum allowable power when the battery is discharged to cut the peak during the time interval.

[0156] Optionally, when the peak shaving algorithm model performs dispatching control of peak shaving according to the electric energy generated by the photovoltaic system and the energy threshold, it includes:

[0157] The electrical energy E generated by the photovoltaic system PV (t) is less than the energy threshold E th In the case of PV (t) is used to satisfy the load L(t); if E PV (t) is less than L(t), then E PV (t) After the load is applied, the insufficient part is supplemented by the battery discharge, and the discharge amount is L(t)-E PV (t), otherwise the remaining power E PV (t)-L(t) is sent to the public grid, and the peak shaving amount of battery discharge is the load L(t) and E PV The difference of (t);

[0158] The electrical energy E generated by the photovoltaic system PV (t) is greater than or equal to the energy threshold E th In the case of E PV (t) is equal to the energy threshold E th The amount of electricity is used to meet the load or to be fed into the public grid; if L(t) is less than E th , then E th -L(t) is fed into the public grid, otherwise it is discharged by the battery L(t)-E th Or purchase electricity L(t)-E th To meet the load;

[0159] E PV (t) is equal to the energy threshold E th The amount of electricity used to meet the load or sent to the public grid is the energy E PV (t)-E th Stored in the battery, the storage capacity is Among them, C is the rated capacity of the battery.

[0160] It should be noted that the embodiment of the present invention may also be applicable to other scheduling models.

[0161] Embodiment 2 of the present invention provides a prosumer electricity optimization system, which runs the prosumer electricity optimization method based on the load clustering algorithm as described in Embodiment 1, and the system includes:

[0162] A collection module is used to collect electricity consumption data, load data and photovoltaic power generation data of prosumers within a preset period of time, and generate load curve data in time series;

[0163] A feature extraction module, for extracting features from the load curve data to obtain the load features of each prosumer, wherein the load features include at least one of the following: consumption pattern, peak demand time, change in power consumption, load factor, peak load position and negative peak load position and standard deviation;

[0164] A classification module is used to classify the load characteristics of each prosumer based on a clustering algorithm to obtain load patterns that respond to different energy demands;

[0165] A calculation module is used to calculate the user self-sufficiency rate and the total cost of power products under various scheduling models of prosumers based on the classified load patterns, wherein the user self-sufficiency rate is used to characterize the proportion of renewable energy in the user load;

[0166] The dispatch optimization module is used to calculate the comprehensive index value based on the user's self-sufficiency rate and the total cost of power products, and select the dispatch model with the highest comprehensive index value to optimize the electricity consumption of producers and consumers.

[0167] Embodiment 3 of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is loaded into the processor, the method for optimizing electricity consumption of producers and consumers based on a load clustering algorithm as described in Embodiment 1 is implemented.

[0168] Embodiment 4 of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for optimizing electricity consumption of producers and consumers based on a load clustering algorithm according to Embodiment 1 is implemented.

[0169] Compared with the prior art, the beneficial effects of the present invention include at least:

[0170] Existing literature mainly associates prosumers with photovoltaic home battery energy storage systems, and only uses the average historical time interval load curves of various countries without analyzing the load pattern. Therefore, the present invention identifies and analyzes the load pattern through clustering and demand curve analysis, using K-means clustering and Gaussian mixture model techniques to address these limitations.

[0171] The present invention constructs a prosumer cost optimization model using a data exploration and analysis method (i.e., K-means clustering algorithm and Gaussian mixture model to analyze power load demand) for self-consumption of BESS prosumers.

[0172] This paper proposes a new strategy for the prosumer dispatch model for all load modes to increase the income of surplus energy: a prosumer self-consumption optimization dispatch model and a peak shaving algorithm dispatch model are compared. The model with the prosumer self-consumption optimization dispatch method provides moderate profits, while the model with the peak shaving algorithm provides higher profits and self-sufficiency.

[0173] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0174] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0175] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0176] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0177] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations. The present invention optimizes the existing mobile energy storage equipment, and uses a centralized control center control unit during optimization, and no other systems and equipment need to be protected.

Claims

1. A method for optimizing electricity consumption of prosumers based on load clustering algorithm, characterized in that: Includes the following: Collect electricity consumption data, load data and photovoltaic power generation data of producers and consumers within a preset time period, and generate load curve data in time series; Extracting features from the load curve data to obtain load features of each prosumer, the load features comprising at least one of the following: consumption pattern, peak demand time, change in power consumption, load factor, peak load position, negative peak load position, and standard deviation; Based on the clustering algorithm, the load characteristics of each prosumer are classified to obtain the load patterns in response to different energy demands; Based on the classified load patterns, the user self-sufficiency rate and the total cost of power products under various dispatch models of prosumers are calculated respectively. The user self-sufficiency rate is used to characterize the proportion of renewable energy in the user load. The comprehensive index value is calculated based on the user's self-sufficiency rate and the total cost of electricity products, and the dispatching model with the highest comprehensive index value is selected to optimize the electricity consumption of producers and consumers.

2. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 1, characterized in that: Based on the established Gaussian mixture model, the load curve data is subjected to feature extraction, including: Each data point on the load curve is modeled as a combination of multiple Gaussian distributions, each of which has a mean, covariance, and corresponding weight. A Gaussian mixture model is established for the data point x according to the following formula: in, x is a data point in the load curve, M is the number of Gaussian distributions in the combination that models x as a Gaussian distribution, μ i is the mean of the ith Gaussian distribution, C i is the covariance matrix of the i-th Gaussian distribution, π i is the weight corresponding to the i-th Gaussian distribution, p i (x; μ i ,C i ) is the mean value μ i , the covariance is C i Gaussian distribution of The Gaussian mixture model is used to assign each data point in the load curve to the corresponding Gaussian distribution, and the characteristics of each data point are obtained as the load characteristics of each prosumer.

3. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 2 is characterized in that: Based on the clustering algorithm, the load characteristics of each prosumer are classified to obtain load patterns that respond to different energy demands, including: The objective function of the K-means clustering algorithm is established according to the following formula: in, m is the number of data points, indicating the total number of load characteristic data points involved in clustering. n is the number of clusters, which means that m data points are divided into n clusters. x i is the ith data point, ||x i -c a || 2 is the relationship between each data location and each cluster centroid c a The distance between Based on the K-means clustering algorithm, the categories of the clusters to which the data points belong are adjusted until the objective function reaches the minimum value, and the load patterns that respond to different energy demands are obtained.

4. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 1, characterized in that: Before extracting features from the load curve to obtain the load features of each prosumer, the method further includes: The data in the load curve is preliminarily classified according to the time scenario to obtain the load curve data corresponding to multiple time scenarios.

5. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 1, characterized in that: The comprehensive index value is calculated based on the user's self-sufficiency rate and the total cost of power products according to the following formula: L=w1×SSR+w2×LCOE 2 Among them, L is the comprehensive index value, SSR is the user self-sufficiency rate, LCOE is the total cost of electricity products, w1 and w2 are the weights of user self-sufficiency rate and total cost of electricity products respectively.

6. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 5 is characterized in that: The user self-sufficiency rate is calculated according to the following formula: Among them, SSR is the user self-sufficiency rate, E du E is the energy directly supplied to the load by photovoltaic power generation within the preset time period. dis is the actual discharge amount of the battery within the preset time period, E load is the total annual user-side load of the system, E PV It is the amount of electricity generated by the photovoltaic power generation system in a given period of time, DOD is the depth of discharge of the battery, and C is the rated capacity of the battery.

7. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 5, characterized in that: The total cost of electricity products is calculated according to the following formula: CRF=i(1+i n ) / ((1+i n )-1) in, LCOE is the total cost of electricity products, which is used to express the cost of producing one kilowatt-hour of electricity. A is the cost generated by the scheduling model, and E bygrid ×γ retail Refers to the market value fuel cost, E sellsback ×γ sell To provide value for electricity sales, E load is the total annual user-side load of the system, I PV is the photovoltaic system capacity, I BES is the capacity of the battery energy storage system, OM is the proportion of annual operation and maintenance costs to the cost generated by the dispatch model, CRF is the capital recovery factor, n is the service life of the photovoltaic power generation equipment, i is the interest rate, and E sellsback E is the remaining energy of the photovoltaic system after self-use of energy. PV,DC is the total amount of DC power generated by the photovoltaic system, Eff BES is the charging efficiency of the battery energy storage system, Eff inv is the inverter efficiency, E sum is the total energy supported by the battery energy storage system, E sc,dc is the self-used energy with battery energy storage system, E sc,dc,0 is the self-consumed energy without battery energy storage system, E SC is the energy self-consumption when there is a battery energy storage system, E SC,initial It is the self-consumption of energy without battery energy storage system.

8. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 1, characterized in that: The dispatch model includes a prosumer self-consumption optimization dispatch model, the prosumer self-consumption optimization dispatch model includes a solar rooftop photovoltaic maximum self-consumption algorithm and an optimal battery energy storage system dispatch algorithm, the dispatch setting of the battery energy storage system capacity is the prosumer maximum energy self-consumption, when the power generated by the photovoltaic system is greater than the household demand, the battery will continue to charge, otherwise, the battery will discharge until the minimum allowed level; The maximum battery discharge power P in each time interval in the prosumer self-consumption optimization scheduling model dsc,i It is expressed as follows: In the prosumer self-consumption optimization scheduling model, the battery charging power is maximized P chr,i It is expressed as follows: The battery state of charge SOC in the prosumer self-consumption optimization scheduling model i It is expressed as follows: The battery discharge state SOC in the prosumer self-consumption optimization scheduling model i It is expressed as follows: Among them, SOC i is the battery charge and discharge state at the i-th moment, SOC i-1 is the battery charge and discharge state at the i-1th moment, Eff BES is the charging efficiency of the inverter, P BES is the efficiency of the battery energy storage system, δt is the time interval, P BES,cap represents the power that cannot be utilized due to the capacity limitation of the battery energy storage system, p charge is the charging power of the energy storage device, p dis is the discharge power of the energy storage device, η charge is the charging efficiency of the energy storage device, η dis is the discharge efficiency of the energy storage device, E EES The amount of charge or discharge during a time interval.

9. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 1, characterized in that: The scheduling model includes a peak shaving algorithm scheduling model, wherein the peak shaving algorithm scheduling model is a scheduling control for peak reduction according to the electric energy generated by the photovoltaic system and the energy threshold. The energy threshold in the peak shaving algorithm scheduling model is expressed as follows: AND th =(∑P Th )δ t -AND BES,cap The discharge peak shaving Φ in the peak shaving algorithm model dsc,i Calculated as follows: F dsc,i =[min(+P Th,i ,P max,dsc,i )] The battery charging power Φ in the peak shaving algorithm model chr,i Calculated as follows: F chr,i =[min(-P Th,i ,P max,dsc,i )] Among them, E th is the energy threshold, P Th represents the rate of electrical energy generation, δ t Indicates the time interval, E BES,cap represents the energy that cannot be utilized due to the capacity limitation of the battery energy storage system, P Th,i Indicates the amount of power that needs to be released from the battery in order to shave the peak during the time interval. It refers to the part of the electricity generated by the photovoltaic system that can meet the load demand after being inverted. PV,DC Indicates the DC power actually generated by the photovoltaic system; P max,dsc,i It is the maximum allowable power when the battery is discharged to cut the peak during the time interval.

10. The method for optimizing electricity consumption of prosumers based on load clustering algorithm according to claim 9, characterized in that: The peak shaving algorithm model performs dispatching control of peak shaving according to the electric energy generated by the photovoltaic system and the energy threshold, including: The electrical energy E generated by the photovoltaic system PV (t) is less than the energy threshold E th In the case of PV (t) is used to satisfy the load L(t); if E PV (t) is less than L(t), then E PV (t) After the load is applied, the insufficient part is supplemented by the battery discharge, and the discharge amount is L(t)-E PV (t), otherwise the remaining power E PV (t)-L(t) is sent to the public grid, and the peak shaving amount of battery discharge is the load L(t) and E PV The difference of (t); The electrical energy E generated by the photovoltaic system PV (t) is greater than or equal to the energy threshold E th In the case of E PV (t) is equal to the energy threshold E th The amount of electricity is used to meet the load or to be fed into the public grid; if L(t) is less than E th , then E th -L(t) is fed into the public grid, otherwise it is discharged by the battery L(t)-E th Or purchase electricity L(t)-E th To meet the load; E PV (t) is equal to the energy threshold E th The amount of electricity used to meet the load or sent to the public grid is the energy E PV (t)-E th Stored in the battery, the storage capacity is Among them, C is the rated capacity of the battery.

11. A prosumer electricity optimization system using the prosumer electricity optimization method based on load clustering algorithm according to any one of claims 1 to 10, characterized in that: The system comprises: A collection module is used to collect electricity consumption data, load data and photovoltaic power generation data of prosumers within a preset period of time, and generate load curve data in time series; A feature extraction module, for extracting features from the load curve data to obtain the load features of each prosumer, wherein the load features include at least one of the following: consumption pattern, peak demand time, change in power consumption, load factor, peak load position and negative peak load position and standard deviation; A classification module is used to classify the load characteristics of each prosumer based on a clustering algorithm to obtain load patterns that respond to different energy demands; A calculation module is used to calculate the user self-sufficiency rate and the total cost of power products under various scheduling models of prosumers based on the classified load patterns, wherein the user self-sufficiency rate is used to characterize the proportion of renewable energy in the user load; The dispatch optimization module is used to calculate the comprehensive index value based on the user's self-sufficiency rate and the total cost of power products, and select the dispatch model with the highest comprehensive index value to optimize the electricity consumption of producers and consumers.

12. An electronic device, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the method for optimizing electricity consumption of prosumers based on a load clustering algorithm according to any one of claims 1 to 10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for optimizing electricity consumption of producers and consumers based on a load clustering algorithm as described in any one of claims 1 to 10 are implemented.

Images