Joint generation method of regional multi-user load power consumption scenarios

Abstract

The present invention discloses a method for jointly generating a regional multi-user load power usage scenario, which mainly includes: dividing users in a region into different categories of household types, and establishing a regional electrical appliance information database; selecting household samples for each type of household in the region by simple random sampling, and establishing a regional electrical appliance usage behavior knowledge base for electrical appliances that need to be manually controlled to turn on and off; establishing a power change pattern sample library based on the complete working cycle power data in various working modes of the electrical appliances; determining the number of households of each type in generating a multi-user load power usage scenario power data set; and generating power data of multiple user load power usage scenarios in the data set. The present invention can ensure that the generated user load power usage scenario power data covers a variety of electrical appliances, the number of data samples is sufficient and has tag information, and can reflect the user's real power usage habits and electrical appliance power consumption mode as much as possible, and can support relevant theoretical method research and product testing and evaluation.

Description

Technical Field

[0001] The present invention relates to the field of load power consumption data generation, and in particular to a method for jointly generating regional multi-user load power consumption scenarios. Background Art

[0002] The theoretical method research and product testing and evaluation of non-intrusive load monitoring, load forecasting and electricity consumption big data analysis technology have high requirements for user load power consumption scene power data. In order to ensure the effectiveness of algorithm design and model training, as well as the fairness and credibility of test results, user load power consumption scene power data often needs to meet the following conditions: ① Include multiple user scenarios with different load compositions, covering a variety of electrical appliances; ② Conform to the normal user usage habits, and be able to reflect the typical combination and rules of actual electrical appliances in mixed operation; ③ Ensure that the number of user load power data samples is sufficient and has labeling information; ④ Consider the impact of different regions and seasons on the scene load composition and user usage habits. Obviously, it is difficult to establish a data set that meets the above requirements by simply relying on long-term continuous field collection of a large number of real users. In order to provide a realistic and feasible solution, it is necessary to study the generation method of multi-user load power consumption scene power data.

[0003] The generation of power data for user load power consumption scenarios involves two parts: the simulation of appliance usage behavior and the simulation of power consumption patterns. Among them, the simulation of appliance usage behavior models the three variables of appliance usage frequency, start-up time, and usage duration to characterize the distribution characteristics or laws of whether the appliance is running over time; the simulation of appliance power consumption patterns is used to characterize the power consumption data of appliances under different working modes.

[0004] At present, in the research on power data generation of user load power consumption scenarios at home and abroad, there are different strategies for simulating the use behavior of electrical appliances. One of the strategies is that the use behaviors of different electrical appliances do not affect each other and the use behavior variables of each electrical appliance satisfy independent uniform distribution or normal distribution. However, considering that there are correlations between different electrical appliances in real life (such as washing machines and dryers) and that the behavior variables do not simply satisfy independent uniform distribution and normal distribution, it is necessary to conduct correlation analysis on the use behavior of electrical appliances and obtain the original approximate distribution of the use behavior variables. According to experience, it is also necessary to consider the impact of region, season and user type on the use behavior habits of electrical appliances.

[0005] Regarding the simulation of electrical appliance power consumption, in the current research, one solution is to directly measure the power consumption data samples of the working cycle of the appliance in different working modes to characterize the different power consumption modes of the appliance. In order to meet the requirements for the use time, the power data samples that have been measured are directly resampled to shorten the time or interpolated to increase the time. However, this method will destroy the inherent power consumption mode of some appliances, such as changing the time scale of the power fluctuation of the washing machine. In addition, this solution does not take into account the randomness and variability of power consumption changes. Another solution is to define four basic power consumption parameter models through the analysis and summary of some electrical appliance power consumption data samples, which are used to characterize the power consumption change characteristics of different types of electrical appliances. The specific power consumption model of each appliance can be a composite of the basic model. This method can change the generated power consumption curve by adjusting and controlling the basic model parameters, so that the simulated power shows a certain degree of randomness. However, some electrical appliances have great differences between different brands and different technical schools, and it is not suitable to characterize and represent a type of electrical appliances with one or several mathematical models. Otherwise, when generating power data of user load power consumption scenarios, the real power consumption characteristics of electrical appliances will not be reproduced, such as variable frequency electrical appliances.

[0006] Whether it is the algorithm design or effect testing of non-intrusive load monitoring, load forecasting and electricity consumption big data analysis technology, it will not focus on a single family or specific scenario, but rather on a certain representative user load power consumption scenario power data set, while the above existing research is more focused on generating user load power consumption data for a certain user. When it comes to data sets, in addition to considering the rationality of the generated single user load data, it is also necessary to consider the diversity of user types contained in the set and the differences between their corresponding load compositions, so that a variety of electrical appliances need to be covered.

[0007] Although there have been many studies on non-invasive load monitoring, load forecasting and electricity consumption big data analysis technologies at home and abroad in recent years, few people have developed a method for generating power data of regional multi-user load power consumption scenarios. In order to meet the above needs and combined with existing problems, the present invention designs a method for generating power data of regional multi-user load power consumption scenarios that can comprehensively consider multiple influencing factors such as region and season. Summary of the invention

[0008] Taking into account the shortcomings of the existing technology and in order to further promote the development of non-invasive load monitoring, load forecasting and electricity consumption big data analysis technology, the present invention designs a method for generating regional multi-user load power consumption scenario power data that can comprehensively consider multiple influencing factors such as region and season, aiming to provide a method for generating user load power data sets that can generate user load power consumption scenario power data with complexity, user type diversity and load composition differences that are consistent with actual conditions.

[0009] In order to meet the above needs, the present invention proposes a method for jointly generating regional multi-user load power usage scenarios, which mainly includes: dividing users in the region into different categories of household types, and establishing a regional electrical appliance information database; selecting household samples for each type of household in the region by simple random sampling, and establishing a regional user electrical appliance usage behavior knowledge base for electrical appliances that require manual control of their on and off; establishing an electrical appliance power change pattern sample library based on the complete working cycle power data in various working modes of the electrical appliances; determining the number of users of each type of household in the power data set for generating multi-user load power usage scenarios; and generating power data for multiple user load power usage scenarios in the data set. The specific steps are as follows:

[0010] Step 1: users in a region are divided into different types of families according to the number, age, family income, housing type and housing area of ​​family members, and an electrical appliance information database of the region is established, wherein the electrical appliance information database includes information such as the proportion of the number of users in different types of families, the types of electrical appliances in different types of families, the penetration rate of various electrical appliances in different types of families, and the market share of corresponding brands and models;

[0011] Step 2: Select household samples by simple random sampling from various types of households in the region. According to whether the appliances need to be turned on or off manually, the appliances are divided into normally-on appliances and manually-controlled appliances. For manually-controlled appliances, a knowledge base of regional users' appliance usage behavior is established. The steps are as follows:

[0012] Step 2-1) The usage frequency F of each manually controlled electrical appliance in the household sample req , start time T s , Usage time T on To conduct statistics;

[0013] Step 2-2) Fit the distribution functions of the above three statistics based on the normal function. For single-peak distribution, use the normal function for fitting; for multi-peak distribution, use formula (1) for fitting:

[0014]

[0015] In formula (1), a1, a2, ..., a n is the weight, and a1+a2+…+a n =1, x is the frequency of use F req , start time T s , Usage time T on Statistical variables, μ1, μ2, ... μ n ,σ1,σ2,……σ n is the corresponding distribution parameter;

[0016] Step 2-3) Use the frequent pattern mining method shown in formula (2) to determine whether there is a correlation between the usage behaviors of different electrical appliances in the same household.

[0017]

[0018] In formula (2), couut(T si )、count(T sj ) represent the number of times appliances i and j are turned on, ΔT s(i，j) is the start-up time difference of appliances i and j, ΔT th To determine whether the two are adjacent to each other, count(|ΔT s(i，j) |<ΔT th ) is the number of times the time difference between the start-up time of appliances i and j is less than the time threshold, support th is the support, D is the number of statistical days, and k is the corresponding coefficient threshold; if there is a usage behavior association, the opening time difference ΔT between appliances i and j is used s(i，j) and the operating time T of appliance i oni and the running time T of appliance j onj The ratio m represents the operating relationship between the two, m = T oni / T onj , and according to formula (1) the variable ΔT s(i，j) Fit with m;

[0019] Step 3: Establish a sample library of electrical power change patterns. The steps are as follows:

[0020] 3-1) According to all the working modes of a certain electrical appliance, the power data of complete working cycles of the electrical appliance of different durations under different working modes are measured to obtain the combination of all power change modes of the electrical appliance;

[0021] 3-2) If the ratio of the operating time between different power change modes in the combination mode of the power change mode of the electrical appliance is adjustable, the power data of the complete working cycle of the electrical appliance is segmented to obtain the power change mode sample of the electrical appliance; if the ratio of the operating time between different power change modes in the combination mode of the power change mode of the electrical appliance is fixed, the power data of the complete working cycle of the electrical appliance is used as its power change mode sample;

[0022] 3-3) The power change pattern samples of all electrical appliances form an electrical appliance power change pattern sample library;

[0023] Step 4: Determine the number of users of each type of household in the power data set for generating a multi-user load power consumption scenario according to the demand and the proportion of the number of users of each type of household in the electrical appliance information database of the region;

[0024] Step 5: Generate power data of each user load power consumption scenario in the data set. The process is as follows:

[0025] 5-1) Generate a load power consumption scenario for each user under each household type, and determine the user's electrical appliance composition according to the electrical appliance penetration rate of users of this household type in the electrical appliance information database;

[0026] 5-2) Determine the usage behavior of each appliance: For normally-on appliances, their usage behavior runs all day long; for manually controlled appliances, the steps to determine their usage behavior are as follows:

[0027] First, a household sample containing the appliance is randomly selected from the household type of users in the regional user usage behavior knowledge base. Then, a derivative transformation is performed on the statistical distribution of appliance usage behavior in the household sample, and the parameters of the derivative distribution are determined according to formula (3):

[0028]

[0029] In formula (3), x is the relative parameter of the original distribution, x′ is the parameter derived from parameter x, and ε is the predefined proportional coefficient; the statistical distribution of the corresponding appliance usage behavior is constructed according to the parameters of the derived distribution, and Monte Carlo random sampling is performed based on it to determine the usage behavior of the appliance in the scenario;

[0030] 5-3) Select the corresponding appliance brands and models based on the appliance models and brand market shares of the household type of the user in the appliance information database; for normally-on appliances, select power samples from the appliance power change pattern sample library, and make them run all day as the working cycle of the appliance; for manually controlled appliances, randomly determine the working mode used each time, generate a series of working cycles under the working mode based on the appliance power change pattern sample library, and select the working cycle that meets the appliance usage behavior determined in step 5-2) from the series of working cycles;

[0031] 5-4) Superimposing the working cycle power of all electrical appliances of the user to obtain the total load power data of the user;

[0032] Step 6: Repeat step 5 to finally generate power data for the multi-user load power consumption scenario in the area.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] (1) Based on the number, age, and income of family members, as well as the type and area of ​​residence, users with different electricity usage habits in the region are classified. Based on the regional electrical appliance information database, user load power usage scenarios are generated by type, so that the constructed user load power usage scenario power data set can reasonably reflect the diverse user types and corresponding power usage behaviors in the region.

[0035] (2) Based on a limited number of household samples, the distribution parameters of variables related to their appliance usage behavior are reasonably derived to form a personalized distribution of user appliance usage behavior, thereby ensuring that users of the same type generated by this sampling present differentiated appliance usage behaviors that are consistent with reality.

[0036] (3) Considering that the basic power consumption units contained in electrical appliances will exhibit different power consumption modes when the appliances are used due to their different working principles or control methods, using the typical power change pattern samples of the basic power consumption units of electrical appliances as the power change pattern samples of electrical appliances can more flexibly generate power samples of the working cycles of electrical appliances whose power consumption complexity is consistent with the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is the framework of the method for jointly generating regional multi-user load power consumption scenarios of the present invention;

[0038] Figure 2 It is a flow chart of the method for jointly generating regional multi-user load power consumption scenarios of the present invention;

[0039] Figure 3-1 It is the power consumption data of a refrigerator that is always on for one day generated in the example;

[0040] Figure 3-2 It is the power consumption data of the electric kettle, a manually controlled electrical appliance, generated during the implementation for one day;

[0041] Figure 4-1 , Figure 4-2 , Figure 4-3 , Figure 4-4 and Figure 4-5 The following are schematic diagrams of electricity consumption of 5 single-working households and 5 elderly households on a working day; (a) shows the electricity consumption of single-working households, and (b) shows the electricity consumption of elderly households. DETAILED DESCRIPTION

[0042] The present invention is further described below in conjunction with the accompanying drawings and specific research materials, but the following research materials are by no means limiting the present invention in any way.

[0043] like Figure 1As shown, the framework of the method for jointly generating regional multi-user load power consumption scenarios of the present invention is generally divided into two stages:

[0044] First, prior information collection and sample measurement. First, establish a regional electrical appliance information database in a given season, including the proportion of the number of users in different types of households, the types of electrical appliances in different types of households, the penetration rate of various electrical appliances in different types of households, and the market share of corresponding brands and models. Secondly, build a knowledge base of electrical appliance usage behavior in the region, select a certain number of household samples from each type of user, obtain their time variables related to electrical appliance usage behavior for multiple days through measurement (or survey), and form corresponding statistical distribution laws. Finally, build a sample library of electrical appliance power change patterns, and measure the power consumption data of complete working cycles of different brands and models of electrical appliances with different durations. Affected by the working mode of the electrical appliance and the user's usage habits, the types and combinations of power change patterns in the working cycle of the electrical appliance are also different. Therefore, the power change pattern samples of the electrical appliance are obtained through pattern segmentation as its basic power consumption model, so as to reproduce the complex characteristics of electrical appliance power consumption in real power consumption scenarios. It is worth noting that unlike the use of electrical appliances which is affected by the user's region, the power change pattern is an inherent electrical characteristic of the appliance. For appliances of the same brand and model, the power change pattern samples measured in a certain area can be used globally across regions, that is, used to generate power data sets for household load power usage scenarios in other different regions.

[0045] Second, the user load power consumption data set is generated. First, the number of users of different types of households to be generated is determined according to the demand and the proportion of users of each type of households in the electrical appliance information database of the area. Then, for a certain user to be generated under a given type of household, according to the load composition information of this type of user in the electrical appliance information database (or obtained by survey), on the one hand, the type of electrical appliances contained in the user is determined, and the statistical distribution and its derivative laws in the electrical appliance usage behavior knowledge base are used to simulate the control mode of each electrical appliance and the user's usage habits; on the other hand, the specific brand and model of the electrical appliances contained in the user to be generated are determined, and the power change pattern samples of the corresponding electrical appliances are randomly selected from the power change pattern sample library. The electrical appliance working cycle power data samples that meet the requirements are generated by combination, and the working cycle power data samples of each electrical appliance are arranged and superimposed according to the simulated electrical appliance usage behavior to generate the total load power consumption data of the user.

[0046] The above framework is used to implement a method for jointly generating regional multi-user load power usage scenarios, which mainly includes: dividing users in the region into different types of households and establishing a regional electrical appliance information database; using simple random sampling to select household samples for each type of household in the region, and establishing a regional user electrical appliance usage behavior knowledge base for electrical appliances that require manual control of the on and off of the appliances; establishing an electrical appliance power change pattern sample library based on the complete working cycle power data in various working modes of the appliances; determining the number of users of each type of household in the power data set for generating multi-user load power usage scenarios; and generating power data for multiple user load power usage scenarios in the data set. Figure 2 As shown, the specific steps are as follows:

[0047] Step 1: users in a region are divided into different types of families according to the number, age, family income, housing type and housing area of ​​family members, and an electrical appliance information database of the region is established, wherein the electrical appliance information database includes information such as the proportion of the number of users in different types of families, the types of electrical appliances in different types of families, the penetration rate of various electrical appliances in different types of families, and the market share of corresponding brands and models;

[0048] Step 2: Select household samples by simple random sampling from various types of households in the area, and divide the electrical appliances into normally-on electrical appliances and manually-controlled electrical appliances according to whether they need to be manually controlled on or off;

[0049] To establish a knowledge base of regional users' electrical appliance usage behavior for manually controlled electrical appliances, the steps are as follows:

[0050] Step 2-1) The usage frequency F of each manually controlled electrical appliance in the household sample req , start time T s , Usage time T on To conduct statistics;

[0051] Step 2-2) Fit the distribution functions of the above three statistics based on the normal function. For single-peak distribution, use the normal function for fitting; for multi-peak distribution, use formula (1) for fitting:

[0052]

[0053] In formula (1), a1, a2, ..., a n is the weight, and a1+a2+…+a n =1, x is the frequency of use F req , start time T s , Usage time T on Statistical variables, μ1, μ2, ... μ n ,σ1,σ2,……σ nis the corresponding distribution parameter;

[0054] Step 2-3) Use the frequent pattern mining method shown in formula (2) to determine whether there is a correlation between the usage behaviors of different electrical appliances in the same household.

[0055]

[0056] In formula (2), couut(T si )、count(T sj ) represent the number of times appliances i and j are turned on, ΔT s(i，j) is the start-up time difference of appliances i and j, ΔT th To determine whether the two are adjacent to each other, count(|ΔT s(i，j) |<ΔT th ) is the number of times the time difference between the start-up time of appliances i and j is less than the time threshold, support th is the support, D is the number of statistical days, and k is the corresponding coefficient threshold;

[0057] If there is a usage behavior association, the start-up time difference ΔT of appliances i and j is used s(i，j) and the operating time T of appliance i oni and the running time T of appliance j onj The ratio m(m=T oni / T onj ) represents the operating relationship between the two, and according to formula (1) the variable ΔT s(i，j) Fit with m;

[0058] Step 3: Establish a sample library of electrical power change patterns. The steps are as follows:

[0059] 3-1) According to all the working modes of a certain electrical appliance, the power data of complete working cycles of the electrical appliance of different durations under different working modes are measured to obtain the combination of all power change modes of the electrical appliance;

[0060] 3-2) If the ratio of the operating time between different power change modes in the combination mode of the power change mode of the electrical appliance is adjustable, the power data of the complete working cycle of the electrical appliance is segmented to obtain the power change mode sample of the electrical appliance; if the ratio of the operating time between different power change modes in the combination mode of the power change mode of the electrical appliance is fixed, the power data of the complete working cycle of the electrical appliance is used as its power change mode sample;

[0061] 3-3) The power change pattern samples of all electrical appliances form an electrical appliance power change pattern sample library;

[0062] Step 4: Determine the number of users of each type of household in the power data set for generating a multi-user load power consumption scenario according to the demand and the proportion of the number of users of each type of household in the electrical appliance information database of the region;

[0063] Step 5: Generate power data of each user load power consumption scenario in the data set. The process is as follows:

[0064] 5-1) Generate a load power consumption scenario for each user under each household type, and determine the user's electrical appliance composition according to the electrical appliance penetration rate of users of this household type in the electrical appliance information database;

[0065] 5-2) Determine the usage behavior of each appliance:

[0066] For normally-on electrical appliances, their use behavior runs all day long;

[0067] For manually controlled appliances, the steps to determine their usage behavior are as follows:

[0068] First, a household sample containing the appliance is randomly selected from the household type of users in the regional user usage behavior knowledge base. Then, a derivative transformation is performed on the statistical distribution of appliance usage behavior in the household sample, and the parameters of the derivative distribution are determined according to formula (3):

[0069]

[0070] In formula (3), x is the parameter corresponding to the original distribution, x′ is the parameter derived from parameter x, and ε is the predefined proportional coefficient;

[0071] Construct the statistical distribution of the corresponding appliance usage behavior based on the parameters of the derived distribution, and perform Monte Carlo random sampling based on it to determine the usage behavior of the appliance in the scenario;

[0072] 5-3) Select the corresponding appliance brands and models based on the appliance models and brand market shares of the household type of the user in the appliance information database; for normally-on appliances, select power samples from the appliance power change pattern sample library, and make them run all day as the working cycle of the appliance; for manually controlled appliances, randomly determine the working mode used each time, generate a series of working cycles under the working mode based on the appliance power change pattern sample library, and select the working cycle that meets the appliance usage behavior determined in step 5-2) from the series of working cycles;

[0073] 5-4) Superimposing the working cycle power of all electrical appliances of the user to obtain the total load power data of the user;

[0074] Step 6: Repeat step 5 to finally generate power data for the multi-user load power consumption scenario in the area.

[0075] Examples of research materials:

[0076] By using the above method flow, a power dataset of user load power consumption scenarios in a certain area in summer is generated.

[0077] The power consumption data curves of the normally open refrigerator and the manually controlled electric kettle generated by the example are as follows: Figure 3-1 and Figure 3-2 As shown. For refrigerators, a complete working cycle of a refrigerator of the corresponding brand and model is selected from the sample library of power change patterns of electrical appliances, and its complete working cycle is arranged throughout the day to generate its power consumption data for one day. For electric kettles, based on the statistical distribution of the user's use behavior of the electric kettle, its usage frequency, start-up time and usage duration are randomly determined for one day, and a series of working cycles of electric kettles of the corresponding brand and model are generated based on the sample library of power change patterns of electrical appliances, and the working cycle that meets the usage duration is selected, and it is operated at the corresponding start-up time to generate its power consumption data for one day.

[0078] The power consumption data curve of some users in one day generated by the example is as follows: Figure 4-1 to Figure 4-5 As shown in the figure, the power consumption data curves of 5 single-working households on a working day are as follows: Figure 4-1 to Figure 4-5 As shown in (a), the power consumption data curves of five elderly households on a weekday are as follows: Figure 4-1 to Figure 4-5 As shown in (b), the daily load curve shows that there are obvious differences in the electricity consumption behaviors of the two types of users. Single office workers basically work outside during the day on weekdays, so there is less use of electrical appliances at home during the day, while the opposite is true for elderly families.

[0079] For regional user load power data used for non-invasive load monitoring, load forecasting and electricity consumption big data analysis technology, although it is possible to obtain a labeled data set of real users over a period of time through on-site invasive measurements, it is very costly to rely on this method to obtain a data set with the required number and diversity of users, and the existing user load power consumption scenario power data generation method cannot simulate the user's electrical appliance usage behavior and electrical appliance power consumption mode well, so that the generated user load power data set is not completely consistent with the real scene. The present invention can simultaneously consider multiple influencing factors that affect the authenticity of the simulation data set, and by establishing an electrical appliance information database, a regional user electrical appliance usage behavior knowledge base, and an electrical appliance power change pattern sample library, the load power data complexity, user type diversity, and load composition differences of each simulation data set are consistent with the actual situation.

[0080] According to the above example, the present invention establishes an electrical appliance information database, a regional user electrical appliance usage behavior knowledge base and an electrical appliance power conversion mode sample library. Since the electrical appliance usage behavior knowledge base and the electrical appliance power sample library retain the basic characteristics of user usage behavior and electrical appliance power consumption mode, the user's electrical appliance usage behavior for multiple days and the electrical appliance working cycle power are simulated based on the electrical appliance usage behavior knowledge base and the electrical appliance power conversion mode sample library, thereby ensuring that the load data of the generated data set is consistent with the actual situation, providing support for theoretical method research and product testing and evaluation of non-invasive load monitoring, load forecasting and electricity consumption big data analysis technology.

[0081] Although the present invention has been described above in conjunction with the accompanying drawings, the present invention is not limited to the above-mentioned specific embodiments, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can make many modifications without departing from the purpose of the present invention, all of which are within the protection of the present invention.

Claims

1. A method for jointly generating regional multi-user load power consumption scenarios, characterized in that: Mainly include: Divide users in the region into different household types and establish a regional electrical appliance information database; Simple random sampling was used to select household samples of all types of households in the region, and a regional user electrical appliance usage behavior knowledge base was established for electrical appliances that require manual control of the on and off of the appliances; To establish a knowledge base of regional users' electrical appliance usage behavior for manually controlled electrical appliances, the steps are as follows: Step 2-1) The usage frequency F of each manually controlled electrical appliance in the household sample req , start time T s , Usage time T on To conduct statistics; Step 2-2) Fit the distribution functions of the above three statistics based on the normal function. For single-peak distribution, use the normal function for fitting; for multi-peak distribution, use formula (1) for fitting: In formula (1), a1, a2, ..., a n is the weight, and a1+a2+…+a n =1, x is the frequency of use F req , start time T s , Usage time T on Statistical variables, μ1, μ2, ... μ n ,σ1,σ2,……σ n is the corresponding distribution parameter; Step 2-3) Use the frequent pattern mining method shown in formula (2) to determine whether there is a correlation between the usage behaviors of different electrical appliances in the same household. In formula (2), count(T si )、count(T sj ) represent the number of times appliances i and j are turned on, ΔT s(i,j) is the start-up time difference of appliances i and j, ΔT th To determine whether the two are adjacent to each other, count(|ΔT s(i,j) |<ΔT th ) is the number of times the time difference between the start-up time of appliances i and j is less than the time threshold, support th is the support, D is the number of statistical days, and k is the corresponding coefficient threshold; If there is a usage behavior association, the start-up time difference ΔT of appliances i and j is used s(i,j) and the operating time T of appliance i oni and the running time T of appliance j onj The ratio m represents the operating relationship between the two, m = T oni / T onj , and according to formula (1) the variable ΔT s(i,j) Fit with m; Establish a sample library of electrical appliance power change patterns based on complete working cycle power data in various working modes of electrical appliances; Determine the number of users of each type of households in generating a multi-user load power consumption scenario power data set; Generate power data of multiple user load power usage scenarios in the data set.

2. The method for jointly generating regional multi-user load power consumption scenarios according to claim 1, characterized in that: The specific steps are as follows: Step 1: users in a region are divided into different types of families according to the number, age, family income, housing type and housing area of ​​family members, and an electrical appliance information database of the region is established, wherein the electrical appliance information database includes information such as the proportion of the number of users in different types of families, the types of electrical appliances in different types of families, the penetration rate of various electrical appliances in different types of families, and the market share of corresponding brands and models; Step 2: Select household samples by simple random sampling from various types of households in the area, and divide the electrical appliances into normally-on electrical appliances and manually-controlled electrical appliances according to whether they need to be manually controlled on or off; Step 3: Establish a sample library of electrical power change patterns. The steps are as follows: 3-1) According to all the working modes of a certain electrical appliance, the power data of complete working cycles of the electrical appliance of different durations under different working modes are measured to obtain the combination of all power change modes of the electrical appliance; 3-2) If the ratio of the operating time between different power change modes in the combination mode of the power change mode of the electrical appliance is adjustable, the power data of the complete working cycle of the electrical appliance is segmented to obtain the power change mode sample of the electrical appliance; if the ratio of the operating time between different power change modes in the combination mode of the power change mode of the electrical appliance is fixed, the power data of the complete working cycle of the electrical appliance is used as its power change mode sample; 3-3) The power change pattern samples of all electrical appliances form an electrical appliance power change pattern sample library; Step 4: Determine the number of users of each type of household in the power data set for generating a multi-user load power consumption scenario according to the demand and the proportion of the number of users of each type of household in the electrical appliance information database of the region; Step 5: Generate power data of each user load power consumption scenario in the data set. The process is as follows: 5-1) Generate a load power consumption scenario for each user under each household type, and determine the user's electrical appliance composition according to the electrical appliance penetration rate of users of this household type in the electrical appliance information database; 5-2) Determine the usage behavior of each appliance: For normally-on electrical appliances, their use behavior runs all day long; For manually controlled appliances, the steps to determine their usage behavior are as follows: First, a household sample containing the appliance is randomly selected from the household type of users in the regional user usage behavior knowledge base. Then, a derivative transformation is performed on the statistical distribution of appliance usage behavior in the household sample, and the parameters of the derivative distribution are determined according to formula (3): In formula (3), x is the parameter corresponding to the original distribution, x ′ is the parameter derived from parameter x, and ε is the predefined proportional coefficient; Construct the statistical distribution of the corresponding appliance usage behavior based on the parameters of the derived distribution, and perform Monte Carlo random sampling based on it to determine the usage behavior of the appliance in the scenario; 5-3) Select the corresponding appliance brands and models based on the appliance models and brand market shares of the household type of the user in the appliance information database; for normally-on appliances, select power samples from the appliance power change pattern sample library, and make them run all day as the working cycle of the appliance; for manually controlled appliances, randomly determine the working mode used each time, generate a series of working cycles under the working mode based on the appliance power change pattern sample library, and select the working cycle that meets the appliance usage behavior determined in step 5-2) from the series of working cycles; 5-4) Superimposing the working cycle power of all electrical appliances of the user to obtain the total load power data of the user; Step 6: Repeat step 5 to finally generate power data for the multi-user load power consumption scenario in the area.

Images