Building load prediction method, device, equipment and medium
By determining scene information in building load prediction, performing Gaussian curve fitting and Monte Carlo simulation, combined with scene feature parameters, the load prediction problem of incomplete factors in the existing technology is solved, and faster and more accurate load prediction is achieved .
Patent Information
- Application Number
- CN202111262932.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-28
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2041-10-28
AI Technical Summary
The prior art has the problem of inaccurate prediction of building load due to incomplete considerations.
By determining the scene information of the current building, obtaining user electricity consumption data, performing Gaussian curve fitting, determining the Monte Carlo simulation algorithm parameters, combining scene feature parameters for simulation, and optimizing load prediction.
It realizes faster and more accurate prediction of building load, takes into account the characteristics of user electricity consumption behavior, and improves the accuracy of prediction results.
Smart Images

Figure CN113887832B_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present invention relate to the field of electric power engineering technology, and in particular to a building load prediction method, device, equipment and medium. Background Art
[0002] With the proposal of my country's "carbon peak" and "carbon neutrality" goals, technologies such as smart buildings and low-carbon buildings are increasingly attracting widespread attention from building investors and integrated energy service providers. In order to effectively manage building energy and improve the operating efficiency of building microgrid systems, it is necessary to predict the load of buildings and provide basic data reference for efficient building energy management and low-carbon operation.
[0003] Generally, there are many factors that affect building load, such as time period, outdoor temperature, power consumption, and pedestrian flow. In the prior art, time series analysis, multiple regression analysis, or artificial neural network are used to predict building load based on the above factors.
[0004] When using existing solutions for forecasting, factors are not considered comprehensively, resulting in inaccurate load forecasting. Summary of the invention
[0005] The embodiments of the present invention provide a building load prediction method, device, equipment and medium, which can optimize the existing implementation scheme for predicting building load.
[0006] In a first aspect, an embodiment of the present invention provides a building load prediction method, comprising:
[0007] Determine the scene information of the current building, and obtain the user power consumption data of the current building according to the scene information;
[0008] Performing Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user's electricity consumption data to obtain a target Gaussian distribution probability density function corresponding to the scene information;
[0009] Determining algorithm parameters of a Monte Carlo simulation algorithm, and determining scene characteristic parameters of the current building according to the user power consumption data;
[0010] The target Gaussian distribution probability density function, the algorithm parameters and the scene characteristic parameters are input into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building.
[0011] In a second aspect, an embodiment of the present invention provides a building load prediction device, comprising:
[0012] A scene determination module, used to determine scene information of a current building, and obtain user power consumption data of the current building according to the scene information;
[0013] A curve fitting module, used to perform Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user's power consumption data, to obtain a target Gaussian distribution probability density function corresponding to the scene information;
[0014] A parameter determination module, used to determine the algorithm parameters of the Monte Carlo simulation algorithm, and determine the scene characteristic parameters of the current building according to the user's electricity consumption data;
[0015] The load prediction module is used to input the target Gaussian distribution probability density function, the algorithm parameters and the scene feature parameters into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building.
[0016] In a third aspect, an embodiment of the present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the building load forecasting method provided in the embodiment of the present invention is implemented.
[0017] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a building load prediction method as provided in an embodiment of the present invention.
[0018] The building load prediction scheme provided in the embodiment of the present invention first determines the scene information of the current building, and obtains the user electricity consumption data of the current building according to the scene information; and performs Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user electricity consumption data to obtain the target Gaussian distribution probability density function corresponding to the scene information; then determines the algorithm parameters of the Monte Carlo simulation algorithm, and determines the scene feature parameters of the current building according to the user electricity consumption data; finally, the target Gaussian distribution probability density function, algorithm parameters and scene feature parameters are input into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building. By adopting the above technical scheme, different user electricity consumption data are obtained according to different scene information of the building, and the user electricity consumption behavior characteristics are taken into account. Further Gaussian fitting is performed on different user electricity consumption data to show the user electricity consumption behavior characteristics, and then the Monte Carlo algorithm is used in combination with different scene feature parameters for simulation, so as to achieve a technical effect of faster load prediction of the current building and more accurate prediction results. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 A schematic diagram of a flow chart of a building load prediction method provided in Embodiment 1 of the present invention;
[0020] Figure 2 A schematic diagram of a flow chart of a building load prediction method provided in Embodiment 2 of the present invention;
[0021] Figure 3 A structural block diagram of a building load prediction device provided in Embodiment 3 of the present invention;
[0022] Figure 4 This is a structural block diagram of a computer device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION
[0023] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and through specific implementation methods. It is understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It is also necessary to explain that, for ease of description, only parts related to the present invention are shown in the accompanying drawings, rather than all structures.
[0024] It should be mentioned before discussing the exemplary embodiments in more detail that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the steps as sequential processes, many of the steps therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the steps can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.
[0025] Embodiment 1
[0026] Figure 1 The present invention provides a flow chart of a method for implementing building load prediction, which can be executed by a building load prediction device to predict building loads. The device can be implemented by software and / or hardware and can generally be integrated into a computer device such as a server. Figure 1 As shown, the method includes:
[0027] S110. Determine the scene information of the current building, and obtain the user power consumption data of the current building according to the scene information.
[0028] The scene information of the above-mentioned building can be divided into: office scene and / or residential scene according to the current usage of the building. It is easy to understand that when the usage scene of the building is different, the corresponding user electricity consumption data is also different.
[0029] For example, in an office building, the user's electricity usage data may be air conditioning electricity usage, lighting electricity usage, and computer electricity usage, etc.; in a residential building, the user's electricity usage data may be household appliances electricity usage, etc.; in a building with both office and residential scenes, the user's electricity usage data may be air conditioning electricity usage, lighting electricity usage, computer electricity usage, and household appliances electricity usage, etc.
[0030] Therefore, when obtaining the user power consumption data of the current building according to the scene information, the power consumption data of the user sockets under different scene information can be intelligently collected through the intelligent collection terminal or intelligent sensor. For example, the intelligent collection terminal or intelligent sensor can collect the power consumption data of each room in the current building, such as the time when the daily electrical appliance is connected to the socket to start using power, the time when the power is stopped, the power consumption during the power consumption period, and the average power during the power consumption period.
[0031] Optionally, the above-mentioned intelligent data collection terminal or intelligent sensor may have a functional module for uploading data information, which can upload the acquired user power consumption data to the server and store it in the database corresponding to the scene information of the current building. The data in the database can indicate the user power consumption behavior data in the current building scene. The advantage of this is that by counting the actual power consumption data of users in the current building scene information, the prediction accuracy of the building load can be improved in the subsequent analysis based on the actual power consumption data of users.
[0032] Furthermore, the data in the database can be preprocessed, that is, invalid and obviously erroneous user electricity consumption data can be eliminated, for example, the collected data is empty, or the data obviously exceeds the normal range. The preprocessed user electricity consumption data can be further statistically analyzed, and the user electricity consumption data can be summarized and counted according to different actual needs. For example, the number of electrical appliances connected to the socket for power consumption at different time segments in a day, the proportion of electrical appliances connected in each time segment, and the number of times a single socket is connected to an electrical appliance in a day can be summarized and obtained.
[0033] S120 . Perform Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user's electricity consumption data to obtain a target Gaussian distribution probability density function corresponding to the scene information.
[0034] The embodiments of the present invention can abstract the problem of predicting the load power of sockets connected at different time segments in a day into a probability problem. The current building load can be predicted by the power of sockets connected at different time periods in the corresponding scene information of the current building. Therefore, the Gaussian distribution probability density function can be used to fit the relevant parameters.
[0035] Among them, the initial Gaussian distribution probability density function represents a Gaussian distribution probability density function containing unknown parameters. By using a preset method to calculate the user's electricity consumption data, the values of the unknown parameters related to the initial Gaussian distribution probability density function can be obtained. Exemplarily, the preset method can be a nonlinear least squares method or a least squares method. The specific preset method is not limited here, as long as the unknown parameters of the Gaussian distribution probability density function can be calculated.
[0036] Substituting the calculated unknown parameters into the Gaussian distribution probability density function, the initial Gaussian distribution probability density function can be obtained. The initial Gaussian distribution probability density function is further fitted with a Gaussian curve through the user's electricity consumption data. The fitting method can be to determine the relevant value of the initial Gaussian distribution probability density function, obtain the target Gaussian distribution probability density function corresponding to the scene information, and determine whether the value corresponding to the current target Gaussian distribution probability density function is close to the actual value. If not, continue fitting until the value of the obtained target Gaussian distribution probability density function meets the preset accuracy requirement, indicating that the fitting of the target Gaussian distribution probability density function corresponding to the current scene information is completed.
[0037] The relevant values of the initial Gaussian distribution probability density function may include the number of fitting segments and the fitting accuracy threshold.
[0038] S130, determining algorithm parameters of the Monte Carlo simulation algorithm, and determining scene characteristic parameters of the current building according to user power consumption data.
[0039] The Monte Carlo simulation algorithm is an algorithm that repeatedly generates time series by setting random processes, calculating the estimates and statistics of related parameters, and then studying the distribution characteristics of the object. Its principle can be understood as follows: when the research object itself has probabilistic characteristics, the sampling results can be generated by computer simulation methods, and the values of statistics or parameters can be calculated based on the sampling. As the number of simulations increases, a stable conclusion can be obtained by averaging the estimated values of each statistic or parameter.
[0040] Among them, the algorithm parameters of the Monte Carlo simulation algorithm may include: the number of Monte Carlo simulation simulations, the number of simulated intraday periods, the simulation interval and the number of simulation periods.
[0041] Preferably, the number of Monte Carlo simulations can be greater than 100 and less than 500. The number of simulated intraday time periods is 1440 points by default, that is, the time interval is 1 minute (min). The relationship between the number of simulation time periods and the time interval is as follows:
[0042]
[0043] In the above formula, T represents the number of simulation periods, T step Indicates the simulation time interval in minutes.
[0044] It should be noted that the algorithm parameters of the above Monte Carlo simulation algorithm need to be set according to the specific requirements of the scene information of the current building, and are not limited to the above examples.
[0045] Furthermore, the scene characteristic parameters of a building may vary according to the scene information of the current building. For example, if the scene information of the current building is an office scene and / or a residential scene, the corresponding characteristic parameters may be: the number of sockets in the office area and / or the residential area, the mean and variance of the power load of the office area and / or the residential area, the temperature range outside the building, and the temperature range of the comfort setting inside the building, etc. The advantage of setting the scene characteristic parameters of a building is that the power consumption behavior of various electrical appliances in the scene information of the current building can be effectively reflected through the characteristic parameters of different scenes.
[0046] It should be noted that, in the building load prediction method provided by the embodiment of the present invention, the specific content included in the scene characteristic parameters of the current building is not limited to the above specific content, but is subject to the actual needs of researchers.
[0047] S140, inputting the target Gaussian distribution probability density function, algorithm parameters and scene feature parameters into a Monte Carlo simulation algorithm for simulation to obtain a load prediction result for the current building.
[0048] According to step S120, the target Gaussian distribution probability density function corresponding to the scene information is obtained, and according to step S130, the algorithm parameters of the Monte Carlo simulation algorithm are determined, and the scene characteristic parameters of the current building are obtained, so that the target Gaussian distribution probability density function, algorithm parameters and scene characteristic parameters can be input into the Monte Carlo simulation algorithm, and the Monte Carlo simulation algorithm is executed for simulation to obtain the load prediction result of the current building.
[0049] Due to the load prediction method for a building provided by the embodiment of the present invention, based on the input scene characteristic parameters, the situation of the electrical appliances in the current building being connected to the sockets is predicted, and the prediction results may include information such as when the electrical equipment is connected and the power of the electrical equipment connected in the current period, the load prediction results of the current building that take into account the user's electricity consumption behavior can be obtained efficiently and quickly.
[0050] The building load prediction method provided by the embodiment of the present invention first determines the scene information of the current building, and obtains the user electricity consumption data of the current building according to the scene information; and performs Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user electricity consumption data to obtain the target Gaussian distribution probability density function corresponding to the scene information; then determines the algorithm parameters of the Monte Carlo simulation algorithm, and determines the scene feature parameters of the current building according to the user electricity consumption data; finally, the target Gaussian distribution probability density function, the algorithm parameters and the scene feature parameters are input into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building. By adopting the above technical scheme, different user electricity consumption data are obtained according to different scene information of the building, and the user electricity consumption behavior characteristics are taken into account. Further Gaussian fitting is performed on different user electricity consumption data to show the user electricity consumption behavior characteristics, and then the Monte Carlo algorithm is used to simulate in combination with different scene feature parameters, so as to achieve a technical effect of faster load prediction of the current building and more accurate prediction results.
[0051] Embodiment 2
[0052] The embodiment of the present invention is further optimized on the basis of the above embodiment, and optimizes that before the Gaussian curve fitting is performed on the initial Gaussian distribution probability density function according to the user's electricity consumption data, it also includes: using the nonlinear least squares method to calculate the user's electricity consumption data to obtain the unknown parameters; substituting the unknown parameters into the preset probability density function to obtain the initial Gaussian distribution probability density function.
[0053] The step of obtaining the load forecast result of the current building is also optimized, including: dividing the time of user electricity consumption according to the user electricity consumption data to obtain multiple time segments; based on the multiple time segments, inputting the target Gaussian distribution probability density function, the algorithm parameters and the scene characteristic parameters into the Monte Carlo simulation algorithm for segmented simulation to obtain the daily load forecast curve corresponding to the current building; when the number of simulations of each of the time segments reaches a preset threshold, summarizing the daily load forecast curves corresponding to the current building to obtain the daily load forecast result of the current building.
[0054] like Figure 2 As shown, Figure 2 A flow chart of a building load prediction method provided in Embodiment 2 of the present invention is provided. Specifically, the method comprises the following steps:
[0055] S210. Determine the scene information of the current building, and obtain the user power consumption data of the current building according to the scene information.
[0056] S220: Calculate the user's electricity consumption data using a nonlinear least squares method to obtain parameters to be determined.
[0057] The nonlinear least squares method is a parameter estimation method that estimates the parameters of nonlinear static models based on the criterion of minimizing the sum of squares of errors.
[0058] Before obtaining the pending parameters, it is necessary to analyze the actual user power consumption data collected in the corresponding time period, as shown in Table 1 and Table 2 below:
[0059] Table 1: Number of office appliances connected to sockets during the corresponding time period and the proportion of each time period
[0060]
[0061]
[0062] Table 2: Number of times household appliances were connected to sockets during the corresponding time periods and the proportion of each time period
[0063] Time period Number of accesses The proportion Time period Number of accesses The proportion 0:00-1:00 66 0.033 12:00-13:00 76 0.038 1:00-2:00 61 0.031 13:00-14:00 72 0.036 2:00-3:00 53 0.026 14:00-15:00 108 0.054 3:00-4:00 55 0.028 15:00-16:00 141 0.071 4:00-5:00 54 0.027 16:00-17:00 118 0.059 5:00-6:00 53 0.026 17:00-18:00 118 0.059 6:00-7:00 54 0.027 18:00-19:00 103 0.051 7:00-8:00 56 0.028 19:00-20:00 73 0.037 8:00-9:00 81 0.040 20:00-21:00 92 0.046 9:00-10:00 120 0.060 21:00-22:00 75 0.038 10:00-11:00 107 0.053 22:00-23:00 76 0.038 11:00-12:00 117 0.058 23:00-24:00 72 0.036
[0064] The “number of connections” in the above table indicates the number of electrical appliances connected in the current time period, and the “proportion” indicates the proportion of the number of electrical appliances connected in the current time period to the total number of electrical appliances connected in a day.
[0065] Therefore, based on the measured user electricity consumption data in Table 1 and Table 2 above, the undetermined parameters contained in the probability density function can be solved using the nonlinear least squares method.
[0066] S230, substituting the undetermined parameters into a preset probability density function to obtain an initial Gaussian distribution probability density function.
[0067] The preset probability density function may be an expression related to the Gaussian distribution probability density function, which may be expressed as follows:
[0068]
[0069] In the above formula, a n 、b n 、c n is the unknown parameter of the initial Gaussian distribution probability density function, k represents the number of fitting segments, and n represents the order of the unknown parameter.
[0070] According to the nonlinear least squares method, the unknown parameter a is solved n 、b n 、c n Then, the unknown parameters are substituted into formula (2) to obtain the initial Gaussian distribution probability density function.
[0071] The above formula (2) represents the probability density function of Gaussian distribution in discrete state. Assuming that the observed data (t i ,yi ) satisfies the Gaussian distribution, the Gaussian distribution probability density function in the continuous state can be expressed by the following formula:
[0072]
[0073] In the above formula, i represents the variable, t represents the time period, and y represents the Gaussian distribution probability density function value under the continuous state.
[0074] S240, inputting the fitting parameters into the initial Gaussian distribution probability density function to obtain the target Gaussian distribution probability density function, and calculating the fitting result corresponding to the target Gaussian distribution probability density function.
[0075] The fitting parameters include: the number of fitting segments corresponding to the undetermined parameters and the fitting accuracy threshold.
[0076] The number of fitting segments corresponding to the undetermined parameter can be understood as: According to formula (2), the undetermined parameter a n 、b n 、c n It is related to the number of fitting segments k. That is, when k=1, the undetermined parameters a1, b1, c1 need to be determined; when k=2, the undetermined parameters a1, b1, c1, a2, b2, and c2 need to be determined, and so on.
[0077] The number of fitting segments can include: minimum Gaussian fitting segment number K min and the maximum number of Gaussian fitting segments K max ; The fitting accuracy threshold may include: a stop fitting accuracy threshold ε1 and a stop fitting accuracy improvement threshold ε2, wherein ε1 and ε2 satisfy the relationship ε1+ε2=1.
[0078] For example, the values corresponding to the fitting parameters may be as follows: Minimum Gaussian fitting segment number K min =1; Maximum number of Gaussian fitting segments K max =3; stop fitting accuracy threshold ε1 = 99.9%; stop fitting accuracy improvement threshold ε2 = 0.01%. The values of each fitting parameter are not limited to the above examples, but are based on the actual needs of researchers.
[0079] Furthermore, the number of initial fitting segments is k = K min , input the fitting parameters into the initial Gaussian distribution probability density function, and thus obtain the target Gaussian distribution probability density function. Then the target Gaussian distribution probability density function is the value obtained by substituting the fitting parameters and the undetermined parameters into formula (2), and the current value represents the fitting result corresponding to the current fitting segment number.
[0080] Furthermore, the accuracy of the current fitting result can be evaluated, and if the accuracy meets a preset threshold, the current K value is output.
[0081] In an optional embodiment, the building load prediction method provided by the embodiment of the present invention further includes:
[0082] a) Calculate the coefficient of determination of the fitting result corresponding to the number of fitting segments.
[0083] The coefficient of determination (R-square) is selected to evaluate the accuracy of the data fit. It can be expressed as follows:
[0084]
[0085]
[0086]
[0087] In the above formula, ω i represents the weight coefficient, Indicates the corresponding value of the probability density function of the fitted Gaussian distribution, y i represents the original data mean, y i Represents the original data.
[0088] b) Determine whether the coefficient of determination meets the fitting accuracy threshold.
[0089] The normal value range of the R-square coefficient is [0, 1]. The closer the value is to 1, the higher the accuracy of the fitted model and the better the fitting effect. That is, the coefficient is calculated by formula (4)-formula (5) to determine whether it is greater than or equal to the stop fitting accuracy threshold ε1, or less than or equal to the stop fitting accuracy improvement threshold ε2. If yes, execute step c); if not, execute step d).
[0090] c) If the determination coefficient is greater than or equal to the fitting accuracy threshold, the fitting result corresponding to the number of fitting segments is output.
[0091] If the determination coefficient is greater than or equal to the fitting accuracy threshold, it indicates that the fitting result corresponding to the current target Gaussian distribution probability density function is good and meets the preset expectations, and the current fitting result is output.
[0092] d) If the determination coefficient is less than the fitting accuracy threshold, the number of fitting segments and the undetermined parameters are updated, and fitting is performed again until the determination coefficient of the target Gaussian distribution probability density function is greater than or equal to the fitting accuracy threshold, and the fitting result corresponding to the updated number of fitting segments is output.
[0093] If the determination coefficient is less than the fitting accuracy threshold, it means that the fitting result corresponding to the current target Gaussian distribution probability density function does not meet the preset expectations, then the number of fitting segments is updated, that is, k=k+1 or k=k+2, and the corresponding pending number also needs to be updated and recalculated, and step S220 is repeated until the determination coefficient of the target Gaussian distribution probability density function is greater than or equal to the fitting accuracy threshold, and the fitting result corresponding to the updated number of fitting segments is output.
[0094] It should be noted that when updating the number of fitting segments, the specific value of k is not restricted here and is subject to the actual needs of researchers.
[0095] S250, determining algorithm parameters of the Monte Carlo simulation algorithm, and determining scene characteristic parameters of the current building according to user power consumption data.
[0096] S260, inputting the target Gaussian distribution probability density function, algorithm parameters, and scene feature parameters into a Monte Carlo simulation algorithm for simulation.
[0097] S270: Divide the time segment of the user's electricity consumption according to the user's electricity consumption data to obtain multiple time segments.
[0098] The time segmentation of the user's electricity consumption can be performed once every hour based on the user's electricity consumption data, or time segments with different intervals can be obtained based on the time of access to the socket obtained by the smart collection terminal or smart sensor. The specific method of obtaining multiple time segments is not limited here.
[0099] It should be noted that the interval of the current time segment and the number of time segments within a day must be consistent with the time segments of actual electrical appliance access of the current building users obtained during the data collection phase, so as to ensure accurate comparison when predicting the building load in the current time period.
[0100] S280. According to multiple time segments, the target Gaussian distribution probability density function, algorithm parameters, and scene characteristic parameters are input into the Monte Carlo simulation algorithm for segmented simulation to obtain a daily load prediction curve corresponding to the current building.
[0101] Accordingly, based on multiple time segments, the starting power consumption time of the current building can be obtained according to the socket access time distribution, and the load power of the current building can be obtained according to the load power distribution, so that the target Gaussian distribution probability density function, algorithm parameters and scene characteristic parameters are input into the Monte Carlo simulation algorithm for segmented simulation, that is, starting from the starting simulation time, the power consumption status of the sockets in each time segment is simulated. After the simulation of the current time period is completed, the next time segment is transferred to the simulation until all time segments within a day are simulated. The daily load forecast curve corresponding to the current building can be obtained. The current daily load forecast curve describes the power consumption of electrical equipment in the current building within a day.
[0102] S290. When the number of simulations in each time segment reaches a preset threshold, the daily load forecast curve corresponding to the current building is summarized to obtain the daily load forecast result of the current building.
[0103] When each simulation is performed in each time segment, the load forecast result of the current time segment is output. In order to make the simulation result of each time segment sufficiently accurate, the number of simulations for each time segment can reach a preset threshold, for example, 500 times. After the current number of simulations reaches the preset threshold, the average value of the obtained simulation results is calculated to obtain the load forecast result of the current time segment.
[0104] Accordingly, the above method is used to simulate the load forecast results of multiple time segments within a day respectively, and the multiple time segments within a day are summarized and counted, so as to obtain the daily load forecast result of the current building.
[0105] Through the above steps, we can obtain the building load forecasting results that take user behavior into consideration. By utilizing this forecasting result, we can effectively promote the energy management of existing smart buildings and also help provide load forecasting data support for the early planning of building construction projects.
[0106] In an optional embodiment, the scene information of the current building includes an office scene and / or a residential scene. If the scene information of the current building includes an office scene and a residential scene, the user power consumption data includes the user power consumption data of the office scene and the user power consumption data of the residential scene, the target Gaussian distribution probability density function includes the target Gaussian distribution probability density function corresponding to the office scene and the target Gaussian distribution probability density function corresponding to the residential scene, and the scene feature parameters include the scene feature parameters of the office scene and the scene feature parameters of the residential scene. Further, the target Gaussian distribution probability density function, the algorithm parameters, and the scene feature parameters are input into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building, including: inputting the user power consumption data of the office scene, the user power consumption data of the residential scene, the target Gaussian distribution probability density function corresponding to the office scene, the target Gaussian distribution probability density function corresponding to the residential scene, the algorithm parameters, the scene feature parameters of the office scene, and the scene feature parameters of the residential scene into the Monte Carlo simulation algorithm for simulation to obtain the load curve corresponding to the office scene and the load curve corresponding to the residential scene; accumulating the load curve corresponding to the office scene and the load curve corresponding to the residential scene to obtain the load prediction result of the current building.
[0107] It should be noted that when the scene information of the current building is an office scene or a residential scene, the target Gaussian distribution probability density function is the target Gaussian distribution probability density function calculated from the user electricity consumption data corresponding to the office scene or the residential scene, and the corresponding load prediction curve is the load curve corresponding to the office scene or the load curve corresponding to the residential scene; if the scene information of the current building is an office scene and a residential scene, the target Gaussian distribution probability density function is the target Gaussian distribution probability density function calculated from the user electricity consumption data corresponding to the office scene and the residential scene respectively, and the corresponding load prediction curve is the load curve corresponding to the office scene and the load curve corresponding to the residential scene, and the load prediction result of the current building needs to accumulate the load curve corresponding to the office scene and the load curve corresponding to the residential scene.
[0108] On the basis of the above embodiment, the embodiment of the present invention further provides an optional embodiment. The current embodiment takes into account that if the current building contains a large air conditioning refrigeration load, the building load prediction method provided by the embodiment of the present invention, before accumulating the load curve corresponding to the office scene and the load curve corresponding to the residential scene to obtain the load prediction result of the current building, further includes:
[0109] The air conditioning characteristic parameters of the current building are counted, and an air conditioning load model about the air conditioning characteristic parameters is constructed. The air conditioning load model is input into the Monte Carlo simulation algorithm for simulation to obtain the air conditioning load prediction curve corresponding to the current building.
[0110] Air conditioner characteristic parameters may include: number of air conditioners, mean and variance of air conditioner thermal melt and thermal conductivity, mean and variance of air conditioner rated power, etc.
[0111] The expression for constructing the air conditioning load model based on the user's air conditioning characteristic parameters is as follows:
[0112]
[0113] In the above formula, T i and T0 represent the current temperature inside the building and the temperature outside the building respectively, C and K represent the heat transfer and thermal conductivity of the air conditioner inside the building, t represents the time, P e P represents the air conditioning power, s represents the air conditioning state, and α represents the air conditioning energy efficiency ratio. h,other Indicates the heat from other heat sources in the room.
[0114] Correspondingly, the load curve corresponding to the office scene and the load curve corresponding to the residential scene are accumulated to obtain the load prediction result of the current building, including: the load curve corresponding to the office scene, the load curve corresponding to the residential scene and the air conditioning load prediction curve are accumulated to obtain the load prediction result of the current building.
[0115] The process of obtaining the air conditioning load prediction curve may be to substitute the obtained air conditioning characteristic parameters into the above air conditioning load model, and solve the state equation expression by using the Euler method to obtain the corresponding air conditioning load curve.
[0116] The obtained office load curve and / or residential load curve and air conditioning load curve are accumulated to obtain a simulated cumulative load curve within one day, and the current building load curve prediction result is recorded.
[0117] The building load prediction method provided by the embodiment of the present invention uses an intelligent collection terminal or an intelligent sensor to collect a large amount of user electricity consumption data of the current building, which is convenient and fast. The Gaussian fitting distribution is used to show the user electricity consumption behavior characteristics in the corresponding scene of the current building, and the user electricity consumption behavior characteristics are comprehensively considered. It has the advantages of strong applicability and self-determination of the best fitting parameters. Finally, based on the obtained target Gaussian distribution probability density function, algorithm parameters and scene feature parameters, the Monte Carlo simulation algorithm is used for simulation, which can effectively and quickly predict the building load.
[0118] Embodiment 3
[0119] Figure 3 This is a structural block diagram of a building load prediction device provided in the third embodiment of the present invention. The device can be implemented by software and / or hardware, and can generally be integrated in a computer device such as a server. The building load prediction method can be executed to achieve building load prediction. Figure 3 As shown, the device includes: a scene determination module 31, a curve fitting module 32, a parameter determination module 33 and a load prediction module 34, wherein:
[0120] A scene determination module 31 is used to determine scene information of a current building and obtain user power consumption data of the current building according to the scene information;
[0121] A curve fitting module 32, configured to perform Gaussian curve fitting on an initial Gaussian distribution probability density function according to the user power consumption data, to obtain a target Gaussian distribution probability density function corresponding to the scene information;
[0122] A parameter determination module 33, used to determine algorithm parameters of a Monte Carlo simulation algorithm, and to determine scene characteristic parameters of the current building according to the user power consumption data;
[0123] The load prediction module 34 is used to input the target Gaussian distribution probability density function, the algorithm parameters and the scene characteristic parameters into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building.
[0124] The building load prediction device provided in the embodiment of the present invention first determines the scene information of the current building, and obtains the user electricity consumption data of the current building according to the scene information; and performs Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user electricity consumption data to obtain the target Gaussian distribution probability density function corresponding to the scene information; then determines the algorithm parameters of the Monte Carlo simulation algorithm, and determines the scene feature parameters of the current building according to the user electricity consumption data; finally, the target Gaussian distribution probability density function, algorithm parameters and scene feature parameters are input into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building. By adopting the above technical scheme, different user electricity consumption data are obtained according to different scene information of the building, and the user electricity consumption behavior characteristics are taken into account. Further Gaussian fitting is performed on different user electricity consumption data to show the user electricity consumption behavior characteristics, and then the Monte Carlo algorithm is used in combination with different scene feature parameters for simulation, so as to achieve a technical effect of faster load prediction of the current building and more accurate prediction results.
[0125] Optionally, the device further comprises: a calculation module and a substitution module, wherein:
[0126] A calculation module, used to calculate the user's electricity consumption data using a nonlinear least squares method to obtain parameters to be determined;
[0127] The substitution module is used to substitute the undetermined parameters into a preset probability density function to obtain the initial Gaussian distribution probability density function.
[0128] Optionally, the curve fitting module 32 is also used to input the fitting parameters into the initial Gaussian distribution probability density function to obtain the target Gaussian distribution probability density function, and calculate the fitting result corresponding to the target Gaussian distribution probability density function; wherein, the fitting parameters include: the number of fitting segments corresponding to the unknown parameters and the fitting accuracy threshold.
[0129] Optionally, the curve fitting module 32 includes: a calculation unit, an output unit and a fitting unit, wherein:
[0130] A calculation unit, used for calculating the determination coefficient of the fitting result corresponding to the number of fitting segments;
[0131] an output unit, configured to output the fitting result corresponding to the number of fitting segments if the determination coefficient is greater than or equal to the fitting accuracy threshold;
[0132] A fitting unit is used to update the number of fitting segments and the undetermined parameters if the determination coefficient is less than the fitting accuracy threshold, and perform fitting again until the determination coefficient of the target Gaussian distribution probability density function is greater than or equal to the fitting accuracy threshold, and output the fitting result corresponding to the updated number of fitting segments.
[0133] Optionally, the scene information of the current building includes an office scene and / or a residential scene. When the scene information of the current building includes the office scene and the residential scene, the user electricity usage data includes the user electricity usage data of the office scene and the user electricity usage data of the residential scene. The target Gaussian distribution probability density function includes the target Gaussian distribution probability density function corresponding to the office scene and the target Gaussian distribution probability density function corresponding to the residential scene. The scene feature parameters include the scene feature parameters of the office scene and the scene feature parameters of the residential scene.
[0134] The load forecasting module includes: a simulation unit and a load forecasting unit, wherein:
[0135] a simulation unit, configured to input the user electricity consumption data of the office scene, the user electricity consumption data of the residential scene, the target Gaussian distribution probability density function corresponding to the office scene, the target Gaussian distribution probability density function corresponding to the residential scene, the algorithm parameters, the scene characteristic parameters of the office scene, and the scene characteristic parameters of the residential scene into the Monte Carlo simulation algorithm for simulation, so as to obtain a load curve corresponding to the office scene and a load curve corresponding to the residential scene;
[0136] The load prediction unit is used to accumulate the load curve corresponding to the office scene and the load curve corresponding to the residential scene to obtain the load prediction result of the current building.
[0137] Optionally, the load forecasting module further includes: a construction unit and a simulation unit, wherein:
[0138] A construction unit, used to count the user air-conditioning characteristic parameters of the current building and construct an air-conditioning load model related to the user air-conditioning characteristic parameters;
[0139] The simulation unit is used to input the air-conditioning load model into the Monte Carlo simulation algorithm for simulation, so as to obtain the air-conditioning load prediction curve corresponding to the current building.
[0140] The load prediction unit is also used to accumulate the load curve corresponding to the office scene, the load curve corresponding to the residential scene and the air-conditioning load prediction curve to obtain the load prediction result of the current building.
[0141] Optionally, the load forecasting module further includes: a time segmentation unit and a forecast curve acquisition unit, wherein:
[0142] A time segmentation unit, used to segment the time of the user's electricity consumption according to the user's electricity consumption data to obtain multiple time segments;
[0143] A prediction curve obtaining unit, for inputting the target Gaussian distribution probability density function, the algorithm parameters and the scene characteristic parameters into the Monte Carlo simulation algorithm for segmented simulation according to the multiple time segments, to obtain a daily load prediction curve corresponding to the current building;
[0144] The load forecasting unit is also used to summarize the daily load forecasting curve corresponding to the current building when the number of simulations in each time segment reaches a preset threshold, so as to obtain the daily load forecasting result of the current building.
[0145] The building load prediction device provided in the embodiment of the present invention can execute the building load prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.
[0146] Embodiment 4
[0147] An embodiment of the present invention provides a computer device, in which the building load prediction device provided by the embodiment of the present invention can be integrated. Figure 4 The computer device 400 may include a memory 401, a processor 402, and a computer program stored in the memory 401 and executable on the processor. When the processor 402 executes the computer program, the building load prediction method according to the embodiment of the present invention is implemented.
[0148] The computer device provided in the embodiment of the present invention can execute the building load prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.
[0149] Embodiment 5
[0150] An embodiment of the present invention further provides a storage medium containing computer executable instructions, wherein the computer executable instructions are used in a building load forecasting method when executed by a computer processor, the method comprising:
[0151] Determine the scene information of the current building, and obtain the user power consumption data of the current building according to the scene information;
[0152] Performing Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user's electricity consumption data to obtain a target Gaussian distribution probability density function corresponding to the scene information;
[0153] Determining algorithm parameters of a Monte Carlo simulation algorithm, and determining scene characteristic parameters of the current building according to the user power consumption data;
[0154] The target Gaussian distribution probability density function, the algorithm parameters and the scene characteristic parameters are input into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building.
[0155] Storage medium - any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROM, floppy disk or tape device; computer system memory or random access memory, such as DRAM, DDRRAM, SRAM, EDORAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory elements, etc. The storage medium may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the first computer system in which the program is executed, or may be located in a different second computer system, which is connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media that may reside in different locations (for example, in different computer systems connected by a network). The storage medium may store program instructions (for example, embodied as a computer program) that can be executed by one or more processors.
[0156] Of course, the computer executable instructions of a storage medium containing computer executable instructions provided in an embodiment of the present invention are not limited to the building load forecasting operations described above, and can also execute related operations in the building load forecasting method provided in any embodiment of the present invention.
[0157] The building load prediction device, equipment and storage medium provided in the above embodiments can execute the building load prediction method provided in any embodiment of the present invention, and have the corresponding functional modules and beneficial effects of executing the method. For technical details not described in detail in the above embodiments, please refer to the building load prediction method provided in any embodiment of the present invention.
[0158] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims
1. A building load prediction method, characterized in that: include: Determine the scene information of the current building, and obtain the user power consumption data of the current building according to the scene information; Performing Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user's electricity consumption data to obtain a target Gaussian distribution probability density function corresponding to the scene information; Determining algorithm parameters of a Monte Carlo simulation algorithm, and determining scene characteristic parameters of the current building according to the user power consumption data; Inputting the target Gaussian distribution probability density function, the algorithm parameters and the scene feature parameters into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building; Wherein, when the scene information of the current building includes the office scene and the residential scene, the user electricity consumption data includes the user electricity consumption data of the office scene and the user electricity consumption data of the residential scene, the target Gaussian distribution probability density function includes the target Gaussian distribution probability density function corresponding to the office scene and the target Gaussian distribution probability density function corresponding to the residential scene, and the scene feature parameters include the scene feature parameters of the office scene and the scene feature parameters of the residential scene; The step of inputting the target Gaussian distribution probability density function, the algorithm parameters, and the scene feature parameters into the Monte Carlo simulation algorithm for simulation to obtain the load prediction result of the current building includes: Inputting the user electricity consumption data of the office scene, the user electricity consumption data of the residential scene, the target Gaussian distribution probability density function corresponding to the office scene, the target Gaussian distribution probability density function corresponding to the residential scene, the algorithm parameters, the scene characteristic parameters of the office scene and the scene characteristic parameters of the residential scene into the Monte Carlo simulation algorithm for simulation, and obtaining the load curve corresponding to the office scene and the load curve corresponding to the residential scene; wherein the algorithm parameters include: the number of Monte Carlo simulation simulations, the number of simulated time periods within a day, the simulation interval and the number of simulation time periods; The user air conditioning characteristic parameters of the current building are counted, and an air conditioning load model about the user air conditioning characteristic parameters is constructed; the air conditioning load model is input into the Monte Carlo simulation algorithm for simulation, and the air conditioning load prediction curve corresponding to the current building is obtained; wherein the air conditioning characteristic parameters include: the number of air conditioners, the mean and variance of the air conditioning thermal melt and thermal conductivity, and the mean and variance of the air conditioning rated power; the air conditioning load model is constructed by the current temperature inside the building and the temperature outside the building, the thermal melt and thermal conductivity of the air conditioner in the building, time, air conditioning power, air conditioning state, air conditioning energy efficiency ratio, and the heat of other heat sources in the room; The load curve corresponding to the office scene, the load curve corresponding to the residential scene and the air conditioning load prediction curve are accumulated to obtain the load prediction result of the current building.
2. The method according to claim 1, characterized in that: Before performing Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user power consumption data, the method further includes: Calculating the user's electricity consumption data using a nonlinear least squares method to obtain undetermined parameters; Substitute the undetermined parameters into a preset probability density function to obtain the initial Gaussian distribution probability density function.
3. The method according to claim 2, characterized in that The performing Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user power consumption data to obtain the target Gaussian distribution probability density function corresponding to the scene information includes: Inputting the fitting parameters into the initial Gaussian distribution probability density function to obtain the target Gaussian distribution probability density function, and calculating the fitting result corresponding to the target Gaussian distribution probability density function; The fitting parameters include: the number of fitting segments corresponding to the undetermined parameters and a fitting accuracy threshold.
4. The method according to claim 3, characterized in that The method further comprises: Calculating the coefficient of determination of the fitting result corresponding to the number of fitting segments; If the determination coefficient is greater than or equal to the fitting accuracy threshold, outputting the fitting result corresponding to the number of fitting segments; If the determination coefficient is less than the fitting accuracy threshold, the number of fitting segments and the pending parameters are updated, and fitting is performed again until the determination coefficient of the target Gaussian distribution probability density function is greater than or equal to the fitting accuracy threshold, and the fitting result corresponding to the updated number of fitting segments is output.
5. The method according to claim 1, characterized in that The obtaining of the load forecast result of the current building includes: Divide the time segment of the user's electricity consumption according to the user's electricity consumption data to obtain multiple time segments; According to the multiple time segments, the target Gaussian distribution probability density function, the algorithm parameters and the scene characteristic parameters are input into the Monte Carlo simulation algorithm for segmented simulation to obtain a daily load forecast curve corresponding to the current building; When the number of simulations of each of the time segments reaches a preset threshold, the daily load forecast curves corresponding to the current building are summarized to obtain the daily load forecast results of the current building.
6. A building load prediction device, characterized in that: include: A scene determination module, used to determine scene information of a current building, and obtain user power consumption data of the current building according to the scene information; A curve fitting module, used to perform Gaussian curve fitting on the initial Gaussian distribution probability density function according to the user's power consumption data, to obtain a target Gaussian distribution probability density function corresponding to the scene information; A parameter determination module, used to determine the algorithm parameters of the Monte Carlo simulation algorithm, and determine the scene characteristic parameters of the current building according to the user's electricity consumption data; A load forecasting module, used for inputting the target Gaussian distribution probability density function, the algorithm parameters and the scene characteristic parameters into the Monte Carlo simulation algorithm for simulation to obtain the load forecasting result of the current building; Wherein, when the scene information of the current building includes the office scene and the residential scene, the user electricity consumption data includes the user electricity consumption data of the office scene and the user electricity consumption data of the residential scene, the target Gaussian distribution probability density function includes the target Gaussian distribution probability density function corresponding to the office scene and the target Gaussian distribution probability density function corresponding to the residential scene, and the scene feature parameters include the scene feature parameters of the office scene and the scene feature parameters of the residential scene; The load forecasting module includes: a simulation unit, a construction unit and a load forecasting unit, wherein: The simulation unit is used to input the user power consumption data of the office scene, the user power consumption data of the residential scene, the target Gaussian distribution probability density function corresponding to the office scene, the target Gaussian distribution probability density function corresponding to the residential scene, the algorithm parameters, the scene characteristic parameters of the office scene and the scene characteristic parameters of the residential scene into the Monte Carlo simulation algorithm for simulation, so as to obtain the load curve corresponding to the office scene and the load curve corresponding to the residential scene; wherein the algorithm parameters include: the number of Monte Carlo simulation simulations, the number of simulated time periods within a day, the simulation interval and the number of simulation time periods; A construction unit, used to count the user air-conditioning characteristic parameters of the current building and construct an air-conditioning load model related to the user air-conditioning characteristic parameters; A simulation unit, used for inputting the air conditioning load model into the Monte Carlo simulation algorithm for simulation, and obtaining the air conditioning load prediction curve corresponding to the current building; wherein the air conditioning characteristic parameters include: the number of air conditioners, the mean and variance of the air conditioning thermal melt and thermal conductivity, and the mean and variance of the air conditioning rated power; the air conditioning load model is constructed by the current temperature inside the building and the temperature outside the building, the thermal melt and thermal conductivity of the air conditioner in the building, time, air conditioning power, air conditioning state, air conditioning energy efficiency ratio, and the heat of other heat sources in the room; The load prediction unit is used to accumulate the load curve corresponding to the office scene, the load curve corresponding to the residential scene and the air-conditioning load prediction curve to obtain the load prediction result of the current building.
7. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.
