Load prediction and internet access control method and device based on electricity consumption information data

Abstract

The invention relates to the technical field of power electronic prediction, and discloses a load prediction and internet access control method and device based on electricity utilization information data, and the method comprises the steps: obtaining historical electricity utilization information data, and calculating corresponding historical load data; sorting all historical load data in two preset periods before the to-be-predicted day by taking the same type of day as a priority; calculating smoothing values of preset moments of different types of days in the first period and the second period based on the historical load data, and obtaining load values of preset moments of the same type of days in the first period; based on the smooth values of the preset moments of different types of days and the load values of the preset moments of the same type of days in the first period, performing power utilization load prediction on the to-be-predicted days; and performing power generation load prediction on the photovoltaic power generation system based on the power utilization load prediction curve so as to perform network access control on the photovoltaic power generation system. According to the method, the problem of relatively large power grid voltage fluctuation caused by incapability of accurately predicting the electrical load and the photovoltaic power generation load is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power electronic prediction, and in particular to a method and device for load prediction and online control based on electricity consumption information data. Background Art

[0002] With the diversification of electrical equipment in recent years, electricity demand has both certain regularities and a large number of random factors, resulting in uncertainty. Load forecasting, based on the inherent regularity of load changes and taking into account the various random factors that affect load changes, aims to accurately predict the load for a certain period in the future. This has important practical significance for improving the efficiency, safety, economy, and reliability of power system operations and enhancing the power supply quality of the power grid.

[0003] With the increase in electrical equipment and load demands in substations, such as the installation of charging stations and large welding machines, the reserved redundancy in substation transformers is becoming increasingly limited. However, with the installation of photovoltaic power generation equipment, controlling the access of photovoltaic power to the grid to supplement load during peak hours has become a worthy research topic. Forecasting photovoltaic power generation load and adjusting the access to the grid in advance based on forecasted power and photovoltaic loads is also crucial to reduce the impact of high-current equipment on the grid and improve power distribution safety. When high-power equipment starts up, the high initial current can cause a direct drop in grid voltage. Low voltage can cause equipment failure, impacting normal operation. Therefore, load forecasting is essential for both power system planning and automated operation. Summary of the Invention

[0004] In view of this, the present invention provides a load forecasting and online control method and device based on electricity consumption information data to solve the problem of large grid voltage fluctuations caused by the inability to accurately predict electricity load and photovoltaic power generation load.

[0005] In a first aspect, the present invention provides a method for load forecasting and online control based on electricity consumption information data, the method comprising: Obtain historical electricity consumption information data, and calculate corresponding historical load data based on the historical electricity consumption information data; Select the load forecasting day and the preset period before the forecasting day, and sort all historical load data within the preset period before the forecasting day with the same type of days as the priority order, the preset period including the first period and the second period; Based on the sorted historical load data, the exponential smoothing method is used to calculate the smoothed values ​​at the preset time of different types of days in the first cycle and the second cycle, and the load values ​​at the preset time of the same type of days in the first cycle are obtained; Based on the smoothed values ​​at preset times of different types of days in the first and second cycles and the load values ​​at preset times of the same type of days in the first cycle, an electricity load forecast curve is obtained for the electricity load forecast of the to-be-forecasted day; Based on the electricity load prediction curve, the long short-term memory neural network is used to predict the power generation load of the photovoltaic power generation system, and the power generation load prediction curve is obtained to control the access to the grid of the photovoltaic power generation system.

[0006] The load forecasting and access control method based on electricity usage information data provided by the present invention uses periodic analysis and calculation of collected data sequences to predict loads at future time points. Using historical electricity usage data collected over two preset periods, the complete electricity load for the forecasted day can be predicted, generating a prediction curve. Based on the electricity load prediction curve, a long-short-term memory neural network is used to predict the generation load of the photovoltaic power generation system, generating a generation load prediction curve for access control of the photovoltaic power generation system. The electricity load prediction curve and the generation load prediction curve are used to predict the trends and patterns of power load fluctuations over a period of time. This helps power companies to effectively manage power production and allocation, formulate peak-shaving and allocation plans in advance, meet future power demand, and ensure stable operation of the power system. This solves the problem of large grid voltage fluctuations caused by the inability to accurately predict electricity load and photovoltaic power generation load.

[0007] In an optional implementation, obtaining historical electricity usage information data and calculating corresponding historical load data based on the historical electricity usage information data includes: The electricity consumption information data is obtained according to the preset sampling time, the average load within the preset sampling time is calculated based on the electricity consumption information data, and the average load is sorted in the order of historical load days and stored as historical load data.

[0008] The load forecasting and online control method based on electricity consumption information data provided by the present invention obtains electricity consumption information data according to a preset sampling time, calculates the average load within the preset sampling time based on the electricity consumption information data, sorts the average load in the order of historical load days and stores it as historical load data. The obtained historical compliance data is the power load of each sampling point in the historical load day, which provides conditions for subsequent prediction of electricity load according to historical load data.

[0009] In an optional embodiment, sorting all historical load data within a preset period before the day to be predicted based on the same type of days as a priority order includes: The first day of the first cycle and the first day of the second cycle are taken as the same type of days as the first day to be predicted, the second day of the first cycle and the second day of the second cycle are taken as the same type of days as the second day to be predicted, and so on, until all the same type of days as the first and second cycles are determined; The same type of day as the first day to be predicted in the first cycle is ranked first, the same type of day as the first day to be predicted in the second cycle is ranked second, and so on. After the historical load data of the same type of day are sorted, the remaining historical load days are sorted according to their distance from the base date. The base date is the starting date for counting historical load days.

[0010] The load forecasting and online control method based on electricity consumption information data provided by the present invention ranks the same type of day as the first day to be predicted in the first cycle in the first place, ranks the same type of day as the first day to be predicted in the second cycle in the second place, and so on. After the historical load data of the same type of day are sorted, the remaining historical load days are sorted according to their distance from the base date. The base date is the starting date for counting the historical load days. The historical load data is used to find the patterns in its time series, which provides conditions for subsequent electricity load forecasting.

[0011] In an optional embodiment, the exponential smoothing method is used to calculate the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle based on the sorted historical load data, including: Based on the historical load data of the corresponding day in the first cycle and the preset smoothing coefficient, the exponential smoothing method is used to calculate the smoothed values ​​of the preset time of different types of days in the first cycle; Based on the historical load data of the corresponding days in the second period and the preset smoothing coefficient, the exponential smoothing method is used to calculate the smoothing values ​​of the preset moments of different types of days in the second period.

[0012] In an optional embodiment, the power load forecast curve is obtained by performing power load forecasting on the to-be-forecasted day based on the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle and the load values ​​at preset times of the same type of days in the first cycle. The power load forecast curve includes: Calculate the ratio of the smoothed value at the preset time of different types of days in the first period to the smoothed value at the preset time of different types of days in the second period; The power load of the to-be-forecasted day is predicted based on the ratio and the load value at a preset time on the same type of day in the first cycle, and a power load prediction curve is obtained.

[0013] The load forecasting and access control method based on electricity usage data provided by this invention generates a prediction curve based on time-series load data from similar days. This prediction curve predicts the trends and patterns of power load fluctuations over a period of time. This helps power companies optimize power production and allocation, formulate peak load regulation and allocation plans in advance, and meet future power demand, ensuring stable operation of the power system.

[0014] In an optional embodiment, before using a long short-term memory neural network to perform power generation prediction for a photovoltaic power generation system based on a power load prediction curve, the load prediction and access control method based on power consumption information data further includes: Select the corresponding mode coefficient based on the selection mode of the day to be predicted, and the selection mode of the day to be predicted includes working days, holidays and weekends; Obtain the temperature difference between the temperature corresponding to the power load forecast and the temperature standard value; The power load forecast is corrected based on the power load forecast, mode coefficient and temperature difference.

[0015] The load forecasting and grid-connected control method based on electricity consumption information data provided by the present invention uses electricity load forecasting theory to calculate the load of the day to be predicted, and then, based on the corresponding mode of the day to be predicted and the selected mode (holidays, weekends, weekdays), retrieves historical data stored in the database to calculate the corresponding mode coefficient, calculates according to the correction formula, and adds the influence of the difference between the temperature on the prediction day and the experimental temperature to correct the load prediction value. The resulting electricity load is more accurate, providing an accurate data basis for photovoltaic grid-connected control.

[0016] In an optional embodiment, using a long short-term memory neural network to predict the power generation load of a photovoltaic power generation system based on the power load prediction curve to obtain the power generation load prediction curve for controlling the photovoltaic power generation system to be connected to the grid includes: The average load of all photovoltaic devices in the photovoltaic power generation system and the preset sampling time are summarized and summed up, and then aligned according to the time scale to obtain the photovoltaic power generation load summary data; Taking the photovoltaic power generation load summary data as input data, the key characteristics that affect the photovoltaic power generation load are selected and the long short-term memory neural network is used to predict the power generation load of the photovoltaic power generation system, and the power generation load prediction curve is obtained. The photovoltaic power generation system is controlled to be connected to the grid based on the power load forecast curve and the power generation load forecast curve.

[0017] The load forecasting and access control method based on electricity consumption information data provided by the present invention summarizes and sums the average loads of all photovoltaic devices in the photovoltaic power generation system and the preset sampling time, and then aligns them according to the time scale to obtain photovoltaic power generation load summary data; the photovoltaic power generation load summary data is used as input data, and the key characteristics that affect the photovoltaic power generation load are selected to use a long short-term memory neural network to predict the power generation load of the photovoltaic power generation system, thereby achieving accurate prediction of the photovoltaic power generation load. By utilizing on-site collected data and predicted load data, combined with the operation conditions of various distributed energy sources, the final energy management and control decision is made through linear programming technology.

[0018] In a second aspect, the present invention provides a load forecasting and online control device based on electricity consumption information data, the device comprising: A historical data acquisition module is used to obtain historical electricity consumption information data and calculate corresponding historical load data based on the historical electricity consumption information data; A historical load data sorting module is used to select a load forecasting day and a preset period before the forecasting day, and sort all historical load data within the preset period before the forecasting day based on the same type of days as the priority order, wherein the preset period includes the first period and the second period; a smoothing processing module for calculating the smoothed values ​​at preset times of different types of days in the first and second periods respectively using an exponential smoothing method based on the sorted historical load data, and obtaining the load values ​​at preset times of the same type of days in the first period; An electricity load forecasting module is configured to perform electricity load forecasting on a to-be-forecasted day based on the smoothed values ​​at preset times on different types of days in the first and second cycles and the load values ​​at preset times on the same type of days in the first cycle, to obtain an electricity load forecasting curve; The online control module is used to predict the power generation load of the photovoltaic power generation system using a long short-term memory neural network based on the power load prediction curve, and obtain the power generation load prediction curve to control the online access of the photovoltaic power generation system.

[0019] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to thereby execute the load forecasting and Internet access control method based on electricity consumption information data of the above-mentioned first aspect or any corresponding embodiment thereof.

[0020] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the load forecasting and Internet access control method based on electricity consumption information data of the above-mentioned first aspect or any corresponding embodiment thereof.

[0021] In a fifth aspect, the present invention provides a computer program product comprising computer instructions for enabling a computer to execute the load forecasting and Internet access control method based on electricity consumption information data of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 1 is a flow chart of a method for load forecasting and online control based on electricity consumption information data according to an embodiment of the present invention; Figure 2 is a flow chart of another method for load forecasting and online control based on electricity consumption information data according to an embodiment of the present invention; Figure 3 1 is a flow chart of another method for load forecasting and online control based on electricity consumption information data according to an embodiment of the present invention; Figure 4 is a schematic diagram of a day indication of the same type according to an embodiment of the present invention; Figure 5 is a schematic diagram of a flow chart of power load forecast correction according to an embodiment of the present invention; Figure 6 is a schematic diagram of an overview of power load collection according to an embodiment of the present invention; Figure 7 is a schematic diagram of power load prediction according to an embodiment of the present invention; Figure 8 is a schematic diagram of photovoltaic power generation load waveform monitoring according to an embodiment of the present invention; Figure 9 is a schematic diagram of photovoltaic power generation load prediction according to an embodiment of the present invention; Figure 10 2 is a structural block diagram of a load forecasting and online control device based on power consumption information data according to an embodiment of the present invention; Figure 11 Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0025] According to an embodiment of the present invention, an embodiment of a load forecasting and Internet access control method based on electricity consumption information data is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0026] In this embodiment, a load forecasting and Internet access control method based on electricity consumption information data is provided, which can be used in the above-mentioned computer terminals, such as central processing units, servers, etc. Figure 1 FIG. 1 is a flow chart of a method for load forecasting and online control based on power consumption information data according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps: Step S101: Acquire historical electricity usage information data, and calculate corresponding historical load data based on the historical electricity usage information data.

[0027] Specifically, the front-end processor of the power consumption collection system collects and reports all measurement data from smart energy meters connected to terminal devices such as concentrators and / or gateways at preset time intervals. The smart energy meter measurement data, namely, power information data, includes voltage, current, active power, reactive power, and power consumption. The historical load data corresponding to the power information data is categorized and organized according to different reporting data types and corresponding date and time protocols, and stored in a database. The power consumption collection system refers to the source-grid-load-storage microgrid platform on the power consumption collection front-end processor, which obtains smart energy meter measurement data through wireless communication with terminal devices such as concentrators and / or gateways.

[0028] Step S102 , selecting a load forecasting day and a preset period before the forecasting day, and sorting all historical load data within the preset period before the forecasting day with the same type of days as priority, wherein the preset period includes a first period and a second period.

[0029] Specifically, there is no standard for selecting the day to be predicted, but in order to ensure the accuracy of the prediction, historical load data saved for the 14 dates before the day to be predicted must be used. Taking the working day selection mode as an example, the preset period of this embodiment is selected as 14 working days, with 7 working days as a cycle, that is, the first cycle is the first seven days of the preset period, and the second cycle is the last seven days of the preset period. ‌ In load forecasting, similar days refer to historical load days that are similar to the day to be predicted in terms of week type, environment, weather, lunar calendar solar terms, holidays, etc. ‌ These historical load days with a high degree of similarity are called similar days. This embodiment determines similar days based on the high similarity of date and time.

[0030] Step S103 , based on the sorted historical load data, exponential smoothing method is used to calculate the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle, and the load values ​​at preset times of the same type of days in the first cycle are obtained.

[0031] Specifically, the exponential smoothing method is a special weighted moving average method used to predict time series data. The calculation formula of the exponential smoothing method is: ‌ (1); in, represents the smoothed value of period t, represents the actual value of period t-1, represents the smoothed value of period t-1, is the smoothing coefficient, which is used to adjust the impact of recent and long-term data on the forecast value.

[0032] Assume that the preset time is time t, and the load value at the preset time of the same type of day in the first cycle is expressed as P 1t .

[0033] Based on the sorted historical load data, the exponential smoothing method is used to calculate the smoothed values ​​A of different types of preset time periods in the first and second cycles. 1t and A 2t .

[0034] Step S104 , based on the smoothed values ​​at preset times on different types of days in the first and second cycles and the load values ​​at preset times on the same type of days in the first cycle, power load is predicted for the to-be-predicted day to obtain a power load prediction curve.

[0035] Specifically, based on the smoothed values ​​A of different types of days at time t in the first and second cycles 1t and A 2t And the load value P at time t of the same type of day in the first cycle 1t Calculate the power load for the forecasted day , predict the electricity load at multiple times t on multiple dates, and fit the multiple electricity loads to form an electricity load prediction curve.

[0036] Step S105 , based on the power load prediction curve, a long short-term memory neural network is used to predict the power generation load of the photovoltaic power generation system, and a power generation load prediction curve is obtained to control the photovoltaic power generation system to be connected to the grid.

[0037] Specifically, the Long Short-Term Memory (LSTM) neural network is a time-recurrent neural network designed to solve the long-term dependency problem of general RNN (Recurrent Neural Network). All RNNs have a chain form of repeated neural network modules.

[0038] The LSTM model is introduced, and a predicted load curve is generated by inputting parameters such as the monitored power generation load, temperature, humidity, and wind speed. This model provides precise predictions and, by learning relevant parameters, can increasingly accurately predict power generation load trends, generating a load forecast curve that can be used to control the accessibility of photovoltaic power generation systems.

[0039] The load forecasting and access control method based on electricity information data provided in this embodiment uses a collected data sequence to perform periodic analysis and calculation to predict the load at a future time point. Using historical electricity consumption data collected over two cycles of a preset number of days, the complete electricity load for the predicted day can be predicted, generating a prediction curve. Based on the electricity load prediction curve, a long-short-term memory neural network is used to predict the power generation load of the photovoltaic power generation system, resulting in a power generation load prediction curve for access control of the photovoltaic power generation system. The electricity load prediction curve and the power generation load prediction curve are used to predict the trends and patterns of power load fluctuations over a period of time in the future. This helps power companies to effectively manage power production and allocation, formulate peak-shaving and allocation plans in advance, meet future power demand, and ensure stable operation of the power system. This solves the problem of large grid voltage fluctuations caused by the inability to accurately predict electricity load and photovoltaic power generation load.

[0040] In this embodiment, a load forecasting and Internet access control method based on electricity consumption information data is provided, which can be used in computer terminals such as central processing units, servers, etc. Figure 2 FIG. 1 is a flow chart of a method for load forecasting and online control based on power consumption information data according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps: Step S201: Acquire historical electricity usage information data, and calculate corresponding historical load data based on the historical electricity usage information data.

[0041] Specifically, the above step S201 includes: Step a. Obtain electricity usage information data according to a preset sampling time, calculate the average load within the preset sampling time based on the electricity usage information data, sort the average load in chronological order of historical load days and store them as historical load data.

[0042] Specifically, the preset sampling time is selected according to the actual situation and is not specifically limited here. The preset sampling time in this embodiment is every minute. Assume that the electricity consumption information data of the smart energy meter is collected every minute and stored in the database. The electricity consumption information data is recorded as E ( t ),in t Indicates time (minutes). Use the sliding difference method to calculate the average power, with a window size of T minutes. P ( t , T ) indicates the time from t The average power in the first T minutes is: (2); The average power within the current slip time is stored in date and time order. This average power is the historical load data that needs to be called later.

[0043] Step S202 , selecting a load forecasting day and a preset period before the forecasting day, and sorting all historical load data within the preset period before the forecasting day with the same type of days as priority, wherein the preset period includes a first period and a second period.

[0044] Specifically, the above step S202 includes: In step S2021, the first day of the first cycle and the first day of the second cycle are taken as days of the same type as the first day to be predicted, the second day of the first cycle and the second day of the second cycle are taken as days of the same type as the second day to be predicted, and so on, until all days of the same type as the days to be predicted in the first cycle and the second cycle are determined.

[0045] For example, Figure 4 As shown in the figure, take the forecast date as the 15th day, and push back 14 days in the past, with each 7 days as a cycle. For example, if the forecast date is the 28th of a month, and the 18th and 19th of the month are weekends, then the working days pushed back 14 days are: The first cycle is D17, D16, D15, D14, D13, D12, D11; The second cycle is D27, D26.D25.D24.D23.D22.D21.

[0046] Among them, D11 and D21, D12 and D22, D13 and D23, D14 and D24, D15 and D25, D16 and D26, D17 and D27 are all called days of the same type.

[0047] The power corresponding to each day in the first cycle is: P7t, P6t, P5t P4t, P3t, P2t, P1t; The power corresponding to each day in the second cycle is: P14t, P13t, P12t, P11t, P10t, P9t, P8t.

[0048] In step S2022, the same type of day as the first day to be predicted in the first cycle is ranked first, the same type of day as the first day to be predicted in the second cycle is ranked second, and so on. After the historical load data of the same type of day are sorted, the remaining historical load days are sorted according to their distance from the base date, and the base date is the starting date for counting historical load days.

[0049] Specifically, based on the historical load data corresponding to the electricity consumption information data, the starting point of the historical date counting is taken as the base date. Starting from the base date, with a 7-day cycle, count forward one day in sequence. Let the load at the t-th moment on the d-th day be P dt , where d = 1, 2…, n, t = 1, 2,… T (T is the number of sampling points per day, i.e., the sampling time interval). The daily load curve is expressed as a set as follows: C d =[P d1 ,P d2 ,…,P dt ](3); All C d The set of is called the historical load sample set.

[0050] All historical load samples are sorted in order of priority based on the same type of days, that is, the load days with the same type as the forecast day in the first cycle are ranked first, the same type of days in the second cycle are ranked second, and so on; when the same type of days are sorted, the remaining historical load days are sorted according to their distance from the base day. Thus, the original n-day historical load curves C1, C2…C n Sort by D1, D2…, D n Obviously, any C (d = 1, 2…, n) must correspond to a certain D (i = 1, 2…, n), and the two are in a one-to-one correspondence. Therefore, in the set of historical load samples D, the smaller the subscript i, the higher the degree of correlation with the day to be predicted. Di = [P i1 ,P i2 ,…P it ], the characteristic parameters of the i-th day are: Maximum daily load P imax = max(P i1 ,P i2 ,…P it ).

[0051] Daily minimum load P imin = min(P i1 ,P i2 ,…P it ).

[0052] Daily average load P iavg = average(P i1 ,P i2 ,…P it ).

[0053] Step S203 : Based on the sorted historical load data, exponential smoothing method is used to calculate the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle, and the load values ​​at preset times of the same type of days in the first cycle are obtained.

[0054] Specifically, the above step S203 includes: Step S2031 : Based on the historical load data of the corresponding days in the first cycle and the preset smoothing coefficient, the exponential smoothing method is used to calculate the smoothed values ​​at preset times of different types of days in the first cycle.

[0055] Specifically, the load forecast value at a certain moment on the forecast day is the result of an exponential smoothing of the load values ​​at the same moment on the relevant recent days.

[0056] Select relevant historical load data for n=14 days, and the smoothed value calculation formula at time t of different types of days in the first cycle is as follows: A 1t =βP 3t +β(1-β)P 4t +…+β(1-β) 5 P 8t (4); Among them, β is the smoothing coefficient of point-by-point load, and the general value of β is [0.2, 0.6]. 3t 、P 4t 、P 8t These are the historical load data at time t corresponding to the 3rd, 4th and 8th days respectively.

[0057] Step S2032 : Based on the historical load data of the corresponding days in the second period and the preset smoothing coefficient, the exponential smoothing method is used to calculate the smoothed values ​​at preset times of different types of days in the second period.

[0058] Specifically, the calculation formula for the smoothed value at the preset time of different types of days in the second period is as follows: A 2t =βP 9t +β(1-β)P 10t +…+β(1-β) 5 P 14t (5); Among them, P 9t 、P 10t 、P 14t These are the historical load data at time t corresponding to the 9th, 10th and 14th days respectively.

[0059] Step S204 , based on the smoothed values ​​at preset times on different types of days in the first and second cycles and the load values ​​at preset times on the same type of days in the first cycle, power load is predicted for the to-be-predicted day to obtain a power load prediction curve.

[0060] Specifically, the above step S204 includes: Step S2041 , calculating the ratio of the smoothed values ​​at the preset time of different types of days in the first cycle to the smoothed values ​​at the preset time of different types of days in the second cycle.

[0061] Specifically, the value P of the same type of day at time t in the first cycle 1t , then: (6).

[0062] Step S2042 : Based on the ratio and the load value at a preset time on the same type of day in the first cycle, the power load is predicted for the to-be-predicted day to obtain a power load prediction curve.

[0063] Specifically, the power load forecast value at time t on the forecasted day is obtained according to formula (6), namely: (7); in, Select the electricity load forecast value for the weekday mode.

[0064] Step S205: Based on the power load forecast curve, the long short-term memory neural network is used to forecast the power generation load of the photovoltaic power generation system, and the power generation load forecast curve is obtained to control the photovoltaic power generation system to be connected to the grid. Figure 1 Step S105 of the illustrated embodiment will not be described in detail here.

[0065] The load forecasting and access control method based on electricity information data provided in this embodiment obtains electricity information data according to a preset sampling time, calculates the average load within the preset sampling time based on the electricity information data, sorts the average load in the order of historical load days and stores it as historical load data. The historical compliance data obtained is the power load of each sampling point in the historical load day. The load forecasting and access control method based on electricity information data ranks the same type of day as the first day to be predicted in the first cycle in the first place, ranks the same type of day as the first day to be predicted in the second cycle in the second place, and so on. After the historical load data of the same type of day is sorted, the remaining historical load days are sorted according to their distance from the base date. The base date is the starting date for counting historical load days. The historical load data is used to find the patterns in its time series, providing conditions for subsequent electricity load forecasting.

[0066] In this embodiment, a load forecasting and Internet access control method based on electricity consumption information data is provided, which can be used in computer terminals such as central processing units, servers, etc. Figure 3 FIG. 1 is a flow chart of a method for load forecasting and online control based on power consumption information data according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps: Step S301: Obtain historical electricity usage information data, and calculate corresponding historical load data based on the historical electricity usage information data. Figure 2 Step S201 of the illustrated embodiment will not be described in detail here.

[0067] Step S302: Select the load forecast day and the preset period before the forecast day, and sort all historical load data within the preset period before the forecast day with the same type of days as the priority order. The preset period includes the first period and the second period. Figure 2 Step S202 of the illustrated embodiment will not be described in detail here.

[0068] Step S303: Based on the sorted historical load data, the exponential smoothing method is used to calculate the smoothed values ​​of different types of days at preset times in the first and second cycles, and the load values ​​of the same type of days at preset times in the first cycle are obtained. Figure 2 Step S203 of the illustrated embodiment will not be described in detail here.

[0069] Step S304: Based on the smoothed values ​​at preset times of different types of days in the first and second cycles and the load values ​​at preset times of the same type of days in the first cycle, the power load is predicted for the to-be-predicted day to obtain a power load prediction curve. Figure 2 Step S204 of the illustrated embodiment will not be described in detail here.

[0070] Step S305, selecting a corresponding mode coefficient based on the mode of the day to be predicted, where the mode of the day to be predicted includes weekdays, holidays, and weekends (also known as weekends); obtaining the temperature difference between the temperature corresponding to the power load forecast and the temperature standard value; and correcting the power load forecast based on the power load forecast, the mode coefficient, and the temperature difference.

[0071] Specifically, after the prediction result is calculated, the prediction value is corrected according to the mode selected by the user. Assume that the weekday mode is marked with modelflg 0, the weekend mode is marked with modelflg 1, and the holiday mode is marked with modelflg 2. The power load forecast correction process is as follows: Figure 5 shown.

[0072] According to the mode selected on the forecast day, the historical load data Pth (Ptw) at the time t when the previous mode modelflg is equal is called back, and the historical load data Pth (Ptw) at the time t when modelflg = 0 on the previous day is called back. tu If the prediction day mode is a holiday, then modelflg = 2, and the modified power load curve coefficient K = P th / P tu , the empirical value range of K is K∈[0.548,0.601]. If the forecast day mode is weekend, modelflg=1, the weekend curve coefficient α=P tw / P tu , the empirical value range of α is α∈[0.705,0.813].

[0073] According to the temperature corresponding to the season and the expected power load forecast and the laboratory temperature standard value (23°C), the power load of the park increases by 1% for every 1°C increase or decrease in temperature. Based on the temperature difference, the load curve shifts up by 1%. The corrected value of the power load forecast is: or (8); Among them, ΔT is the difference between the predicted temperature and the standard temperature value, K is the holiday curve coefficient, and α is the weekend curve coefficient. If holidays and weekends overlap, the holiday coefficient shall prevail and the holiday mode shall take precedence.

[0074] This process is repeated for time t=1, 2,…, T in sequence to obtain the power load forecast curve for each time of the day.

[0075] Step S306 , based on the power load prediction curve, a long short-term memory neural network is used to predict the power generation load of the photovoltaic power generation system, and a power generation load prediction curve is obtained to control the photovoltaic power generation system to be connected to the grid.

[0076] Specifically, the above step S306 includes: Step S3061 , summing up all photovoltaic devices in the photovoltaic power generation system and the average load within a preset sampling time, and aligning them according to the time scale to obtain photovoltaic power generation load summary data.

[0077] Specifically, the calculation formula for the summary data of photovoltaic power generation load is as follows: (9); Where T is the preset sampling time, and N is the number of photovoltaic devices.

[0078] Step S3062: Taking the photovoltaic power generation load summary data as input data, selecting key characteristics that affect the photovoltaic power generation load, and using a long short-term memory neural network to predict the photovoltaic power generation load, a power generation load prediction curve is obtained.

[0079] Specifically, the above-mentioned data is used as input variables, and according to the load characteristics, key features that affect the photovoltaic load, such as temperature, humidity, wind speed, etc., are selected. Correlation analysis and principal component analysis are used to select features, and the LSTM (long short-term memory neural network) model is used to obtain the predicted load curve. , the formula is: (10).

[0080] Step S3063: Controlling the photovoltaic power generation system to be connected to the grid based on the power load forecast curve and the power generation load forecast curve.

[0081] Specifically, by predicting the power consumption and power generation loads, recording the startup, operation, and termination characteristics of heavy-load equipment on the grid, and switching photovoltaic circuit breakers during peak hours, the integration of photovoltaic power generation into the grid can effectively offset the impact of high loads on the grid. The predicted photovoltaic power generation load can determine which micro-photovoltaic circuit breakers to close based on the power consumption and the load value to be supplemented, enabling more accurate power replenishment. The control logic remotely closes or opens the photovoltaic circuit breakers based on the predicted power load, controlling the photovoltaic power generation system to supplement the grid and ensuring 100% photovoltaic power supply replenishment. During low power consumption periods, the photovoltaic circuit breakers can be promptly disconnected to ensure stable grid operation.

[0082] The load forecasting and access control method based on electricity information data provided in this embodiment uses electricity load forecasting theory to calculate the load for the day to be predicted. Then, based on the corresponding mode of the day to be predicted, historical data stored in the database is retrieved according to the selected mode (holiday, weekend, weekday) to calculate the corresponding mode coefficient. The correction formula is used to perform the calculation, and the influence of the difference between the predicted day temperature and the experimental temperature is added to correct the load forecast value, resulting in a more accurate electricity load. The average load of all photovoltaic devices in the power grid and the load within a preset sampling time are summarized and summed, and then aligned according to the time scale to obtain photovoltaic power generation load summary data. The photovoltaic power generation load summary data is used as input data, and key characteristics that affect the photovoltaic power generation load are selected to use a long short-term memory neural network to predict the power generation load of the photovoltaic power generation system. This achieves accurate prediction of the power generation load of the photovoltaic power generation system. The final energy management and control decision is made using linear programming technology, utilizing field-collected data and predicted load data, combined with the operating conditions of various distributed energy resources.

[0083] As one or more specific application embodiments of the present invention, combined with Figures 7 to 10 The load forecasting and access control method based on power consumption information data provided by the present invention is further described in detail as follows: This example only forecasts industrial park electricity consumption and photovoltaic power generation within the park, and does not consider the complex grid supply scenarios supported by wind and thermal power. Based on the park's electricity usage characteristics, forecast modes for weekdays, weekends, and holidays are added. Furthermore, considering that the predicted daily electricity consumption characteristics may be affected by seasonal temperature fluctuations, the predicted daily temperature input is added to achieve the purpose of revising the prediction curve.

[0084] In this embodiment, the power consumption information collection system server is connected to the gateway or collection terminal in a wireless manner, and the electricity meter and photovoltaic micro circuit breaker communicate with the gateway or collection terminal in a wired manner (such as RS485 or Modbus), ensuring that the power consumption information collection system can collect the equipment power consumption information data.

[0085] The gateway or collection terminal reports the collected data at the default interval (default 15 minutes), and the electricity consumption information collection system stores the data in the database in the order of the date and time of receipt.

[0086] Step 1) Overview of electricity consumption information collection Figure 6 As shown: Historical data (database storage data) can be viewed through the Source-Grid-Load-Storage Microgrid Platform. Enter any date before the current date on the Query Electricity Parameter Details page. The Source-Grid-Load-Storage Microgrid Platform will display the historical load curve based on the selected date and data type on the details interface.

[0087] Step 2) The schematic diagram of power load forecast is as follows Figure 7 As shown below: Enter the power load forecast service window, select the date for load forecasting (i.e., the day to be forecasted), the mode to be selected for the day to be forecasted, and the expected temperature for the day to be forecasted. The Source-Grid-Load-Storage Microgrid Platform uses the time series method to calculate the forecast curve based on the power load forecast model and displays it on the forecast page. The curve hash point interval is based on the reporting interval of the data collection terminal.

[0088] Step 3) Power generation load monitoring data: Monitor the voltage, current, active power, reactive power, power factor, frequency, power consumption and other data of the photovoltaic equipment main line, and report to the platform every 15 minutes.

[0089] Step 4) Power generation load prediction: Based on the LSTM model, the power generation load is predicted by inputting parameters such as temperature, humidity, and wind speed. By learning related parameters, the trend change of power generation load can be predicted more and more accurately. The schematic diagrams of photovoltaic power generation load waveform monitoring and photovoltaic power generation load prediction are shown as follows: Figure 8 and Figure 9 shown.

[0090] This embodiment uses time series data to identify patterns within the historical load data. Using this data, it forecasts load demand based on time series data. Using the auto-regressive averaging method, it can predict the curve 15 days out. The predicted load value is then corrected based on parameters such as holidays, weekend patterns, and expected temperature, resulting in a relatively accurate prediction of the future load curve. Compared to methods that directly calculate predictions based on holiday data from the previous two cycles (14 periods), this method offers the advantage of using a benchmark that is closer to the current load situation. The reference data source has a shorter time span, providing greater guidance, and is applicable to forecasts of various load types. An LSTM model is introduced to generate a predicted load curve based on inputs such as monitored power generation load, temperature, humidity, and wind speed. This model provides accurate predictions and, by learning relevant parameters, can increasingly accurately predict power generation load trends. By predicting load demand and recording the startup, operation, and shutdown characteristics of heavy-load equipment on the grid, photovoltaic circuit breakers are activated during peak periods to ensure 100% grid replenishment by photovoltaic power sources. During off-peak periods, photovoltaic circuit breakers are promptly activated to ensure stable grid operation.

[0091] This embodiment also provides a load forecasting and online control device based on electricity usage information data. This device is used to implement the above-mentioned embodiments and preferred embodiments, and the details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0092] This embodiment provides a load forecasting and access control device based on electricity consumption information data, such as Figure 10 As shown, including: The historical data acquisition module 1101 is used to acquire historical power usage information data and calculate corresponding historical load data based on the historical power usage information data.

[0093] The historical load data sorting module 1102 is used to select the load forecast day and the preset period before the forecast day, and sort all historical load data within the preset period before the forecast day based on the same type of days as the priority order. The preset period includes the first period and the second period.

[0094] The smoothing processing module 1103 is used to calculate the smoothed values ​​of different types of days at preset times in the first and second cycles using exponential smoothing method based on the sorted historical load data, and obtain the load values ​​of the same type of days at preset times in the first cycle.

[0095] The power load forecasting module 1104 is configured to forecast the power load for the to-be-forecasted day based on the smoothed values ​​at preset times on different types of days in the first and second cycles and the load values ​​at preset times on the same type of days in the first cycle, to obtain a power load forecasting curve.

[0096] The online control module 1105 is used to predict the power generation load of the photovoltaic power generation system using a long short-term memory neural network based on the power load prediction curve, and obtain the power generation load prediction curve to control the online access of the photovoltaic power generation system.

[0097] In some optional implementations, the historical data acquisition module 1101 includes: The historical load data calculation unit is used to obtain power information data according to a preset sampling time, calculate the average load within the preset sampling time based on the power information data, sort the average load in the order of historical load days and store it as historical load data.

[0098] In some optional implementations, the historical load data sorting module 1102 includes: The same type day determination unit is used to use the first day of the first cycle and the first day of the second cycle as the same type days of the first day to be predicted, and the second day of the first cycle and the second day of the second cycle as the same type days of the second day to be predicted, and so on, until the same type days of all the days to be predicted in the first cycle and the second cycle are determined.

[0099] The historical load data sorting unit is used to sort the same type of days as the first day to be predicted in the first cycle in the first place, and the same type of days as the first day to be predicted in the second cycle in the second place, and so on. When the historical load data of the same type of days are sorted, the remaining historical load days are sorted according to their distance from the base date, and the base date is the starting date for counting historical load days.

[0100] In some optional implementations, the smoothing processing module 1103 includes: The first cycle smoothing value calculation unit is used to calculate the smoothing values ​​of different types of days at preset moments in the first cycle using an exponential smoothing method based on the historical load data of the corresponding day in the first cycle and a preset smoothing coefficient.

[0101] The second period smoothing value calculation unit is used to calculate the smoothing values ​​of different types of days at preset moments in the second period using an exponential smoothing method based on the historical load data of the corresponding day in the second period and a preset smoothing coefficient.

[0102] In some optional implementations, the power load forecasting module 1104 includes: The ratio calculation unit is used to calculate the ratio of the smoothed value at the preset time of different types of days in the first cycle to the smoothed value at the preset time of different types of days in the second cycle.

[0103] The power load forecast value calculation unit is used to predict the power load on the to-be-forecasted day based on the ratio and the load value at a preset time on the same type of day in the first cycle, and obtain a power load forecast curve.

[0104] In some optional implementations, the Internet access control module 1105 includes: The photovoltaic power generation load summary unit is used to summarize and sum all photovoltaic devices in the power grid and the average load within a preset sampling time, and then align them according to the time scale to obtain photovoltaic power generation load summary data.

[0105] The power generation load prediction unit is used to take the photovoltaic power generation load summary data as input data, select key characteristics that affect the photovoltaic power generation load, and use a long short-term memory neural network to predict the power generation load of the photovoltaic power generation system to obtain a power generation load prediction curve.

[0106] The photovoltaic grid-connected control unit is used to control the photovoltaic power generation system to be grid-connected based on the power load forecast curve and the power generation load forecast curve.

[0107] In some optional implementations, the load forecasting and Internet access control device based on electricity consumption information data further includes: The electricity load forecast correction module is used to select the corresponding mode coefficient based on the selection mode of the day to be predicted, which includes weekdays, holidays and weekends; obtain the temperature difference between the temperature corresponding to the electricity load forecast and the temperature standard value; and correct the electricity load forecast based on the electricity load forecast, mode coefficient and temperature difference.

[0108] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0109] The load forecasting and online control device based on electricity consumption information data in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0110] The embodiment of the present invention also provides a computer device having the above Figure 10 The load forecasting and access control device based on electricity consumption information data is shown.

[0111] See also Figure 11 , Figure 11 is a structural diagram of a computer device provided by an optional embodiment of the present invention, such as Figure 11As shown, the computer device includes: one or more processors 10, memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 11 A processor 10 is taken as an example.

[0112] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0113] The memory 20 stores instructions that can be executed by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0114] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0115] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0116] The computer device further includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 11 The bus connection is taken as an example.

[0117] The input device 30 can receive input digital or character information and generate key signal input related to user settings and function control of the computer device. Examples include a touch screen, keypad, mouse, trackpad, touchpad, pointing stick, one or more mouse buttons, trackball, joystick, etc. The output device 40 may include a display device, auxiliary lighting devices (e.g., LEDs), and tactile feedback devices (e.g., vibration motors). Such display devices include, but are not limited to, liquid crystal displays, light emitting diodes, monitors, and plasma displays. In some optional embodiments, the display device may be a touch screen.

[0118] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0119] A portion of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0120] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for load forecasting and online control based on power consumption information data, characterized in that: The method comprises: Acquire historical power consumption information data, and calculate corresponding historical load data based on the historical power consumption information data; Selecting a load prediction day and a preset period before the prediction day, sorting all historical load data within the preset period before the prediction day with the same type of days as the priority order, wherein the preset period includes a first period and a second period; Based on the sorted historical load data, the exponential smoothing method is used to calculate the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle, and the load values ​​at preset times of the same type of days in the first cycle are obtained; Based on the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle and the load values ​​at preset times of the same type of days in the first cycle, the power load forecasting curve is obtained by forecasting the power load on the to-be-forecasted day; Based on the power load prediction curve, the long short-term memory neural network is used to predict the power generation load of the photovoltaic power generation system, and the power generation load prediction curve is obtained to control the access to the grid of the photovoltaic power generation system.

2. The method according to claim 1, characterized in that The acquiring of historical power usage information data and calculating corresponding historical load data based on the historical power usage information data includes: The power consumption information data is acquired at a preset sampling time, the average load within the preset sampling time is calculated based on the power consumption information data, and the average load is sorted in the order of historical load days and stored as historical load data.

3. The method according to claim 2, characterized in that The method of sorting all historical load data within a preset period before the day to be predicted by taking the same type of days as the priority order includes: The first day of the first cycle and the first day of the second cycle are taken as the same type of days as the first day to be predicted, the second day of the first cycle and the second day of the second cycle are taken as the same type of days as the second day to be predicted, and so on, until all the same type of days as the first and second cycles are determined; The same type of day as the first day to be predicted in the first cycle is ranked first, the same type of day as the first day to be predicted in the second cycle is ranked second, and so on. When the historical load data of the same type of days are sorted, the remaining historical load days are sorted according to their distance from the base date, and the base date is the starting date for counting historical load days.

4. The method according to claim 1, characterized in that: The method of using the exponential smoothing method based on the sorted historical load data to calculate the smoothed values ​​of different types of day preset times in the first cycle and the second cycle includes: Based on the historical load data of the corresponding day in the first cycle and the preset smoothing coefficient, the exponential smoothing method is used to calculate the smoothing values ​​at preset times of different types of days in the first cycle; Based on the historical load data of the corresponding day in the second period and the preset smoothing coefficient, the exponential smoothing method is used to calculate the smoothing values ​​at preset times of different types of days in the second period.

5. The method according to claim 4, characterized in that The power load forecasting curve obtained by forecasting the power load for the forecast day based on the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle and the load values ​​at preset times of the same type of days in the first cycle includes: Calculate the ratio of the smoothed value at the preset time of different types of days in the first period to the smoothed value at the preset time of different types of days in the second period; Based on the ratio and the load value at a preset time on the same type of day in the first cycle, the power load on the to-be-predicted day is predicted to obtain a power load prediction curve.

6. The method according to claim 1, characterized in that Before using the long short-term memory neural network to predict photovoltaic power generation based on power load prediction, the method also includes: Selecting a corresponding mode coefficient based on a selection mode of a day to be predicted, wherein the selection mode of a day to be predicted includes working days, holidays and weekends; Obtain the temperature difference between the temperature corresponding to the power load forecast and the temperature standard value; The power load forecast is corrected based on the power load forecast, mode coefficient and temperature difference.

7. The method according to claim 2, characterized in that The method of using a long short-term memory neural network to predict the power generation load of a photovoltaic power generation system based on the power load prediction curve to obtain a power generation load prediction curve to control the photovoltaic power generation system to be online includes: The average loads of all photovoltaic devices in the photovoltaic power generation system and the preset sampling time are summarized and summed up, and then aligned according to the time scale to obtain the photovoltaic power generation load summary data; The photovoltaic power generation load summary data is used as input data, and key characteristics affecting the photovoltaic power generation load are selected to use a long short-term memory neural network to predict the power generation load of the photovoltaic power generation system, thereby obtaining a power generation load prediction curve; The photovoltaic power generation system is controlled to be connected to the grid based on the power load forecast curve and the power generation load forecast curve.

8. A load forecasting and online control device based on power consumption information data, characterized in that: The device comprises: A historical data acquisition module, used to acquire historical power consumption information data, and calculate corresponding historical load data based on the historical power consumption information data; A historical load data sorting module is used to select a load prediction day and a preset period before the prediction day, and sort all historical load data within the preset period before the prediction day with the same type of days as the priority order, wherein the preset period includes a first period and a second period; A smoothing processing module, used to calculate the smoothing values ​​at preset times of different types of days in the first cycle and the second cycle respectively based on the sorted historical load data by using an exponential smoothing method, and obtain the load values ​​at preset times of the same type of days in the first cycle; An electricity load forecasting module is used to forecast the electricity load on the forecast day based on the smoothed values ​​at preset times of different types of days in the first cycle and the second cycle and the load values ​​at preset times of the same type of days in the first cycle, to obtain an electricity load forecasting curve; The online control module is used to predict the power generation load of the photovoltaic power generation system using a long short-term memory neural network based on the power load prediction curve, and obtain the power generation load prediction curve to control the online access of the photovoltaic power generation system.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the load forecasting and Internet access control method based on power consumption information data as described in any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the load forecasting and Internet access control method based on power consumption information data according to any one of claims 1 to 7.

Images