Transformer load prediction method, device, and electronic equipment

Abstract

The present invention discloses a transformer load prediction method, device, and electronic device thereof, relating to the field of power grid safety technology. The prediction method comprises: collecting transformer load data within a target time period, preprocessing the load data, and obtaining preprocessed load data; calculating a trend coefficient for each time point within the target time period based on the preprocessed load data; predicting the trend coefficient to obtain a predicted trend coefficient for each time point in a future time period; calculating a seasonal coefficient for the transformer, and predicting the predicted load data for the transformer in the future time period based on the preprocessed load data, the predicted trend coefficient, and the seasonal coefficient. The present invention solves the technical problem in the related art of using an exponential smoothing method to configure weight values ​​and then predict transformer load, resulting in low prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the field of power grid safety technology, and in particular to a transformer load prediction method, a device thereof, and electronic equipment. Background Art

[0002] Transformer load forecasting is the process of estimating or predicting transformer load demands over a period of time. This process is crucial for power system planning, operation, and dispatch. It helps power systems rationally allocate resources, improve grid efficiency, reduce operating costs, and ensure the reliability and stability of power supply.

[0003] In related technologies, when predicting the load of a transformer, the exponential smoothing method can be used to exponentially reduce the weights of different time points so as to pay more attention to recent data, thereby predicting the transformer load in a short period of time. However, the exponential smoothing method cannot effectively capture long-term change trends, resulting in low accuracy in the results of transformer load prediction.

[0004] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0005] The embodiments of the present invention provide a transformer load prediction method and device, and an electronic device thereof, to at least solve the technical problem in the related art of predicting transformer load by configuring weight values ​​using an exponential smoothing method, resulting in low prediction accuracy.

[0006] According to one aspect of an embodiment of the present invention, a transformer load prediction method is provided, comprising: collecting load data of the transformer within a target time period, and preprocessing the load data to obtain preprocessed load data; calculating a trend coefficient at each time point within the target time period based on the preprocessed load data; predicting the trend coefficient to obtain a predicted trend coefficient at each time point in a future time period; calculating a seasonal coefficient of the transformer, and predicting predicted load data of the transformer in the future time period based on the preprocessed load data, the predicted trend coefficient, and the seasonal coefficient.

[0007] Optionally, the step of calculating the trend coefficient of each time point within the target time period based on the preprocessed load data includes: step one, selecting a time point as the target observation point, and obtaining the load value of the target observation point; step two, obtaining the predicted load value of the historical observation point corresponding to the target observation point; step three, configuring a smoothing coefficient for the load value of the target observation point; step four, calculating the trend coefficient of the target observation point based on the load value of the target observation point, the predicted load value of the historical observation point and the smoothing coefficient; repeating steps one to four until the trend coefficient of each time point within the target time period is calculated.

[0008] Optionally, the step of predicting the trend coefficient to obtain the predicted trend coefficient for each time point in the future time period includes: step one, constructing a linear regression model based on the trend coefficient and the time point corresponding to the trend coefficient; step two, calculating the first regression coefficient and the second regression coefficient of the linear regression model; step three, calculating the predicted trend coefficient of the target time point based on the first regression coefficient, the second regression coefficient and the target time point to be predicted; repeating steps two to three until the residual sum of squares of the linear regression model reaches a minimum value, thereby obtaining the predicted trend coefficient for each time point in the future time period.

[0009] Optionally, the first regression coefficient expression of the linear regression model is: Among them, β1 represents the first regression coefficient, n represents the number of load data in the target time period, X i Indicates the time point, Y i represents the trend coefficient at a time point, represents the average value at a time point, represents the average value of the trend coefficient.

[0010] Optionally, the second regression coefficient expression of the linear regression model is: Among them, β0 represents the second regression coefficient, β1 represents the first regression coefficient, represents the average value at a time point, represents the average value of the trend coefficient.

[0011] Optionally, the step of calculating the seasonal coefficient of the transformer includes: obtaining the average load value of the transformer in a historical period; dividing the historical period according to seasons to obtain the average load value of the transformer in each season; calculating the ratio between the average load value of the transformer in each season and the average load value in the historical period to obtain the seasonal coefficient of the transformer.

[0012] Optionally, the step of predicting the predicted load data of the transformer in the future time period based on the preprocessed load data, the prediction trend coefficient and the seasonal coefficient includes: calculating the initial predicted load data for each time point based on the prediction trend coefficient and the load data; splitting the initial predicted load data into seasonal load data and non-seasonal load data; adjusting the seasonal load data based on the seasonal coefficient to obtain adjusted seasonal load data; and calculating the final predicted load data based on the adjusted seasonal load data and the non-seasonal load data.

[0013] Optionally, the step of preprocessing the load data includes: filling missing values ​​of the load data based on an interpolation method; detecting abnormal values ​​of the load data, and removing the abnormal values.

[0014] According to another aspect of an embodiment of the present invention, a transformer load prediction device is further provided, including: an acquisition unit, used to acquire load data of the transformer within a target time period, and preprocess the load data to obtain preprocessed load data; a calculation unit, used to calculate the trend coefficient of each time point within the target time period based on the preprocessed load data; a first prediction unit, used to predict the trend coefficient to obtain a predicted trend coefficient for each time point in a future time period; a second prediction unit, used to calculate the seasonal coefficient of the transformer, and predict the predicted load data of the transformer in the future time period based on the preprocessed load data, the predicted trend coefficient and the seasonal coefficient.

[0015] Optionally, the calculation unit includes: a first selection module, used for step one, selecting a time point as a target observation point, and obtaining the load value of the target observation point; a first acquisition module, used for step two, obtaining the predicted load value of the historical observation point corresponding to the target observation point; a first configuration module, used for step three, configuring a smoothing coefficient for the load value of the target observation point; a first calculation module, used for step four, calculating the trend coefficient of the target observation point based on the load value of the target observation point, the predicted load value of the historical observation point and the smoothing coefficient; a first repetition module, used to repeat steps one to four until the trend coefficient of each time point in the target time period is calculated.

[0016] Optionally, the first prediction unit includes: a first construction module, used for step one, to construct a linear regression model based on the trend coefficient and the time point corresponding to the trend coefficient; a second calculation module, used for step two, to calculate the first regression coefficient and the second regression coefficient of the linear regression model; a third calculation module, used for step three, to calculate the predicted trend coefficient of the target time point based on the first regression coefficient, the second regression coefficient and the target time point to be predicted; a second repetition module, used to repeat steps two to three until the residual sum of squares of the linear regression model reaches a minimum value, thereby obtaining the predicted trend coefficient of each time point in the future time period.

[0017] Optionally, the first regression coefficient expression of the linear regression model is: Among them, β1 represents the first regression coefficient, n represents the number of load data in the target time period, X i Indicates the time point, Y i represents the trend coefficient at a time point, represents the average value at a time point, represents the average value of the trend coefficient.

[0018] Optionally, the second regression coefficient expression of the linear regression model is: Among them, β0 represents the second regression coefficient, β1 represents the first regression coefficient, represents the average value at a time point, represents the average value of the trend coefficient.

[0019] Optionally, the second prediction unit includes: a second acquisition module, used to obtain the average load value of the transformer in a historical period; a first division module, used to divide the historical period according to seasons to obtain the average load value of the transformer in each season; and a fourth calculation module, used to calculate the ratio between the average load value of the transformer in each season and the average load value in the historical period to obtain the seasonal coefficient of the transformer.

[0020] Optionally, the second prediction unit also includes: a fifth calculation module, used to calculate the initial predicted load data at each time point based on the prediction trend coefficient and the load data; a first splitting module, used to split the initial predicted load data into seasonal load data and non-seasonal load data; a first adjustment module, used to adjust the seasonal load data based on the seasonal coefficient to obtain adjusted seasonal load data; a sixth calculation module, used to calculate the final predicted load data based on the adjusted seasonal load data and the non-seasonal load data.

[0021] Optionally, the acquisition unit further includes: a first filling module, configured to fill missing values ​​of the load data based on an interpolation method; and a first elimination module, configured to detect abnormal values ​​of the load data and eliminate the abnormal values.

[0022] According to another aspect of an embodiment of the present invention, an electronic device is also provided, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned transformer load prediction methods.

[0023] In this application, the following steps are taken: first, the load data of the transformer within the target time period is collected, and the load data is preprocessed to obtain the preprocessed load data; then, the trend coefficient of each time point within the target time period is calculated based on the preprocessed load data; then, the trend coefficient is predicted to obtain the predicted trend coefficient of each time point in the future time period; finally, the seasonal coefficient of the transformer is calculated; and the predicted load data of the transformer in the future time period is predicted based on the preprocessed load data, the predicted trend coefficient and the seasonal coefficient.

[0024] In this application, for the load data in the historical time period, the trend coefficient of each time point is calculated and the trend coefficient is predicted. At the same time, seasonal changes are taken into account, and the seasonal coefficient is calculated through seasonal load data. The transformer load in the future time period is comprehensively predicted based on the seasonal coefficient, the original load data and the predicted trend coefficient. The trend coefficient can well reflect the changes of the transformer over a long period of time. Combined prediction with the seasonal coefficient can improve the accuracy of load prediction, thereby solving the technical problem in the related technology of low accuracy of prediction results in predicting the transformer load by configuring weight values ​​through exponential smoothing method. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0026] Figure 1 is a flow chart of an optional transformer load prediction method according to an embodiment of the present invention;

[0027] Figure 2 is a schematic diagram of an optional transformer load prediction process according to an embodiment of the present invention;

[0028] Figure 3 is a schematic diagram of an optional transformer load prediction device according to an embodiment of the present invention;

[0029] Figure 4 The figure is a hardware structure block diagram of an electronic device (or mobile device) for a transformer load prediction method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only 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 ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] It should be noted that the transformer load prediction method and device in this application can be used in the field of power grid safety technology when predicting the load value of the transformer, and can also be used in any field other than the field of power grid safety technology when predicting the load value of the transformer. This application does not limit the application field of the transformer load prediction method and device.

[0033] The following embodiments of the present invention can be applied to various transformer load forecasting systems, applications, and devices. The present invention utilizes trend and seasonal information in historical data to forecast transformer loads, achieving high prediction accuracy and stability. It can more accurately capture long-term trends and seasonal variations in load data, thereby improving forecast accuracy. This data-driven approach can better adapt to the variability and uncertainty of load data, flexibly adapt to different types of load data, and is not restricted by data distribution and characteristics. It can process data of different scales, frequencies, and time periods, and is applicable to complex variations in transformer load patterns.

[0034] The present invention will be described in detail below with reference to various embodiments.

[0035] Example 1

[0036] According to an embodiment of the present invention, an embodiment of a transformer load prediction method 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.

[0037] Figure 1 is a flow chart of an optional transformer load prediction method according to an embodiment of the present invention, such as Figure 1 As shown, the method includes the following steps:

[0038] Step S101, collecting load data of the transformer within a target time period, and preprocessing the load data to obtain preprocessed load data;

[0039] Step S102, calculating the trend coefficient of each time point within the target time period based on the preprocessed load data;

[0040] Step S103, predicting the trend coefficient to obtain the predicted trend coefficient at each time point in the future time period;

[0041] Step S104 , calculating the seasonal coefficient of the transformer, and predicting the predicted load data of the transformer in a future time period based on the preprocessed load data, the prediction trend coefficient and the seasonal coefficient.

[0042] Through the above steps, first, the load data of the transformer in the target time period is collected, and the load data is preprocessed to obtain the preprocessed load data. Then, the trend coefficient of each time point in the target time period is calculated based on the preprocessed load data. Then, the trend coefficient is predicted to obtain the predicted trend coefficient of each time point in the future time period. Finally, the seasonal coefficient of the transformer is calculated, and the predicted load data of the transformer in the future time period is predicted based on the preprocessed load data, the predicted trend coefficient and the seasonal coefficient.

[0043] In this embodiment, for the load data in the historical time period, the trend coefficient of each time point is calculated and the trend coefficient is predicted. At the same time, seasonal changes are taken into account, and the seasonal coefficient is calculated through the seasonal load data. The transformer load in the future time period is comprehensively predicted based on the seasonal coefficient, the original load data and the predicted trend coefficient. The trend coefficient can well reflect the changes of the transformer over a long period of time. Combined prediction with the seasonal coefficient can improve the accuracy of the load prediction, thereby solving the technical problem in the related technology of low accuracy of the prediction result of configuring weight values ​​by exponential smoothing method to predict the transformer load.

[0044] The embodiment of the present invention is described in detail below in conjunction with the above steps.

[0045] It should be noted that transformer load data refers to the current and power loads borne by the transformer during operation. This data is crucial for evaluating transformer performance. Load data can include active power and reactive power, which are used to reflect the load variation patterns of the transformer in different time periods, helping to obtain the operating status and efficiency of the transformer. Transformer load data can also be used to evaluate transformer characteristics. For example, the resistive and inductive parameters obtained through open-circuit and short-circuit tests can be used to evaluate the transformer's efficiency, temperature rise, lifespan, and parallel operation capability. In short, transformer load data is an important means of evaluating transformer characteristics. By analyzing and applying load data, the operating efficiency and reliability of the transformer can be improved, ensuring the stable operation of the power system and enhancing the stability of the power system.

[0046] During the planning, operation and dispatching of the power system, transformer load forecasting can predict the status and characteristic changes of the transformer in advance, and then assist power workers in rationally allocating resources, thereby improving the efficiency of power grid operation and reducing operating costs.

[0047] Step S101 : collecting load data of the transformer within a target time period, and preprocessing the load data to obtain preprocessed load data.

[0048] It should be noted that transformers, as electrical devices, are used to increase or decrease the values ​​of AC voltage and current based on the principle of electromagnetic induction, while also transmitting electrical energy. They are indispensable components for power transmission and distribution. Load data can include: load current, active power, reactive power, apparent power, and load factor. Load current is the value of the current in each phase line during transformer operation. Active power refers to the actual electrical energy consumed by the transformer. Reactive power is related to active power but does not actually consume electrical energy. Apparent power represents the total power capacity of the transformer, combining active and reactive power. The load factor represents the ratio of the actual load of the transformer to the rated capacity and can be used to evaluate the operating efficiency of the transformer.

[0049] It should be noted that after collecting the load data of the transformer, the load data needs to be preprocessed first to process the missing values ​​and abnormal values ​​in the load data to ensure the integrity and accuracy of the load data.

[0050] Optionally, the step of preprocessing the load data includes: filling missing values ​​of the load data based on an interpolation method; detecting abnormal values ​​of the load data, and removing the abnormal values.

[0051] It should be noted that, in the embodiment of the present invention, interpolation method can be used to fill in missing values ​​of load data to ensure data continuity and integrity. The interpolation formula is expressed as: Among them, Y t is a missing value, Y t-1 and Y t+1 are the observations closest to the missing value.

[0052] It should be noted that, in the embodiment of the present invention, a box plot can be used to detect outliers to select and remove or correct outliers in the load data. The box plot formula is expressed as:

[0053] IQR=Q3-Q1

[0054] Lower_Bound=Q1-1.5×IQR

[0055] Upper_Bound=Q3-1.5×IQR

[0056] Where IQR is the interquartile range, Q1 is the first quartile, Q3 is the third quartile, Lower_Bound is the lower limit, Upper_Bound is the upper limit, and outliers are defined as observations below the lower limit or above the upper limit.

[0057] Step S102 : calculating the trend coefficient of each time point within the target time period based on the pre-processed load data.

[0058] It should be noted that the embodiment of the present invention predicts the transformer load value by calculating the trend coefficient at each time point within the target time period. The trend coefficient can reflect the load change of the transformer.

[0059] Optionally, the step of calculating the trend coefficient of each time point within the target time period based on the preprocessed load data includes: step one, selecting a time point as the target observation point and obtaining the load value of the target observation point; step two, obtaining the predicted load value of the historical observation point corresponding to the target observation point; step three, configuring a smoothing coefficient for the load value of the target observation point; step four, calculating the trend coefficient of the target observation point based on the load value of the target observation point, the predicted load value of the historical observation point and the smoothing coefficient; repeat steps one to four until the trend coefficient of each time point within the target time period is calculated.

[0060] Specifically, when calculating the trend coefficient, first select a time point as the target observation point, then obtain the load value of the target observation point, and at the same time obtain the predicted load value of the previous time point corresponding to the target observation point. According to the pre-set smoothing coefficient (which can be a value between 0-1), the trend coefficient of the target observation point is comprehensively calculated, and the trend coefficient is used to quantify the change trend of the transformer at the target observation point. Repeat the above steps until the trend coefficients of all time points in the target time period are calculated, and the trend coefficient of each time point in the target time period is obtained.

[0061] Specifically, the calculation formula of the trend coefficient is expressed as: T t =α·Y t +(1-α)·T t-1 Among them, T t is the trend coefficient of the transformer at the target observation point t, Y t represents the load data of the transformer at the target observation point, α is the smoothing coefficient, which usually takes a value between 0 and 1, indicating the weight of the past observation values, T t-1 Indicates the trend coefficient of the previous observation point of the target observation point.

[0062] Step S103: predict the trend coefficient to obtain the predicted trend coefficient at each time point in the future time period.

[0063] It should be noted that the embodiment of the present invention predicts the trend coefficient and can obtain the predicted trend coefficient at each time point in the target time period. Then, based on the load data at each time point in the target time period, the predicted load value at each time point in the future period is obtained through comprehensive calculation, thereby realizing the prediction of the transformer load value. The predicted trend coefficient is not restricted by the data distribution and characteristics, and can flexibly adapt to different types of load data, thereby improving the scope of application of the prediction and the accuracy of the prediction results.

[0064] Optionally, the step of predicting the trend coefficient to obtain the predicted trend coefficient for each time point in the future time period includes: step one, constructing a linear regression model based on the trend coefficient and the time point corresponding to the trend coefficient; step two, calculating the first regression coefficient and the second regression coefficient of the linear regression model; step three, calculating the predicted trend coefficient of the target time point based on the first regression coefficient and the second regression coefficient and the target time point to be predicted; repeating steps two to three until the residual sum of squares of the linear regression model reaches a minimum value, thereby obtaining the predicted trend coefficient for each time point in the future time period.

[0065] It should be noted that when predicting the trend coefficient at each time point, the trend coefficient is first used as the dependent variable, and the time point corresponding to the trend coefficient is used as the independent variable to construct a linear regression model. The linear regression model expression contains the independent variable, the dependent variable, the first regression coefficient and the second regression coefficient. Then, the values ​​of the first regression coefficient and the second regression coefficient are calculated according to the calculation formula of the regression coefficient. The values ​​of the first regression coefficient and the second regression coefficient are substituted into the linear regression model to calculate the predicted trend coefficient corresponding to each time point. The above steps are repeated for iterative calculation until the residual sum of squares of the model is minimized, and the final predicted trend coefficient of each time point is obtained.

[0066] Specifically, in the embodiment of the present invention, the trend coefficient can be used as the dependent variable and the time point as the independent variable, and the trend coefficient of each time point can be predicted by fitting a linear regression model. The linear regression model formula is:

[0067] Y=β0+β1X+∈

[0068] Where Y is the dependent variable (trend coefficient), X is the independent variable (time point), β0 (first regression coefficient) and β1 (second regression coefficient) are regression coefficients, and ∈ is the error term;

[0069] Optionally, the first regression coefficient expression of the linear regression model is Among them, β1 represents the first regression coefficient, n represents the number of load data in the target time period, X i Indicates the time point, Y i represents the trend coefficient at a time point, represents the average value at a time point, represents the average value of the trend coefficient.

[0070] Optionally, the second regression coefficient expression of the linear regression model is: Among them, β0 represents the second regression coefficient, β1 represents the first regression coefficient, represents the average value at a time point, represents the average value of the trend coefficient.

[0071] When predicting the trend coefficient at each time point based on the linear regression model, the following formula is used to calculate the values ​​of the regression coefficients β0 and β1 to minimize the residual sum of squares of the model and obtain the final predicted trend coefficient at each time point. The calculation formulas for the second regression coefficient and the first regression coefficient are respectively expressed as:

[0072]

[0073] in, and are the means of the independent and dependent variables, respectively, and n is the sample size.

[0074] Step S104 , calculating the seasonal coefficient of the transformer, and predicting the predicted load data of the transformer in a future time period based on the preprocessed load data, the prediction trend coefficient and the seasonal coefficient.

[0075] After predicting the trend coefficient at each time point, the embodiment of the present invention uses seasonal factors as one of the indicators for load value prediction. By calculating the seasonal coefficient and capturing the seasonal factors in the historical load data, the accuracy of the calculation results can be improved. The nodularity coefficient, the load data at each time point, and the predicted trend coefficient are used as load value prediction indicators to comprehensively calculate the predicted load data in the future time period.

[0076] Optionally, the step of calculating the seasonal coefficient of the transformer includes: obtaining the average load value of the transformer in a historical period; dividing the historical period according to seasons to obtain the average load value of the transformer in each season; calculating the ratio between the average load value of the transformer in each season and the average load value in the historical period to obtain the seasonal coefficient of the transformer.

[0077] It should be noted that when calculating the seasonal coefficient, the average load value of each season in an observation period (corresponding to the above-mentioned historical period) is first calculated, and the observation period is divided into multiple seasons with one season as the unit, and the average load value in each season is calculated. The ratio between the average load value in each season and the average load value in the historical period is calculated to obtain the seasonal coefficient corresponding to each season.

[0078] Optionally, the step of predicting the predicted load data of the transformer in a future time period based on the preprocessed load data, the prediction trend coefficient and the seasonal coefficient includes: calculating the initial predicted load data for each time point based on the prediction trend coefficient and the load data; splitting the initial predicted load data into seasonal load data and non-seasonal load data; adjusting the seasonal load data based on the seasonal coefficient to obtain adjusted seasonal load data; and calculating the final predicted load data based on the adjusted seasonal load data and non-seasonal load data.

[0079] It should be noted that when comprehensively predicting the load value of the transformer in the short future time based on the pre-processed load data, prediction trend coefficient and seasonal coefficient, the prediction trend coefficient and load data at each time point are first accumulated to obtain the initial prediction load data at each time point, and then the initial prediction load data is adjusted according to the seasonal coefficient. Specifically, the calculated initial prediction overload data is split into seasonal load data and non-seasonal load data, and then the seasonal coefficient corresponding to the target time period is obtained, and the product of the seasonal coefficient and the seasonal load data is calculated to obtain the adjusted seasonal load data. Finally, the adjusted seasonal load data and the non-seasonal load data are accumulated to obtain the final prediction load data, thereby completing the prediction of the transformer load data.

[0080] The following describes in detail another optional specific implementation.

[0081] Figure 2 is a schematic diagram of an optional transformer load prediction process according to an embodiment of the present invention, such as Figure 2 As shown in Figure 2, the transformer load prediction process includes:

[0082] Step 1: Collect historical transformer load data, including hourly or daily load values, perform missing value processing and outlier detection on the historical load data, fill in missing values ​​and remove outliers;

[0083] Collect historical transformer load data, including hourly or daily loads, which can be obtained through transformer monitoring systems, sensors, or historical records.

[0084] Interpolation is used to fill missing values ​​to ensure data continuity and integrity. In this embodiment, linear interpolation is used, and the linear interpolation formula is as follows: Among them, Y t is a missing value, Y t-1 and Y t+1 are the observations closest to the missing value.

[0085] Outliers are detected using statistical methods (such as box plots or Z-score) or machine learning methods (such as isolation forests or clustering-based methods). Detected outliers can be removed or corrected. In this embodiment, the box plot is used to remove outliers as an example. The box plot formula is expressed as:

[0086] IQR=Q3-Q1

[0087] Lower_Bound=Q1-1.5×IQR

[0088] Upper_Bound=Q3-1.5×IQR

[0089] Where IQR is the interquartile range, Q1 is the first quartile, Q3 is the third quartile, Lower_Bound is the lower limit, Upper_Bound is the upper limit, and outliers are defined as observations below the lower limit or above the upper limit.

[0090] The missing values ​​are filled through the above steps to ensure the continuity and integrity of the data, and then outliers are detected and removed to reduce interference with subsequent model building.

[0091] Step 2: Estimate the trend of the original load data and calculate the trend coefficient at each time point;

[0092] When calculating the trend coefficient at each time point, first select an initial value as the starting point (corresponding to the target observation point mentioned above), then calculate a new estimate based on the load data at the starting point and the estimate at the previous time point, and use the following formula to calculate the trend estimate at each time point:

[0093] T t =α·Y t +(1-α)·T t-1

[0094] Among them, T t is the trend estimate at time t, Y t is the original load value at time t, α is the smoothing coefficient, which usually takes a value between 0 and 1, indicating the weight of past observations. The trend estimate obtained is the trend coefficient, which indicates the trend at each time point.

[0095] Step 3: predict the trend coefficient, identify the long-term trend, and obtain the predicted trend coefficient at each time point;

[0096] The embodiment of the present invention uses a linear regression model to predict the trend coefficient to identify long-term trends. In this embodiment, the trend coefficient is regarded as the dependent variable and time is regarded as the independent variable. The linear regression model is fitted to predict the future trend coefficient. The linear regression model formula is expressed as:

[0097] Y=β0+β1X+∈

[0098] Where Y is the dependent variable (trend coefficient), X is the independent variable (time), β0 and β1 are the first and second regression coefficients respectively, and ∈ is the error term.

[0099] Use the following formula to estimate the values ​​of the regression coefficients β0 and β1, so as to minimize the residual sum of squares of the model, and continuously fit to obtain the final predicted trend coefficient at each time point:

[0100]

[0101] in, and are the means of the independent and dependent variables, respectively, and n is the sample size.

[0102] Step 4: Add the forecast trend coefficient to the original load data to obtain the adjusted load data;

[0103] The predicted trend coefficient obtained in step 3 is matched with the time point, and then the predicted trend coefficient is added to the original load data to obtain the adjusted load data. The calculation formula is expressed as:

[0104]

[0105] in, is the adjusted load data, i.e. the initial load forecast value of the transformer at time t; Y t is the original load value at time t; T t is the predicted trend coefficient at time t.

[0106] Through the above steps, the predicted trend is added to the original load data to obtain more accurate adjusted load data. This process is logically related to the trend prediction step and provides the basis for the final load forecast.

[0107] Step 5: seasonally decompose the adjusted load data to capture seasonal variations and obtain seasonal data (corresponding to the above seasonal load data);

[0108] In this embodiment, the adjusted load data is decomposed into trend, seasonality, and residual to capture seasonal changes. The seasonal decomposition formula is expressed as:

[0109]

[0110] in, is the adjusted initial forecast load data at time t; T t is the trend part; S t is the seasonal part; R t is the residual part.

[0111] Specifically, the trend part T is estimated by moving average or other smoothing methods. t Then, by calculating the difference between the original load data and the trend part, the remaining part, namely Y′, is obtained. t -T t Finally, seasonal analysis is performed on the remaining part to obtain the seasonal part S t (corresponding to the seasonal data above).

[0112] Through this step, the initial forecast load data of the transformer can be seasonally decomposed using an additive model to capture seasonal variations.

[0113] Step 6: Calculate the seasonal coefficient and adjust the seasonal data based on the seasonal coefficient to obtain the final predicted load data.

[0114] Select an observation period, for example, a whole year as an observation period, calculate the average load value of the transformer during the observation period, divide the observation period according to season, obtain the average load value of each season, calculate the ratio of the average load value of each season to the average load value of the observation period, and obtain the seasonal coefficient of each season.

[0115] After obtaining the seasonal coefficient, the seasonal data is adjusted using the seasonal coefficient to ensure the accuracy of seasonal changes. The expression for seasonal adjustment is:

[0116]

[0117] Among them, S′ t is the seasonally adjusted data, S t is the raw seasonal data at time t, I t is the seasonal coefficient at time t, is the average of all seasonal indices.

[0118] Finally, the adjusted seasonal data is added to the trend part and the residual part obtained by splitting to calculate the final forecast load data.

[0119] The embodiment of the present invention uses trend and seasonal information in historical data to predict transformer load, has high prediction accuracy and stability, and can more accurately capture long-term trends and seasonal changes in load data, thereby improving the accuracy of the prediction. This data-driven method can better adapt to the changes and uncertainties of load data, flexibly adapt to different types of load data, is not restricted by data distribution and characteristics, can process data of different scales, different frequencies and different time periods, and can be applied to complex change patterns of transformer load.

[0120] The following describes it in detail with reference to another embodiment.

[0121] Example 2

[0122] A transformer load prediction device provided in this embodiment includes multiple implementation units, each implementation unit corresponds to each implementation step in the above-mentioned embodiment 1. Its specific implementation method and beneficial effects can refer to the above-mentioned method embodiment and will not be repeated here.

[0123] Figure 3 is a schematic diagram of an optional transformer load prediction device according to an embodiment of the present invention, such as Figure 3 As shown, the transformer load prediction device may include: a collection unit 31, a calculation unit 32, a first prediction unit 33, and a second prediction unit 34, wherein:

[0124] The collection unit 31 is used to collect the load data of the transformer within the target time period and preprocess the load data to obtain the preprocessed load data;

[0125] A calculation unit 32, configured to calculate a trend coefficient at each time point within a target time period based on the preprocessed load data;

[0126] The first prediction unit 33 is used to predict the trend coefficient to obtain the predicted trend coefficient at each time point in the future time period;

[0127] The second prediction unit 34 is configured to calculate the seasonal coefficient of the transformer and predict the predicted load data of the transformer in a future time period based on the preprocessed load data, the prediction trend coefficient and the seasonal coefficient.

[0128] The above-mentioned transformer load prediction device collects the load data of the transformer within the target time period through the collection unit 31, and preprocesses the load data to obtain the preprocessed load data; calculates the trend coefficient of each time point in the target time period based on the preprocessed load data through the calculation unit 32; predicts the trend coefficient through the first prediction unit 33 to obtain the predicted trend coefficient for each time point in the future time period; calculates the seasonal coefficient of the transformer through the second prediction unit 34, and predicts the predicted load data of the transformer in the future time period based on the preprocessed load data, the predicted trend coefficient and the seasonal coefficient.

[0129] In this embodiment, for the load data in the historical time period, the trend coefficient of each time point is calculated and the trend coefficient is predicted. At the same time, seasonal changes are taken into account, and the seasonal coefficient is calculated through the seasonal load data. The transformer load in the future time period is comprehensively predicted based on the seasonal coefficient, the original load data and the predicted trend coefficient. The trend coefficient can well reflect the changes of the transformer over a long period of time. Combined prediction with the seasonal coefficient can improve the accuracy of the load prediction, thereby solving the technical problem in the related technology of low accuracy of the prediction result of configuring weight values ​​by exponential smoothing method to predict the transformer load.

[0130] Optionally, the calculation unit 32 includes: a first selection module, used for step one, selecting a time point as a target observation point and obtaining the load value of the target observation point; a first acquisition module, used for step two, obtaining the predicted load value of the historical observation point corresponding to the target observation point; a first configuration module, used for step three, configuring a smoothing coefficient for the load value of the target observation point; a first calculation module, used for step four, calculating the trend coefficient of the target observation point based on the load value of the target observation point, the predicted load value of the historical observation point and the smoothing coefficient; a first repetition module, used to repeat steps one to four until the trend coefficient of each time point in the target time period is calculated.

[0131] Optionally, the first prediction unit 33 includes: a first construction module, used for step one, to construct a linear regression model based on the trend coefficient and the time point corresponding to the trend coefficient; a second calculation module, used for step two, to calculate the first regression coefficient and the second regression coefficient of the linear regression model; a third calculation module, used for step three, to calculate the predicted trend coefficient of the target time point based on the first regression coefficient and the second regression coefficient and the target time point to be predicted; a second repetition module, used to repeat steps two to three until the residual sum of squares of the linear regression model reaches a minimum value, thereby obtaining the predicted trend coefficient of each time point in the future time period.

[0132] Optionally, the first regression coefficient expression of the linear regression model is: Among them, β1 represents the first regression coefficient, n represents the number of load data in the target time period, X i Indicates the time point, Y i represents the trend coefficient at a time point, represents the average value at a time point, represents the average value of the trend coefficient.

[0133] Optionally, the second regression coefficient expression of the linear regression model is: Among them, β0 represents the second regression coefficient, β1 represents the first regression coefficient, represents the average value at a time point, represents the average value of the trend coefficient.

[0134] Optionally, the second prediction unit 34 includes: a second acquisition module, used to obtain the average load value of the transformer in the historical period; a first division module, used to divide the historical period according to seasons to obtain the average load value of the transformer in each season; a fourth calculation module, used to calculate the ratio between the average load value of the transformer in each season and the average load value in the historical period to obtain the seasonal coefficient of the transformer.

[0135] Optionally, the second prediction unit 34 also includes: a fifth calculation module, used to calculate the initial predicted load data at each time point based on the prediction trend coefficient and the load data; a first splitting module, used to split the initial predicted load data into seasonal load data and non-seasonal load data; a first adjustment module, used to adjust the seasonal load data based on the seasonal coefficient to obtain adjusted seasonal load data; a sixth calculation module, used to calculate the final predicted load data based on the adjusted seasonal load data and non-seasonal load data.

[0136] Optionally, the acquisition unit 31 further includes: a first filling module, configured to fill missing values ​​of the load data based on an interpolation method; and a first eliminating module, configured to detect abnormal values ​​of the load data and eliminate the abnormal values.

[0137] The above-mentioned transformer load prediction device may also include a processor and a memory. The above-mentioned acquisition unit 31, calculation unit 32, first prediction unit 33, second prediction unit 34, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0138] The processor includes a core, which retrieves the corresponding program unit from the memory. One or more cores can be provided, and the load value of the transformer can be predicted by adjusting the core parameters.

[0139] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0140] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any of the above-mentioned transformer load prediction methods.

[0141] According to another aspect of an embodiment of the present invention, an electronic device is also provided, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement any one of the above-mentioned transformer load prediction methods.

[0142] According to another aspect of an embodiment of the present invention, a computer program product is further provided. The computer program product includes a computer program, wherein when the computer program is executed by a processor, any one of the above transformer load prediction methods is implemented.

[0143] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: collecting load data of the transformer within a target time period, and preprocessing the load data to obtain preprocessed load data; calculating the trend coefficient at each time point within the target time period based on the preprocessed load data; predicting the trend coefficient to obtain a predicted trend coefficient at each time point in the future time period; calculating the seasonal coefficient of the transformer, and predicting the predicted load data of the transformer in the future time period based on the preprocessed load data, the predicted trend coefficient and the seasonal coefficient.

[0144] Optionally, the step of calculating the trend coefficient of each time point within the target time period based on the preprocessed load data includes: step one, selecting a time point as the target observation point and obtaining the load value of the target observation point; step two, obtaining the predicted load value of the historical observation point corresponding to the target observation point; step three, configuring a smoothing coefficient for the load value of the target observation point; step four, calculating the trend coefficient of the target observation point based on the load value of the target observation point, the predicted load value of the historical observation point and the smoothing coefficient; repeat steps one to four until the trend coefficient of each time point within the target time period is calculated.

[0145] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: predicting the trend coefficient, and the step of obtaining the predicted trend coefficient for each time point in the future time period includes: step one, constructing a linear regression model based on the trend coefficient and the time point corresponding to the trend coefficient; step two, calculating the first regression coefficient and the second regression coefficient of the linear regression model; step three, calculating the predicted trend coefficient of the target time point based on the first regression coefficient and the second regression coefficient and the target time point to be predicted; repeating steps two to three until the residual sum of squares of the linear regression model reaches a minimum value, thereby obtaining the predicted trend coefficient for each time point in the future time period.

[0146] The present application also provides a computer program product, which, when executed on a data processing device, is adapted to execute a program for initializing the following method steps: the first regression coefficient of the linear regression model is expressed as: Among them, β1 represents the first regression coefficient, n represents the number of load data in the target time period, X i Indicates the time point, Y i represents the trend coefficient at a time point, represents the average value at a time point, represents the average value of the trend coefficient.

[0147] The present application also provides a computer program product, which, when executed on a data processing device, is adapted to execute a program for initializing the following method steps: the second regression coefficient of the linear regression model is expressed as: Among them, β0 represents the second regression coefficient, β1 represents the first regression coefficient, represents the average value at a time point, represents the average value of the trend coefficient.

[0148] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: the step of calculating the seasonal coefficient of the transformer includes: obtaining the average load value of the transformer in a historical period; dividing the historical period according to seasons to obtain the average load value of the transformer in each season; calculating the ratio between the average load value of the transformer in each season and the average load value in the historical period to obtain the seasonal coefficient of the transformer.

[0149] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: the steps of predicting the predicted load data of the transformer in a future time period based on preprocessed load data, a predicted trend coefficient and a seasonal coefficient include: calculating the initial predicted load data for each time point based on the predicted trend coefficient and the load data; splitting the initial predicted load data into seasonal load data and non-seasonal load data; adjusting the seasonal load data based on the seasonal coefficient to obtain adjusted seasonal load data; and calculating the final predicted load data based on the adjusted seasonal load data and non-seasonal load data.

[0150] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: the step of preprocessing the load data includes: filling the missing values ​​of the load data based on the interpolation method; detecting abnormal values ​​of the load data and eliminating the abnormal values.

[0151] Figure 4 FIG. 1 is a hardware structure block diagram of an electronic device (or mobile device) for a transformer load prediction method according to an embodiment of the present invention. Figure 4 As shown, the electronic device may include one or more ( Figure 4(As shown in the figure, 402a, 402b, ..., 402n are used) processor 402 (processor 402 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), memory 404 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a keyboard, a power supply and / or a camera. It will be understood by those skilled in the art that Figure 4 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 4 More or fewer components than shown, or with Figure 4 Different configurations shown.

[0152] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0153] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0154] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0155] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0156] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0157] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc. Various media that can store program codes.

[0158] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A transformer load prediction method, characterized in that: include: Collecting load data of the transformer within a target time period, and preprocessing the load data to obtain preprocessed load data; Calculating a trend coefficient at each time point within a target time period based on the preprocessed load data; Predicting the trend coefficient to obtain a predicted trend coefficient at each time point in a future time period; Calculate the seasonal coefficient of the transformer, and predict the predicted load data of the transformer in the future time period based on the preprocessed load data, the predicted trend coefficient and the seasonal coefficient, including: calculating the initial predicted load data at each time point based on the predicted trend coefficient and the load data; splitting the initial predicted load data into seasonal load data and non-seasonal load data; adjusting the seasonal load data based on the seasonal coefficient to obtain adjusted seasonal load data; and calculating the final predicted load data based on the adjusted seasonal load data and the non-seasonal load data.

2. The prediction method according to claim 1, characterized in that The step of calculating the trend coefficient at each time point within the target time period based on the pre-processed load data includes: Step 1: Select a time point as a target observation point and obtain the load value of the target observation point; Step 2: Obtain the predicted load value of the historical observation point corresponding to the target observation point; Step 3: configuring a smoothing coefficient for the load value of the target observation point; Step 4: calculating the trend coefficient of the target observation point based on the load value of the target observation point, the predicted load value of the historical observation point, and the smoothing coefficient; Repeat steps 1 to 4 until the trend coefficient of each time point in the target time period is calculated.

3. The prediction method according to claim 1, wherein: The step of predicting the trend coefficient to obtain the predicted trend coefficient at each time point in the future time period includes: Step 1: constructing a linear regression model based on the trend coefficient and the time point corresponding to the trend coefficient; Step 2: Calculate the first regression coefficient and the second regression coefficient of the linear regression model; Step 3: calculating a prediction trend coefficient of the target time point based on the first regression coefficient, the second regression coefficient, and the target time point to be predicted; Repeat steps 2 to 3 until the residual sum of squares of the linear regression model reaches a minimum value, and obtain the predicted trend coefficient for each time point in the future time period.

4. The prediction method according to claim 3, characterized in that The first regression coefficient expression of the linear regression model is: Among them, β1 represents the first regression coefficient, n represents the number of load data in the target time period, X i Indicates the time point, Y i represents the trend coefficient at a time point, represents the average value at a time point, represents the average value of the trend coefficient.

5. The prediction method according to claim 4, characterized in that The second regression coefficient expression of the linear regression model is: Among them, β0 represents the second regression coefficient, β1 represents the first regression coefficient, represents the average value at a time point, represents the average value of the trend coefficient.

6. The prediction method according to claim 1, characterized in that The step of calculating the seasonal coefficient of the transformer comprises: Obtaining an average load value of the transformer in a historical period; Dividing the historical period by season to obtain an average load value of the transformer in each season; The ratio between the average load value of the transformer in each season and the average load value in the historical period is calculated to obtain the seasonal coefficient of the transformer.

7. The prediction method according to claim 1, wherein: The step of preprocessing the load data includes: Filling missing values ​​of the load data based on an interpolation method; Detect abnormal values ​​of the load data and remove the abnormal values.

8. A transformer load prediction device, characterized in that: include: The collecting unit is used to collect the load data of the transformer within the target time period and pre-process the load data to obtain the pre-processed load data; a calculation unit, configured to calculate a trend coefficient at each time point within a target time period based on the preprocessed load data; A first prediction unit is used to predict the trend coefficient to obtain a predicted trend coefficient at each time point in a future time period; a second prediction unit, configured to calculate a seasonal coefficient of the transformer, and predict predicted load data of the transformer in the future time period based on the preprocessed load data, the prediction trend coefficient, and the seasonal coefficient; The second prediction unit includes: a fifth calculation module, used to calculate the initial predicted load data at each time point based on the prediction trend coefficient and the load data; a first splitting module, used to split the initial predicted load data into seasonal load data and non-seasonal load data; a first adjustment module, used to adjust the seasonal load data based on the seasonal coefficient to obtain adjusted seasonal load data; a sixth calculation module, used to calculate the final predicted load data based on the adjusted seasonal load data and the non-seasonal load data.

9. An electronic device, characterized in that: The method comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the transformer load prediction method according to any one of claims 1 to 7.

Images