Data-based methods for line loss prediction and order dispatch

Abstract

This invention relates to the field of line loss prediction technology, specifically disclosing a data-based method for line loss prediction and dispatch. The method includes collecting line loss data for day T-2, where T is the current day; calculating the line loss rate for day T-2 based on the data; calculating the predicted line loss rates for days T-1 and T using a grey relational analysis-modified least squares method and a grey prediction GM(1,1) model; comparing the severity of different line loss situations; and dispatching the line loss situation for day T-2 and the predicted line loss situations for days T-1 and T to management personnel in descending order of severity. This invention uses an RPA program to automatically collect line loss data instead of manually. The prediction algorithm is easy to implement, occupies little memory, and provides accurate and reliable prediction results. It also enables intelligent dispatch, allowing management personnel to intuitively understand the line loss situation and take timely and appropriate measures for management.

Description

Technical Field

[0001] This invention relates to the field of line loss prediction technology, and in particular to a data-based method for line loss prediction and order dispatch. Background Technology

[0002] Line loss is one of the core economic and technical indicators for power grid companies. The analysis, prediction, and management of line loss are crucial for reducing carbon emissions and tapping the potential for energy conservation and consumption reduction in the power grid. Currently, line loss management in the power industry has multifaceted significance. For distribution, comparing actual line loss with advanced international and domestic indicators helps identify the main causes of line loss, enabling the development of relevant measures to reduce line loss and improve energy efficiency. For marketing, its significance lies in effectively reducing power line losses and improving power supply quality through scientific planning and comprehensive management, thus providing a stable and reliable power supply for social and economic development and people's lives.

[0003] In power grid companies, line losses are generally divided into two different types: distribution transformer area line losses and distribution line losses. Low-voltage distribution transformer areas are crucial links connecting users and transmission in the power grid system, located at the end of the power supply system. Due to the complex and diverse topology of low-voltage distribution networks, their wide service areas, and the difficulty of management, distribution transformer areas account for a significant portion of power losses. Distribution transformer area line losses mainly refer to the active power losses that occur during the transmission and distribution of electrical energy in the distribution network due to the impedance of the distribution lines and equipment. These losses are typically affected by many factors, including the distribution network topology, load conditions, and weather conditions. The distribution transformer area line loss rate is an important indicator that measures the efficiency and economy of power system operation and is a crucial aspect of energy conservation, loss reduction, and efficiency improvement.

[0004] Distribution line losses refer to the loss of electrical energy during transmission within the entire power distribution network due to factors such as resistive heat loss, induced electromagnetic wave loss, and capacitive current loss caused by current flowing through conductors. Similar to transformer substation losses, distribution line losses are also related to factors such as line length, material, cross-sectional area, and the magnitude of the current flowing through the line. Both types of line losses can be estimated and controlled through certain calculation methods and predictive models, thereby reducing energy loss and improving the operating efficiency and economy of the power system.

[0005] Currently, power grid companies primarily rely on an integrated power consumption and line loss management system to collect data on transformer substation line losses. This system can manually generate six detailed tables representing three indicators—negative loss, high loss, and line loss non-compliance rate—for the line loss of transformer substations and distribution lines on day T-2. However, this system has limitations in line loss analysis, prediction, and early warning, and it cannot promptly inform relevant transformer substation or line management personnel to address line loss issues. Therefore, achieving automated collection, analysis, prediction, and intelligent dispatch of line loss data is crucial for strengthening line loss management, reducing the burden on grassroots staff, and improving operational efficiency. Summary of the Invention

[0006] The purpose of this invention is to provide a data-based method for line loss prediction and dispatch, in order to solve the limitations of existing technologies in line loss analysis, prediction and early warning, as well as the inability to dispatch orders in a timely manner.

[0007] To address the aforementioned technical problems, this invention provides a data-based method for line loss prediction and order dispatch, characterized by the following steps:

[0008] S1. Collect line loss data for day T-2, where T is the current day. The line loss data includes power supply, electricity sales, independent variables affecting line loss, and line loss impact indicators. Calculate the line loss rate for day T-2 based on power supply and electricity sales to obtain the line loss situation for day T-2.

[0009] S2. Based on the independent variables and indicators affecting line loss, the predicted line loss rate for days T-1 and T is calculated using the least squares method improved by grey relational analysis and the grey prediction GM(1,1) model, thus obtaining the predicted line loss situation for days T-1 and T.

[0010] S3. Compare the severity of line loss and assign the line loss situation on day T-2 and the predicted line loss situation on day T-1 and day T to the management personnel in order from most severe to least severe.

[0011] Preferably, the Cyclone RPA designer is used to design the RPA program for the integrated power and line loss management system. The RPA program is used to automatically collect and process line loss data, which is stored in the background server in the form of an Excel spreadsheet.

[0012] Preferably, the formula for calculating the line loss rate is: Line loss rate = (line loss power / power supply) × 100% = (power supply - power sales) / power supply × 100% = (1 - power sales / power supply) × 100%.

[0013] Preferably, the independent variables affecting line loss include voltage, current, power factor, equipment load rate, and grid operation mode, and the indicators affecting line loss include enterprise economic benefits, power quality, equipment operating efficiency, and power supply reliability.

[0014] Preferably, the collected line loss data is parsed. Parsing the line loss data includes data preprocessing and data analysis. Data preprocessing involves removing null values ​​and outliers, and data analysis involves calculating the mean and standard deviation.

[0015] Preferably, the displayData function of the chart.js library is used to convert the parsed line loss data into a chart and display it on the web interface of the integrated power consumption and line loss management system.

[0016] Preferably, line loss prediction is performed using the least squares method improved by grey relational analysis:

[0017] Let n sequences of independent variables affecting line loss form a matrix:

[0018]

[0019] Where m is the number of indicators affecting line loss, we get:

[0020] X i ′=(x i ′(1),x i ′(2),...,x i ′(m)) T (i = 1, 2, ..., n);

[0021] The dimensionless sequence forms a matrix:

[0022]

[0023] Calculate the absolute difference between the corresponding elements of each evaluated object's index sequence and the reference sequence: x0(k) - x i (k)(k=1,...,m,i=1,...,n);

[0024] Calculate the correlation coefficient:

[0025]

[0026] Where ρ is the resolution coefficient, taking values ​​in (0,1), the association order is calculated as follows:

[0027]

[0028] The least squares linear regression model was used for training to obtain the predicted value of the line loss rate, k1.

[0029] Line loss prediction using the grey prediction GM(1,1) model:

[0030] Suppose that n independent variables affecting line loss form a sequence:

[0031] x (0)=[x (0) (1),x (0) (2),...,x (0) (n)];

[0032] Perform an accumulation calculation, denoted as 1 - AGO:

[0033]

[0034] The generated sequence is obtained as follows:

[0035] x (1) =[x (1) (1),x (1) (2),...,x (1) (n)];

[0036] The mean sequence is obtained:

[0037] z(w) = 0.5x(w) + 0.5x(w-1);

[0038] To solve the grey system problem, the first-order differential equation of GM(1,1) is:

[0039]

[0040] List the Y and B matrices:

[0041]

[0042] The parameter vector to be identified is Φ = [au] T And Y = BΦ, the parameter vector obtained by the least squares method is:

[0043]

[0044] The discrete solution is obtained:

[0045]

[0046] The predicted line loss rate k2 is obtained:

[0047]

[0048] The final weighted line loss rate prediction value K is obtained:

[0049] K = g1k1 + g2k2;

[0050] Where g1 and g2 are weights, and g1 + g2 = 1.

[0051] Preferably, g1 = g2 = 0.5.

[0052] Preferably, the line loss situation is obtained by calculating the line loss rate as follows: a line loss rate > 6% is considered high loss, a line loss rate < 0% is considered negative loss, and both high loss and negative loss are considered substandard.

[0053] Preferably, line loss situations are divided into four types: A. There is a failure to meet the standard on day T-2; B. It is predicted that there will be a failure to meet the standard on day T-1 or day T; C. There is a line loss rate exceeding the warning line of 90% on day T-2; D. It is predicted that there will be a line loss rate exceeding the warning line of 90% on day T-1 or day T. The warning line exceeding 90% is defined as a line loss rate > 5.4% or a line loss rate < 0.9%.

[0054] For the four scenarios mentioned above, dispatch weight values ​​of 0.50, 0.25, 0.20, and 0.05 are set respectively. Scenario A, which exists, is set as category A, and scenario B, which does not exist, is set as category B. Category A scenarios have higher dispatch priority than category B scenarios. The dispatch order is as follows: Category A scenarios with higher dispatch priority are notified to the management personnel first, followed by Category B scenarios with lower dispatch priority. When dispatch priorities are the same, the dispatch weight value is calculated, and the scenario with the higher total dispatch weight value is notified first. When the total dispatch weight value is the same within the same category, the scenario with the higher single dispatch weight value is notified first.

[0055] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0056] This invention utilizes RPA (Robotic Process Automation) programs to automatically collect and process line loss data, which is more accurate and reliable than manual methods. Furthermore, the RPA program design employs multiple error prevention and exception handling mechanisms, utilizing multi-function processes and multi-interface configurations to enhance program robustness. In addition, RPA technology itself possesses non-intrusive, non-service interface, and non-data transmission characteristics, enabling agile delivery and development, and shortening the delivery cycle. Secondly, it uses a grey relational analysis-based least squares method and a grey prediction GM(1,1) model, which are computationally simple, easy to implement, require little memory, and have high accuracy, thus compensating for the shortcomings of integrated power and line loss management systems in line loss prediction. The analysis results are displayed digitally on a web interface, allowing managers to intuitively understand the line loss situation. Furthermore, a dispatching method is proposed to promptly notify relevant transformer substations or distribution line managers to address line loss, making line loss notifications more organized, strengthening line loss management, and improving operational efficiency. Attached Figure Description

[0057] Figure 1 This is a flowchart of the RPA program for data acquisition according to an embodiment of the present invention;

[0058] Figure 2 This is a flowchart of a line loss data prediction program according to an embodiment of the present invention;

[0059] Figure 3 This is a flowchart of the dispatching procedure according to an embodiment of the present invention;

[0060] Figure 4 This is an overall flowchart of an embodiment of the present invention. Detailed Implementation

[0061] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0062] like Figure 1-4 As shown, an embodiment of the present invention provides a data-based method for line loss prediction and order dispatch, comprising the following steps:

[0063] S1. Use Cyclone RPA designer to design an integrated power and line loss management system RPA program to automatically collect line loss data of transformer substations and distribution lines on day T-2, where T is the current day. Data collection can be automatically executed at 9:00 every day. RPA programs are more accurate and reliable than manual methods. They adopt multiple error prevention and exception handling mechanisms, and use multi-function processes and multi-interface configurations to improve program robustness. In addition, RPA technology itself has the characteristics of non-intrusiveness, non-service interface and non-data transmission, which can achieve agile delivery and agile development, and shorten the delivery cycle.

[0064] Line loss data includes power supply, electricity sales, independent variables affecting line loss, and indicators affecting line loss. Independent variables affecting line loss include voltage, current, power factor, equipment load rate, and grid operation mode. Indicators affecting line loss include enterprise economic benefits, power quality, equipment operating efficiency, and power supply reliability. The line loss data is processed to remove duplicates and outliers. The line loss rate for day T-2 is calculated based on power supply and electricity sales. The formula for calculating the line loss rate is:

[0065] Line loss rate = (line loss power / power supply) × 100% = (power supply - power sales) / power supply × 100% = (1 - power sales / power supply) × 100%;

[0066] Based on the line loss rate, the line loss situation on day T-2 is obtained. The line loss situation is divided into high loss and negative loss. A line loss rate > 6% is considered high loss, and a line loss rate < 0% is considered negative loss. Both high loss and negative loss are considered substandard. The line loss data is divided into six Excel spreadsheets, named as follows: "Distribution Line Line Loss Substandard Details yyyymmdd", "Distribution Line Line Loss Negative Loss Details yyyymmdd", "Distribution Line Line Loss High Loss Details yyyymmdd", "Transformer Area Line Loss Substandard Details yyyymmdd", and "Transformer Area Line Loss Negative Loss Details yyyymmd". The six Excel spreadsheets, named "Result File yyyy-mm-dd" and "Detailed Data of High Line Loss in Transformer Areas yyyymmdd", are merged into a folder named "Result File yyyy-mm-dd". The system checks if a result file for day T-2 exists on the server. If it does not exist, the "Result File yyyy-mm-dd" is stored on the server. If it exists, the file is deleted and silently overwritten, and the current file is stored on the server. This ensures that there is only one copy of the file on the server, guaranteeing the accuracy and consistency of the data.

[0067] S2. Call the JavaScript library to use the readExcelFile function on the web interface of the integrated power consumption and line loss management system to read the Excel spreadsheet from the server. The readExcelFile function receives the path of the Excel file as a parameter. The read data is preprocessed, including removing null values ​​and outliers, in order to facilitate subsequent line loss data mining and analysis. Analyze the preprocessed line loss data, calculate statistical indicators such as average and standard deviation, and use the displayData function of the chart.js library to convert the calculated line loss data into charts, such as line charts and bar charts, and display the charts on the web interface of the integrated power consumption and line loss management system.

[0068] Based on the independent variables and indicators affecting line loss, the line loss rate is predicted. First, the predicted line loss rate for days T-1 and T is calculated using the least squares method improved by grey relational analysis.

[0069] Let n sequences of independent variables affecting line loss form a matrix:

[0070]

[0071] Where m is the number of indicators affecting line loss, we get:

[0072] X i ′=(x i ′(1),x i ′(2),...,x i ′(m)) T (i = 1, 2, ..., n);

[0073] The dimensionless sequence forms a matrix:

[0074]

[0075] Calculate the absolute difference between the corresponding elements of each evaluated object's indicator sequence and the reference sequence:

[0076] x0(k)-x i (k)(k=1,...,m,i=1,...,n);

[0077] Calculate the correlation coefficient:

[0078]

[0079] Where ρ is the resolution coefficient, taking values ​​in (0,1), the association order is calculated as follows:

[0080]

[0081] Depending on the specific circumstances, a weighted sequence of associations can also be calculated:

[0082]

[0083] Among them W k Assigning weights to each indicator.

[0084] The least squares linear regression model was used to train these key factors to obtain the predicted value of the line loss rate k1;

[0085] Then, the predicted line loss rate for days T-1 and T is calculated using the grey prediction GM(1,1) model:

[0086] Suppose that n independent variables affecting line loss form a sequence:

[0087] x (0) =[x (0) (1),x (0) (2),...,x (0) (n)];

[0088] Perform an accumulation calculation, denoted as 1 - AGO:

[0089]

[0090] The generated sequence is obtained as follows:

[0091] x (1) =[x (1) (1),x (1) (2),...,x (1) (n)];

[0092] The mean sequence is obtained:

[0093] z(w) = 0.5x(w) + 0.5x(w-1);

[0094] To solve the grey system problem, the first-order differential equation of GM(1,1) is:

[0095]

[0096] List the Y and B matrices:

[0097]

[0098] The parameter vector to be identified is Φ = [au] T And Y = BΦ, the parameter vector obtained by the least squares method is:

[0099] The discrete solution is obtained:

[0100]

[0101] The predicted line loss rate k2 is obtained:

[0102]

[0103] Finally, the weighted line loss rate prediction value K is obtained:

[0104] K = g1k1 + g2k2;

[0105] Where g1 and g2 are weights, g1+g2=1. In this embodiment, g1=g2=0.5. Different weight values ​​can also be assigned according to the preference for the prediction model. Based on the predicted line loss rate, the predicted line loss situation on day T-1 and day T is obtained. The grey relational improved least squares method and grey prediction GM(1,1) model used here are simple to operate, easy to implement, occupy little memory space, and have high accuracy, thus making up for the shortcomings of the integrated power and line loss management system in line loss prediction.

[0106] Each algorithm modeling process requires error verification. The relative error value and grade ratio deviation value are checked through model residual verification to evaluate the model's fitting effect. The relative error value should be less than 20%, and the grade ratio deviation value should be less than 0.2.

[0107] After the forecast is completed, the latest forecast results are displayed in chart form on the web interface of the integrated power and line loss management system, so that managers can intuitively understand the line loss situation. Step S2 is repeated every once in a while, and a timer can be set to execute it automatically at 12:00 every day to continuously monitor and analyze line loss data.

[0108] S3. Each manager is responsible for multiple lines in their area. Each line may have different line loss conditions. The severity of the line loss conditions is compared, and the line loss conditions for day T-2 and the predicted line loss conditions for day T-1 and day T are assigned to the managers in order of severity from most severe to least severe.

[0109] First, the existing line loss situations are divided into four types: A. There is a failure to meet the standard on day T-2; B. It is predicted that there will be a failure to meet the standard on day T-1 or day T; C. There is a line loss rate exceeding the 90% warning line on day T-2; D. It is predicted that there will be a line loss rate exceeding the 90% warning line on day T-1 or day T. The line loss rate exceeding the 90% warning line is defined as a line loss rate > 5.4% or a line loss rate < 0.9%.

[0110] Then, assign order weight values ​​of 0.50, 0.25, 0.20, and 0.05 to the four scenarios mentioned above, respectively. Scenario A, which exists, is designated as category A, and scenario B, which does not exist, is designated as category B. Category A scenarios have higher order priority than category B scenarios. The order assignment order ordering ...

[0111] For example, if Manager 1's area has 1 Class A and 4 Class D line losses, Manager 2's area has 1 Class A and 1 Class C line loss, Manager 3's area has 2 Class B and 1 Class C line losses, and Manager 4's area has 10 Class D line losses, then because Class A line losses exist, Managers 1 and 2's line loss situations are classified as Class A, while Managers 3 and 4's are classified as Class B. Managers 1 and 2 have a higher priority for dispatching orders than Managers 3 and 4. Since Managers 1 and 2 have the same dispatching priority, the dispatching weight is calculated. The weights for dispatching orders are calculated as follows: Order #1 = 1 × 0.5 + 4 × 0.05 = 0.7, and Order #2 = 1 × 0.5 + 1 × 0.2 = 0.7. Since the order dispatch weights for managers #1 and #2 are the same, but manager #2's situation C takes precedence over manager #1's situation D, therefore, order #2 is dispatched first. Similarly, the order dispatch weights for managers #3 and #4 are 0.7 and 0.5 respectively, so manager #3 is dispatched first. Therefore, based on the severity of the line loss, the order dispatching order for the four managers is: #2, #1, #3, #4. Furthermore, for each manager, the priority order for line loss management is: A, B, C, D. This order dispatching method allows for timely notification of relevant transformer substations or distribution line managers to manage line losses, making line loss notification more organized, strengthening line loss management, and improving operational efficiency.

[0112] This invention utilizes RPA (Robotic Process Automation) programs to automatically collect and process line loss data, which is more accurate and reliable than manual methods. Furthermore, the RPA program design employs multiple error prevention and exception handling mechanisms, utilizing multi-function processes and multi-interface configurations to enhance program robustness. In addition, RPA technology itself possesses non-intrusive, non-service interface, and non-data transmission characteristics, enabling agile delivery and development, and shortening the delivery cycle. Secondly, it uses a grey relational analysis-based least squares method and a grey prediction GM(1,1) model, which are computationally simple, easy to implement, require little memory, and have high accuracy, thus compensating for the shortcomings of integrated power and line loss management systems in line loss prediction. The analysis results are displayed digitally on a web interface, allowing managers to intuitively understand the line loss situation. Furthermore, a dispatching method is proposed to promptly notify relevant transformer substations or distribution line managers to address line loss, making line loss notifications more organized, strengthening line loss management, and improving operational efficiency.

[0113] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A data-based line loss prediction and dispatch method, characterized in that, The method comprises the following steps: S1. Collecting line loss data on the T-2 day, T being the current day, the line loss data comprising power supply, power sales, line loss influencing self variables and line loss influencing indexes, calculating the line loss rate on the T-2 day according to the power supply and the power sales, and obtaining the line loss situation on the T-2 day; The line loss influencing self variables comprise voltage, current, power factor, equipment load rate and power grid operation mode, and the line loss influencing indexes comprise enterprise economic benefit, power quality, equipment operation efficiency and power supply reliability; S2. Calculating the line loss rate prediction values on the T-1 and T days according to the line loss influencing self variables and the line loss influencing indexes by using the grey correlation improved least square method and the grey prediction GM(1,1) model, and obtaining the predicted line loss situations on the T-1 and T days; The line loss is predicted by using the grey correlation improved least square method: Set a sequence of line loss influencing variables forming a matrix: ； wherein is the number of line loss impact indicators, resulting in: ； The dimensionless sequence forms a matrix: ； The absolute difference value of each evaluated object index sequence and the corresponding element of the reference sequence is calculated one by one: ； The correlation coefficient is calculated: ； wherein are the resolution coefficients, taking values in (0, 1), and the correlation order is calculated as: ； The linear regression model of least square method is used for training to obtain the line loss rate prediction value ; The line loss is predicted by using the grey prediction GM(1,1) model: Set A sequence of line loss influencing variable sequences is formed: ； Once cumulative calculation is performed, denoted as 1-AGO: ； The generated sequence is obtained: ； The mean sequence is obtained: ； The grey system is established for solving, and the GM(1,1) first-order differential equation is: ； The Y and B matrices are listed: ； Parameter vector to be identified , and The parameter vector obtained by the least square method is: ； The discrete solution is obtained: ； obtaining a line loss rate prediction value : ； The final weighted line loss rate prediction value is obtained : ； wherein and are weights; S3. Comparing the line loss situation severity, and assigning the line loss situation on the T-2 day and the predicted line loss situations on the T-1 and T days to the management personnel in the order from heavy to light.

2. The data based line loss prediction and dispatch method according to claim 1, characterized in that, In the step S1, a RPA program of the integrated power and line loss management system is designed by using the Cyclone RPA designer, and the RPA program is used for automatically collecting and arranging the line loss data, and the line loss data is stored in the background server in the form of an Excel table.

3. The data based line loss prediction and dispatch method according to claim 1, characterized in that, In the step S1, the line loss rate calculation formula is: line loss rate=(line loss power / supply power)×100%=(supply power-power sales) / supply power×100%=(1-power sales / supply power)×100%.

4. The data based line loss prediction and dispatch method according to claim 1, characterized in that, In the step S2, the collected line loss data is analyzed, and the analysis of the line loss data comprises data preprocessing and data analysis, the data preprocessing is removing null values and abnormal values, and the data analysis is calculating the mean and the standard deviation.

5. The data based line loss prediction and dispatch method according to claim 4, characterized in that, In the step S2, the analyzed line loss data is converted into a chart by using the displayData function of the chart.js library, and the chart is displayed on the web end of the integrated power and line loss management system.

6. The data based line loss prediction and dispatch method according to claim 1, characterized in that, Said step S2, .

7. The data based line loss prediction and dispatch method according to claim 6, characterized in that, Said step S2, .

8. The data based line loss prediction and dispatch method according to claim 1, characterized in that, In the step S1, the line loss situation is obtained by calculating the line loss rate, and the line loss situation is specifically: high loss when the line loss rate>6%, and negative loss when the line loss rate<0%, both of which are unqualified.

9. The data based line loss prediction and dispatch method according to claim 8, characterized in that, In the step S3, the line loss situation is divided into four types: a. there is unqualified on the T-2 day, b. there is unqualified on the T-1 or T day, c. there is line loss rate exceeding 90% early warning line on the T-2 day, and d. there is line loss rate exceeding 90% early warning line on the T-1 or T day, and the line loss rate exceeding 90% early warning line is line loss rate>5.4% or line loss rate<0.9%. The dispatching weight values of 0.50, 0.25, 0.20 and 0.05 are respectively set for the four cases, the case A is set as a class A, the case B is set as a class B, the dispatching priority of the class A is higher than that of the class B, the dispatching order is: the class A case with higher dispatching priority is preferentially notified to the manager, then the class B case with lower dispatching priority is notified to the manager, when the dispatching priorities are same, the total dispatching weight value is calculated, the line loss case with higher total dispatching weight value is preferentially notified, when the total dispatching weight values are same, the party with higher single dispatching weight value is preferentially notified.

Images