Power outage work planning support system and its program

Abstract

To provide a system and a program therefor, the system being capable of efficiently supporting creation of an outage work plan by accurately performing prediction of distribution line tidal current caused by outage of a transformer and evaluation of feasibility of outage work based on the prediction.SOLUTION: An outage work plan creation support system 1 is constituted by: a calculation device 2 comprising an input unit 2a to which data inside or outside the system is to be input, a teacher data generation unit 2b for generating teacher data, a learned model generation unit 2c for generating a learned model using the teacher data, an evaluation unit 2d for evaluating feasibility of outage work on the basis of the learned model, and an output unit 2e for outputting an evaluation result by the evaluation unit 2d to the outside of the system; and a storage device 3 inside which a substation database 4, a weather information database 5, an input database 6, a teacher data database 7, a learned model database 8, and an output database 9 are constructed.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a power outage work plan creation support system and a program therefor that evaluates the feasibility of power outage work on transformers in substations and supports the creation of work plans. [Background technology]

[0002] When performing outage work on a substation transformer, the feasibility of the outage work is considered based on the status of the distribution line system, the capacity of the distribution line, and past records of the power flow that has occurred on the distribution line in question. If it is predicted that the power flow that will occur on the distribution line due to the outage will exceed the capacity of the remaining transformer, the feasibility of switching that excess power to a nearby substation is also considered. The results of these considerations will be communicated by the relevant parties to the person requesting the outage work, and if there is a possibility that the outage work cannot be carried out as scheduled, a request to change the scheduled date and time of the outage work will be made at the same time as the above notification.

[0003] Here, the procedure for considering whether or not to perform power outage work will be explained using Figure 8. Figure 8 is a diagram showing an example of a power system in two substations, each equipped with two distribution line transformers (hereinafter referred to as Bank 1 transformer and Bank 2 transformer). Note that in Figure 8, the circuit breakers and disconnecting switches marked with an "x" indicate that they are in the "off" state. As shown in Figure 8, substation 50 has bank 1 transformer 50a and bank 2 transformer 50b, and substation 51 has bank 1 transformer 51a and bank 2 transformer 51b. Bank 1 transformer 50a and bank 2 transformer 50b are connected to transmission lines 55a and 55b via primary-side circuit breakers (primary CBs) 52a and 52b and disconnectors 60a and 60b, respectively, and are also connected to buses 56a and 56b via secondary-side circuit breakers (secondary CBs) 53a and 53b, respectively. Furthermore, primary-side circuit breaker 52a and bank 1 transformer 50a are connected to primary-side circuit breaker 52b and bank 2 transformer 50b via electric wire 54a, in which circuit breaker 61a and disconnector 62a are interposed. The busbars 56a and 56b are connected to each other via a busbar connection circuit breaker (busbar connection CB) 57a, and the busbars 56a and 56b are connected to distribution lines 59a and 59b via circuit breakers 58a and 58b, respectively. A consumer 63a and a solar power generation facility 63b are connected to the distribution line 59a.

[0004] Bank 1 transformer 51a and bank 2 transformer 51b are connected to transmission lines 55c and 55d via primary circuit breakers (primary CBs) 52c and 52d and disconnectors 60c and 60d, respectively, and are also connected to busbars 56c and 56d via secondary circuit breakers (secondary CBs) 53c and 53d, respectively. Primary circuit breaker 52c and bank 1 transformer 51a are connected to primary circuit breaker 52d and bank 2 transformer 51b via electric wire 54b, which has circuit breaker 61b and disconnector 62b installed between them. Busbars 56c and 56d are connected to each other via busbar connection circuit breaker (busbar connection CB) 57b, and distribution lines 59c and 59d are connected to busbars 56c and 56d via circuit breakers 58c and 58d, respectively.

[0005] As shown by the "x" mark in Figure 8, when power outage work is performed on Bank 1 transformer 50a of substation 50, primary-side circuit breaker 52a, secondary-side circuit breaker 53a, circuit breaker 61a, and disconnectors 60a and 62a are turned off. As a result, Bank 2 transformer 50b, which is the remaining transformer, is connected to distribution line 59a in addition to the load connected to distribution line 59b, and the load that had previously been applied to Bank 1 transformer 50a is now newly added. In this case, if all the loads connected to distribution lines 59a and 59b are greater than the rated capacity of Bank 2 transformer 50b, Bank 2 transformer 50b will be overloaded, and some of the load will have to be transferred to a transformer in a nearby substation. Specifically, the power flow to Bank 2 transformer 50b that occurs when Bank 1 transformer 50a goes out is predicted from past performance, and if this value does not exceed the rated capacity of Bank 2 transformer 50b, it is determined that the above-mentioned power outage work is possible. On the other hand, if the predicted value of the power flow exceeds the rated capacity of Bank 2 transformer 50b, a survey is conducted to see if there is a transformer at a nearby substation that can handle the load, and if such a transformer is found, it is determined that the above-mentioned power outage work is possible, but if such a transformer is not found, it is determined that the power outage work is impossible, and the person requesting the power outage work is asked to change the date and time of the work. In this way, when considering whether or not to carry out power outage work, it is necessary to predict the power flow in distribution lines based on past performance and compare this prediction with the rated capacity of remaining transformers and transformers at nearby substations. However, conventional systems used to consider whether or not to carry out power outage work do not incorporate information about the transformers and distribution lines at each substation, nor do they incorporate data on past performance data on power flow in distribution lines, so creating a power outage work plan requires a great deal of time and effort.

[0006] For example, Patent Document 1 discloses a technology for selecting a date for a power outage, called a "power outage planning system." This system includes a customer information database storing customer information such as customer locations, power consumption, and electrical equipment; a worker information database storing worker information such as the level of expertise of workers engaged in planned power outages; a map information database storing map information including the planned power outage area; a weather information database storing weather information including the planned power outage area; an input unit for inputting planned power outage conditions such as the planned power outage area and work content; and a planning task for formulating a power outage plan. The planning task calculates a standard work time based on the input planned power outage conditions, corrects the standard work time based on at least one of the customer information, worker information, map information, and weather information, and then selects a day or time period during which the customer's power consumption is low and falls within the calculated work time to formulate a power outage plan. Therefore, this invention accurately calculates work time and appropriately selects a power outage date. Furthermore, Patent Document 2 discloses an invention entitled "Power Outage Plan Evaluation System, Power Outage Plan Evaluation Method, and Power Outage Plan Evaluation Program." This invention relates to power outage plans and is characterized by including a storage means for storing constraints imposed by work at the power station and constraints imposed by the equipment that transmits power from the power station to consumers, and a processing means for determining whether an input power outage plan is acceptable or not based on the constraints stored in the storage means and input consumer constraints. Therefore, this invention makes it possible to evaluate power outage plans taking into account constraints imposed by consumers and constraints on the power supplier side. Furthermore, Patent Document 3 discloses an invention entitled "Method and Apparatus for Setting a Power System Power Outage Work Plan." This invention is characterized in that, when creating a request for a power outage plan for a business's power system, a coordination information database containing a power outage work coordination manual and power outage plans for all business locations is created, and when creating a power outage plan for a business location, the coordination information database is referenced in advance, prior to submitting the request to the supervisory department that oversees each business location, to check at least the items that need to be considered at the power outage work planning stage and consistency with other business locations at the time of planning. Therefore, this invention can simplify the work of coordinating power outage plans at the supervisory department. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Laid-Open No. 2013-123321 [Patent Document 2] Japanese Patent Application Laid-Open No. 2011-61914 [Patent Document 3] Japanese Patent Application Laid-Open No. 2007-244139 Summary of the Invention [Problem to be solved by the invention]

[0008] The invention disclosed in Patent Document 1, which is the above-mentioned prior art, is configured to reflect meteorological information (weather forecast information) stored in a meteorological information database in the adjustment of standard working hours, but it is not configured to predict the distribution line current flow to the remaining transformer of the substation where the power outage work is to be performed, and therefore there is a problem in that it is not possible to compare the distribution line current with the capacity of the remaining transformer to consider the feasibility of performing the power outage work. Furthermore, the inventions relating to the creation of power outage plans disclosed in Patent Documents 2 and 3 are not configured to predict distribution line currents or the amount of power generated by solar power generation facilities based on past performance data, and therefore have the problem of being unable to evaluate whether or not power outage work can be carried out.

[0009] The present invention has been made in response to such conventional circumstances, and aims to provide a system and a program therefor that can efficiently support the creation of power outage work plans by predicting distribution line currents that will occur due to a transformer power outage and evaluating the feasibility of power outage work based on that prediction. [Means for solving the problem]

[0010] In order to achieve the above object, the first invention, a power outage work plan creation support system, is characterized by comprising: a trained power distribution line flow prediction model that is trained by inputting power distribution line flow teacher data in order to evaluate the feasibility of power outage work on transformers at substations; and a calculation device that performs calculations to determine the power distribution line flow based on the planned work timing and weather information forecast data input into the power distribution line flow prediction model, and outputs the calculation results as the predicted power distribution line flow, included in the power outage work plan evaluation data. In the power outage work plan creation support system configured as described above, the planned work timing for power outage work and weather information forecast data are input into the power distribution line current prediction model, and the calculation device calculates the predicted power distribution line current and outputs it as part of the power outage work plan evaluation data.

[0011] The second invention, a power outage work plan creation support system, is characterized in that, in the first invention, it comprises a trained solar power generation prediction model that has been trained by inputting solar power generation teacher data, and the calculation device performs calculations to determine the solar power generation amount based on the weather information forecast data for the scheduled time of the power outage work that has been input into the solar power generation prediction model, and the sum of this predicted value of solar power generation and the predicted value of distribution line flow is set as the predicted distribution line flow. In the present invention, the term "amount of photovoltaic power generation" is used to mean "amount of power generated by a photovoltaic power generation facility." The power outage work plan creation support system configured as described above operates by inputting weather information forecast data for the scheduled time of power outage work into a solar power generation prediction model, and adding the solar power generation amount calculated by the calculation device to the predicted value of distribution line current, and including this in the power outage work plan evaluation data as the predicted distribution line current.

[0012] The third invention, a power outage work plan creation support system, is characterized in that in the first or second invention, the calculation device has an evaluation unit that uses the difference between the predicted distribution line current and the capacity of the remaining transformer at the substation as the transfer load, compares this transfer load with the capacity of other transformers installed in substations nearby the substation, and evaluates the feasibility of power outage work, and outputs the evaluation results as part of the power outage work plan evaluation data. In this invention, "remaining transformer" is used to mean "a transformer that is not subject to power outage work at a substation where power outage work is being performed," and "transferred load amount" is used to mean "the load amount that needs to be transferred to other transformers installed in nearby substations to prevent overloading of the remaining transformer." In the power outage work plan creation support system configured as described above, the evaluation unit compares the difference between the predicted distribution line current and the capacity of the remaining transformer at the substation with the capacity of other transformers installed at nearby substations to evaluate the feasibility of power outage work, and outputs the evaluation results as part of the power outage work plan evaluation data.

[0013] The fourth invention, a power outage work plan creation support program, is a program executed by a computer to evaluate the feasibility of power outage work on transformers at substations, and is characterized in that it has a trained power distribution line flow prediction model that has been trained by inputting power distribution line flow teacher data, performs calculations to determine the power distribution line flow based on the planned work timing and weather information forecast data input into this power distribution line flow prediction model, and outputs the calculation results as the predicted power distribution line flow, included in the power outage work plan evaluation data. In the power outage work plan creation support program configured as described above, the trained power distribution line flow prediction model operates to include the predicted power distribution line flow obtained by inputting the planned work date for the power outage work and forecast data on weather information in the power outage work plan evaluation data and output it.

[0014] The fifth invention is a power outage work plan creation support program, which is characterized in that, in the fourth invention, it includes a trained solar power generation prediction model that has been trained by inputting solar power generation teacher data, performs a calculation to determine the solar power generation amount based on the weather information forecast data for the scheduled time of the power outage work that has been input into this solar power generation prediction model, and determines the predicted power distribution line current as the sum of the predicted solar power generation amount and the predicted value of the distribution line current. The power outage work plan creation support program configured as described above operates by adding the amount of solar power generation calculated by inputting weather information forecast data for the scheduled time of power outage work into a solar power generation prediction model to the predicted value of distribution line current, and including this in the power outage work plan evaluation data as the predicted distribution line current.

[0015] The sixth invention, a power outage work plan creation support program, is characterized in that, in the fourth or fifth invention, the difference between the predicted distribution line current and the capacity of the remaining transformer at the substation is used as the transfer load, and this transfer load is compared with the capacity of other transformers installed at substations nearby the substation to evaluate the feasibility of power outage work, and the evaluation results are output as part of the power outage work plan evaluation data. The power outage work plan creation support program configured as described above evaluates the feasibility of power outage work by comparing the difference between the predicted distribution line current and the capacity of the remaining transformer at the substation with the capacity of other transformers installed at nearby substations, and outputs the evaluation results as part of the power outage work plan evaluation data. [Effects of the Invention]

[0016] According to the power outage work plan creation support system of the first invention, the power flow predicted to occur in the distribution line due to a transformer outage is included in the power outage work plan evaluation data and output, so that the predicted power flow can be compared with the remaining transformer capacity to evaluate the feasibility of power outage work. Furthermore, because the power distribution line power flow prediction accuracy is high, according to the first invention, it is possible to evaluate the feasibility of power outage work with high accuracy and efficiently support the creation of power outage work plans.

[0017] The second invention, the power outage work plan creation support system, adds the predicted value of solar power generation to the predicted value of distribution line current, so that even if the weather on the day of the power outage work is worse than expected and no solar power generation is performed at all, the predicted distribution line current is estimated to be higher so that overload of residual transformers is less likely to occur, making it safer.

[0018] According to the third invention, the power outage work plan creation support system, the possibility of transferring load to other transformers installed in substations near the substation where the power outage work will be performed can be evaluated with even greater accuracy.

[0019] Like the first invention, the fourth invention, the power outage work plan creation support program, can predict the distribution line current caused by a transformer outage and evaluate the feasibility of power outage work based on that prediction with high accuracy, thereby efficiently supporting the creation of power outage work plans.

[0020] In the fifth invention, the power outage work plan creation support program, the predicted value of solar power generation is added to the predicted value of distribution line current. This allows the predicted distribution line current to be estimated higher so that overload of residual transformers is less likely to occur even if the weather on the day of the power outage work is worse than expected and no solar power generation is performed at all, making it as safe as the second invention.

[0021] In the sixth invention, the power outage work plan creation support program, as in the third invention, the feasibility of power outage work can be evaluated with even greater accuracy by considering the possibility of transfer of load to other transformers installed in substations nearby the substation where the power outage work is being performed. [Brief explanation of the drawings]

[0022] [Figure 1] 1 is a block diagram of a power outage work plan creation support system according to an embodiment of the present invention. [Figure 2](a) and (b) are tables showing examples of actual data on distribution line power flow and solar power generation for each substation, and (c) is a table showing an example of weather information. [Figure 3] (a) is a table showing the maximum hourly data for distribution line power flow on weekdays at each substation, divided into the first, middle, and last half of the month. (b) is a table showing the maximum hourly data for distribution line power flow on Saturdays at each substation. (c) is a table showing the maximum hourly data for distribution line power flow on Sundays at each substation. [Figure 4] 10 is a flowchart showing a procedure for generating distribution line power flow data and photovoltaic power generation amount data to be stored in the input database from the tables shown in FIG. 2(b) and FIG. 3(a) to FIG. 3(c). [Figure 5] 10(a) and 10(b) are conceptual diagrams showing the configurations of a trained distribution line power flow prediction model and a trained solar power generation prediction model using neural networks, respectively. [Figure 6] 10 is a flowchart for explaining the flow of processing by an evaluation unit in the power outage work plan creation support system. [Figure 7] 10(a) and 10(b) are diagrams showing an example of power outage work plan evaluation data. [Figure 8] FIG. 1 shows an example of a power system with two substations, each equipped with two distribution transformers. DETAILED DESCRIPTION OF THE INVENTION [Example]

[0023] FIG. 1 is a block diagram of a power outage work plan creation support system 1 according to an embodiment of the present invention. FIGS. 2(a) and 2(b) are tables showing examples of distribution line power flow information and information on the amount of power generated by a solar power generation facility (hereinafter referred to as solar power generation amount), respectively, stored in a substation database 4. FIG. 2(c) is a table showing an example of weather information stored in a weather information database 5. Note that "DB" in FIG. 1 refers to a database. Furthermore, because distribution line power flow is expressed as the amount of power flowing into a power system, kWh (kilowatt-hour) is typically used as a unit. However, in the present specification and drawings, A (ampere), which indicates the amount of current flowing through a transformer per unit time, is used as a unit. Furthermore, VA (volt-ampere) is typically used as a unit for transformer capacity, and W (watt) is typically used as a unit for solar power generation. However, in the present specification and drawings, A (ampere) is used as a unit for both, in line with the power flow. As shown in Figure 1, the power outage work plan creation support system 1 is composed of an arithmetic unit 2 configured as one or more semiconductor chips or computer devices, and a storage device 3 configured as RAM (Random Access Memory) that stores data and computer programs necessary for various processes performed by the arithmetic unit 2. The arithmetic device 2 includes an input unit 2a that receives input of data read from the storage device 3 and data external to the system, a training data generation unit 2b that generates training data, a trained model generation unit 2c that generates a trained model using the training data, an evaluation unit 2d that evaluates the feasibility of power outage work based on the trained model, and an output unit 2e that outputs the evaluation results by the evaluation unit 2d to the outside of the system. Meanwhile, the storage device 3 includes a substation database 4, a weather information database 5, an input database 6, a training data database 7, a trained model database 8, and an output database 9. The input unit 2a is configured with, for example, a keyboard, a mouse, an optical reader, a receiving terminal for an information communication network, etc., and can be used to incorporate a distribution line power flow prediction model 8a (described later) into this system if that prediction model is created outside the system. The input unit 2a is also used to import weather information forecast data (described later) from a communication line (not shown). The output unit 2e is configured by a device such as a display or a printer.

[0024] The substation database 4 readably stores the locations of substations, information about power outages that have been performed in the past and those that are scheduled to be performed in the future, information about the distribution line system, and the types and rated capacities of the transformers installed at each substation. Furthermore, the database 4 readably stores multiple years of actual data on the hourly distribution line power flow at each transformer (see FIG. 2(a)) and multiple years of actual data on the amount of power generated per hour (hereinafter referred to as "photovoltaic power generation amount") by the photovoltaic power generation facility 63b (see FIG. 8) connected to each transformer via buses 56a-56d and distribution lines 59a-59d (see FIG. 2(b)). These data are associated with dates and time periods. The weather information database 5 stores multiple years' worth of actual data 5a (see FIG. 2(c)) and forecast data 5b relating to hourly weather and temperature and daily solar radiation at each substation, each of which is readably associated with a date and time period for the weather, temperature, and solar radiation. Solar radiation refers to the radiant energy that a unit area receives from the sun per unit time, and this solar radiation can be divided into solar radiation from only the solar photosphere (direct solar radiation), solar radiation from areas other than the solar photosphere (diffuse solar radiation), and global solar radiation, which is the sum of these. In this embodiment, global solar radiation is used as the solar radiation contained in the actual data 5a and forecast data 5b of the weather information database 5.

[0025] The input database 6 stores distribution line power flow data 6a and solar power generation data 6b in a readable manner, and the teacher data database 7 stores distribution line power flow teacher data 7a, 7b and solar power generation teacher data 7c, 7d generated by the teacher data generation unit 2b based on the distribution line power flow data 6a and solar power generation data 6b, as will be described later, in a readable manner. The trained model database 8 stores the distribution line power flow prediction model 8a and the solar power generation prediction model 8b generated by the trained model generation unit 2c in a readable manner, and the output database 9 stores the power outage work plan evaluation data 9a in a readable manner. In this embodiment, the trained model generation unit 2c and the teacher data database 7 are configured to be included in the power outage work plan creation support system 1, but the trained model generation unit 2c and the teacher data database 7 may be provided independent of the power outage work plan creation support system 1. For example, a system separate from the power outage work plan creation support system 1 can be configured to use the teacher data database 7 to have the trained model generation unit 2c generate a power distribution line power flow prediction model 8a and a photovoltaic power generation amount prediction model 8b, and store only these prediction models in the trained model database 8.

[0026] Figure 3(a) shows a table created based on the actual data of distribution line power flow on weekdays shown in Figure 2(a). Note that "early," "middle," and "late" in the table refer to the periods "from the 1st to the 10th," "from the 11th to the 20th," and "from the 21st to the end of the month" of each month. The values ​​shown in the table indicate the maximum values ​​for each time period within each period. Therefore, even if the data is for the same period of the same month (either early, middle, or late), if the time period is different, it does not necessarily mean that it is for the same day. Figure 3(b) shows a table created based on the actual data of distribution line power flow on Saturdays shown in Figure 2(a), and the values ​​shown in the table represent the actual maximum values ​​for each hour on Saturdays of each week. Figure 3(c) shows a table created based on the actual data of power flow on distribution lines on Sundays shown in Figure 2(a), and the values ​​shown in the table represent the actual maximum values ​​for each hour on Sundays of each week. To avoid cluttering the diagrams, only data from 2020 is shown in Figures 3(a) to 3(c), but data from multiple years other than 2020 is also included in the tables shown in Figures 3(a) to 3(c). Furthermore, when creating the tables in Figures 3(a) to 3(c), data collected on the day of a substation where power outage work was performed in the past and data collected when the distribution line system was in an unusual state were excluded from the actual data shown in Figure 2(a) based on the power outage work information and distribution line system information stored in the substation database 4. In other words, the tables in Figures 3(a) to 3(c) were created based on actual data when the distribution line was in a normal system and the transformer was not experiencing a power outage.

[0027] The procedure for generating the distribution line power flow data 6a and the solar power generation amount data 6b to be stored in the input database 6 will be described with reference to Fig. 4. Fig. 4 is a flowchart showing the procedure for generating the distribution line power flow data 6a and the solar power generation amount data 6b from the tables shown in Fig. 2(b) and Fig. 3(a) to Fig. 3(c). The distribution line power flow data 6a and the solar power generation amount data 6b are created by extracting the actual data of the distribution line power flow and the solar power generation amount for each transformer in the substation, where the distribution line power flow information and the solar power generation amount information are stored in the substation database 4, from the tables shown in FIG. 2(b) and FIG. 3(a) to FIG. 3(c) according to the procedure shown in FIG. 4. For example, in the case of power outage work on Bank 1 transformer 50a of substation A, it is first determined whether the day the power outage work was performed was a Sunday (step S1 in Figure 4), and if the work day was a Sunday, the maximum value for the relevant time period in the relevant week and the weeks before and after that in the relevant month is found from the table shown in Figure 3(c), and this is used as the actual value of the power flow (hereinafter referred to as distribution line power flow a) for the remaining transformer (Bank 2 transformer 50b) of substation A (step S2 in Figure 4). Note that if the work day was not a Sunday in step S1, step S2 is bypassed and the process proceeds to step S3. To explain step S2 in more detail, for example, if the date and time of the power outage work is from 1:00 PM to 2:00 PM on June 14, 2020, and the work date is Sunday of the second week of June, the power line flow for the time period from 1:00 PM to 2:00 PM on Sundays of the first to third weeks of June is read from the table shown in Figure 3(c), and the maximum value of these three values ​​is taken as the actual value of the power line flow a described above (step S2).Furthermore, by similarly calculating the actual value of power line flow a for other years when power outage work is performed on Sundays, actual data for multiple years can be obtained.

[0028] Next, it is determined whether the day the power outage work was performed is a Saturday or a public holiday (step S3 in FIG. 4). If the work day is a Saturday, the maximum value for the relevant time slot in the relevant week and the weeks before and after that in that month is found from the table shown in FIG. 3(b), and this is used as the actual value of the distribution line power flow a described above (step S4 in FIG. 4). Also, if the work day is a public holiday, there is no correlation between the distribution line power flow and the day of the week, so the data for the Saturday of the week in which the holiday falls is used. In other words, if the work day is a public holiday, the maximum value for the relevant time slot in the relevant week and the Saturday of the weeks before and after that in that month is found from the table shown in FIG. 3(b), and this is used as the actual value of the distribution line power flow a (step S4 in FIG. 4). For example, if the date and time of the power outage work is from 1:00 PM to 2:00 PM on June 13, 2020, the work date is the Saturday of the second week of June. Therefore, the power flow for the time period from 1:00 PM to 2:00 PM on Saturdays of the first to third weeks of June is read from the table shown in FIG. 3(b), and the maximum of the three values ​​is set as the actual value of the power flow a described above (step S4). If the date and time of the power outage work is from 4:00 PM to 5:00 PM on August 10, 2020, the work date is the national holiday of the second week of August. Therefore, the power flow for the time period from 4:00 PM to 5:00 PM on Saturdays of the first to third weeks of August is read from the table shown in FIG. 3(b), and the maximum of the three values ​​is set as the actual value of the power flow a described above (step S4). Furthermore, similar procedures are performed for other years, and the actual values ​​of the power flow a are calculated for cases where power outage work was performed on Saturdays or national holidays, thereby obtaining actual data for multiple years. If it is determined in step S3 that the work day is neither a Saturday nor a public holiday, the process bypasses step S4 and proceeds to step S5.

[0029] In step S5, it is determined whether the day the power outage work was performed is a weekday. If the work day is a weekday, the distribution line power flow for the relevant time period in the relevant period (either the first, middle, or last half of the month) of the relevant month is read from the table shown in Figure 3(a), and this is set as the actual value of the distribution line power flow a (step S6 in Figure 4). For example, if the date and time of the power outage work is from 9:00 to 10:00 on June 9, 2020, the work date is a weekday in early June, so the distribution line power flow for the time period from 9:00 to 10:00 in early June is read from the table shown in Figure 3(a) and this is used as the actual value of the distribution line power flow a described above (step S6).Furthermore, by similarly calculating the actual value of distribution line power flow a for cases in which power outage work was performed on a weekday for other years, actual data for multiple years can be obtained. If the work day is not a weekday in step S5, step S6 is bypassed and the process proceeds to step S7. In step S7, the amount of power generated by the solar power generation facility connected to the distribution line 59b (see FIG. 8) of the remaining transformer (No. 2 bank transformer 50b) of substation A for the relevant time period on the date corresponding to the work day (hereinafter referred to as solar power generation amount b) is read from the table shown in FIG. 2(b), and this is set as the actual value of solar power generation amount b. Then, in step S8, distribution line flow data 6a is created in which the actual value of distribution line flow a obtained in steps S1 to S6 is associated with the time (month and date) when the power outage work was carried out and actual data 5a related to "hourly weather and temperature at each substation" stored in weather information database 5. In addition, solar power generation amount data 6b is created in a state where it is associated with actual data 5a regarding "hourly weather and solar radiation at each substation" in which the actual value of solar power generation amount b obtained in the process of step S7 is stored in the weather information database 5, and is stored in a readable state in the input database 6 together with the distribution line flow data 6a.

[0030] The teacher data generation unit 2b generates distribution line power flow teacher data 7a consisting of actual data 5a on the timing of the power outage work (information on the month and whether the work day was early, middle or late in the month if it was a weekday, and information on which week of the month it was a Saturday, Sunday or public holiday if the work day was not a weekday) and weather and temperature, based on the distribution line power flow data 6a, and also creates distribution line power flow teacher data 7b consisting of actual values ​​of distribution line power flow a corresponding to the distribution line power flow teacher data 7a. In addition, the teacher data generation unit 2b generates solar power generation teacher data 7c consisting of actual data 5a on weather and solar radiation based on the solar power generation data 6b, and generates solar power generation teacher data 7d consisting of actual values ​​of solar power generation b corresponding to the solar power generation teacher data 7c, and stores these in the teacher data database 7 in a readable state together with the distribution line power flow teacher data 7a, 7b and the solar power generation teacher data 7c.

[0031] The trained model generation unit 2c reads out the distribution line power flow teacher data 7a, 7b and the solar power generation teacher data 7c, 7d stored in the teacher data database 7, and then performs deep learning using the distribution line power flow teacher data 7a and the solar power generation teacher data 7c as input teacher data and the distribution line power flow teacher data 7b and the solar power generation teacher data 7d as output teacher data to generate a distribution line power flow prediction model 8a and a solar power generation prediction model 8b, respectively. 5(a) and 5(b) are conceptual diagrams showing the configurations of a trained distribution line power flow prediction model 8a and a trained solar power generation amount prediction model 8b using neural networks, respectively. As shown in FIG. 5(a), the power distribution line power flow prediction model 8a has an input layer including a plurality of neurons to which the planned work month, the planned work date (if it is a weekday, whether it is early, middle, or late in the month; if it is a non-weekday, whether it is a Saturday, Sunday, or holiday) and weather and temperature forecast data 5b are input, and an output layer includes a neuron that outputs a predicted value of the power distribution line power flow a. Also, as shown in FIG. 5(b), the solar power generation forecast model 8b has an input layer including a plurality of neurons to which the weather and solar radiation forecast data 5b is input, and an output layer includes a neuron that outputs a predicted value of the solar power generation amount b. The middle layers of the power distribution line power flow prediction model 8a and the solar power generation forecast model 8b form functions that represent the relationship between the planned work period and weather conditions (weather, temperature, and solar radiation forecast data 5b) and the predicted values ​​of the power distribution line power flow a and the solar power generation amount b (neurons output to the output layer).

[0032] That is, the distribution line flow prediction model 8a is a fully connected neural network that takes the time when the work is scheduled and weather conditions as input variables and the predicted value of the distribution line flow a corresponding to the input as an output variable, and the solar power generation amount prediction model 8b is a fully connected neural network that takes weather conditions as input variables and the predicted value of the solar power generation amount b corresponding to the input as an output variable. In this neural network, an output vector composed of output variables is obtained by performing an affine transformation, or in some cases, a nonlinear transformation, on an input vector composed of input variables using a weight matrix and a bias vector. In this case, the parameters (each component of the weight matrix and bias vector) that define the relationship between the output variable and the input variable are optimized using a technique such as backpropagation. In this way, the learned model generation unit 2c performs deep learning using the distribution line flow teacher data 7a, 7b and the solar power generation teacher data 7c, 7d, thereby generating a distribution line flow prediction model 8a that receives input of forecast data 5b on the time, weather, and temperature for the planned power outage work and outputs a predicted value of distribution line flow a, and a solar power generation prediction model 8b that receives input of forecast data 5b on weather and solar radiation and outputs a predicted value of solar power generation b. When planning power outage work for a transformer installed in a substation where information on the transformer and distribution line system is stored in the substation database 4, this distribution line flow prediction model 8a and solar power generation prediction model 8b are used to predict the current generated in the remaining transformer (distribution line flow a) and the amount of power generated by the solar power generation equipment connected to the distribution line of the remaining transformer (solar power generation b) at the time and during which the work is planned.

[0033] The function of the evaluation unit 2d of the calculation device 2 will be described with reference to Figures 6 and 7. Figure 6 is a flowchart for explaining the processing flow of the evaluation unit 2d in the power outage work plan creation support system 1, and Figures 7(a) and 7(b) are diagrams showing an example of power outage work plan evaluation data 9a. Figure 6 shows the steps executed by a computer for the power outage work plan creation support program of the present invention, and explaining the processing flow of the evaluation unit 2d in the power outage work plan creation support system 1 with reference to Figure 6 will also explain an embodiment of the power outage work plan creation support program. Note that the components shown in Figures 1 to 5 and 8 are assigned the same reference numerals and their explanations will be omitted. As shown in step S1 of FIG. 6, in the power outage work plan creation support system 1, when information (such as names and identification numbers) about the substations and transformers where power outage work is scheduled, the scheduled work period, and forecast data 5b about weather information (hourly weather and temperature, and daily global solar radiation) about the substations are input from the input unit 2a of the computing device 2, the evaluation unit 2d searches the substation database 4 to obtain information about the transformers installed in the substation in question, as well as information about transformers installed in neighboring substations. This identifies remaining transformers (transformers not targeted for power outage work at the substation in question) and transformers installed in neighboring substations that may be able to transfer load (hereinafter referred to as transfer destination transformers) (step S2 of FIG. 6). Furthermore, the evaluation unit 2d reads the rated capacity of the remaining transformers from the substation database 4 and sets 90% of that as the remaining transformer capacity (step S3 of FIG. 6). Next, the evaluation unit 2d inputs the forecast data 5b of the time, weather, and temperature for the planned power outage work, which were input from the input unit 2a in step S1, into the distribution line power flow prediction model 8a, and inputs the weather and solar radiation forecast data 5b into the solar power generation prediction model 8b. This results in predicted values ​​for the distribution line power flow a and the solar power generation amount b. The predicted value for the solar power generation amount b obtained as the output of the solar power generation prediction model 8b reflects the effects of the weather on the planned work day, but there may be cases where the weather on the work day is worse than expected and no solar power generation occurs at all. In such cases, there is a risk that a power flow equivalent to the sum of the predicted value for the distribution line power flow a and the predicted value for the solar power generation amount b will flow through the residual transformer. Therefore, in the power outage work plan creation support system 1, the evaluation unit 2d makes a conservative estimate and determines the sum of the predicted value of the distribution line current a and the predicted value of the solar power generation amount b as the distribution line predicted current c (the current that is expected to flow into the remaining transformer of the substation where power outage work is planned) (step S4 in Figure 6).

[0034] If the capacity of the remaining transformer obtained in step S3 is greater than the distribution line predicted power flow c obtained in step S4, the remaining transformer will not be overloaded, and power outage work can be carried out. On the other hand, if the capacity of the remaining transformer is smaller than the distribution line predicted power flow c, the remaining transformer will be overloaded, and so it will be necessary to transfer the load to a transformer installed in a nearby substation. If transferring the load to another transformer installed in a nearby substation is not possible, power outage work cannot be carried out. Therefore, the evaluation unit 2d subtracts the distribution line predicted power flow c from the remaining transformer capacity to determine the amount of transfer load (step S5 in Figure 6), and compares the remaining transformer capacity with the distribution line predicted power flow c (step S6 in Figure 6). Note that the "transfer load amount" refers to the amount of load that needs to be transferred to another transformer installed in a nearby substation to prevent the remaining transformer from being overloaded. Then, if the capacity of the transformer is greater than the distribution line predicted current c (i.e., if the transfer load amount is positive), the evaluation unit 2d determines that the power outage work can be carried out without the need to transfer the load (step S12 in Figure 6), and creates power outage work plan evaluation data 9a (see Figure 7(a)) that includes the planned date and time of the work, the rated capacity of the remaining transformer, and the distribution line predicted current c, as well as the evaluation result regarding the feasibility of the work (step S13 in Figure 6). In step S6, if the evaluation unit 2d determines that the capacity of the transformer is not greater than the distribution line predicted power flow c (i.e., the transfer load amount is not positive), it reads from the substation database 4 a transformer installed in a nearby substation that is a candidate for the transfer destination transformer, and sets 90% of its rated capacity as the capacity of the transfer destination transformer (step S7 in Fig. 6).Then, for that transformer, predicted values ​​of the distribution line predicted power flow a and solar power generation amount b are calculated as in the case of a transformer undergoing power outage work, and the sum of these values ​​and the transfer load amount obtained in step S5 is set as the distribution line predicted power flow d (step S8 in Fig. 6). In step S9, if the capacity of the transfer destination transformer obtained in step S7 is greater than the distribution line predicted power flow d obtained in step S8, the evaluation unit 2d determines that power outage work can be carried out by transferring the load to the transfer destination transformer (step S11 in FIG. 6), and creates power outage work plan evaluation data 9a (see FIG. 7(b)) including the scheduled work date and time, the rated capacity of the remaining transformer, the distribution line predicted power flow c, the rated capacity of the transfer destination transformer, and the distribution line predicted power flow d, as well as an evaluation result regarding the feasibility of the work (step S13 in FIG. 6). On the other hand, in step S9, if the capacity of the transfer destination transformer is not greater than the distribution line predicted power flow d, the evaluation unit 2d determines that power outage work cannot be carried out because it is not possible to transfer the load to the transfer destination transformer (step S10 in FIG. 6), and creates power outage work plan evaluation data 9a (step S13 in FIG. 6). The power outage work plan evaluation data 9a created in step S13 is stored in the output database 9.

[0035] As described above, the power outage work plan creation support system 1 can predict the distribution line power flow caused by a transformer outage and evaluate the feasibility of power outage work based on that prediction with high accuracy, thereby efficiently supporting the creation of power outage work plans. Furthermore, the power outage work plan creation support system 1 can consider the possibility of transferring load to other transformers installed in substations near the substation where the power outage work is to be performed. In this case, even if the remaining transformers at the substation where the power outage work is planned cannot handle the load alone, the power outage work can be performed by transferring load to the other transformers, thereby increasing the feasibility of the power outage work. In particular, in power outage work planning support system 1, the planned time for power outage work is also input into distribution line power flow prediction model 8a, so the predicted value of distribution line power flow a reflects not only meteorological conditions such as weather and temperature, but also the relationship with the planned month and day of the week for the work. Therefore, power outage work planning support system 1 equipped with distribution line power flow prediction model 8a can predict with high accuracy the above-mentioned distribution line power flow a that may occur due to power outage work on a transformer. Furthermore, because forecast data 5b of temperature and solar radiation amount is input to solar power generation prediction model 8b, the weather conditions of temperature and solar radiation amount are reflected in the prediction of solar power generation amount b. Therefore, power outage work plan creation support system 1 equipped with solar power generation prediction model 8b can predict with high accuracy the amount of power generated by the solar power generation facility connected to the distribution line of the transformer for which power outage work is scheduled (solar power generation amount b).

[0036] The power outage work plan creation support system and power outage work plan creation support program of the present invention are not limited to the configurations shown in the above-described embodiments. For example, the meteorological information used to generate the distribution line power flow teacher data 7a and the distribution line power flow prediction model 8a may be either weather or temperature. Also, the meteorological information used to generate the solar power generation amount teacher data 7c and the solar power generation amount prediction model 8b may be either weather or solar radiation. Furthermore, the predicted distribution line power flows c and d may be calculated based only on the predicted value of the distribution line power flow a, without using the predicted value of the solar power generation amount b. [Industrial Applicability]

[0037] The present invention can be used to create a power outage work plan for a transformer in a substation. [Explanation of symbols]

[0038] 1...Power outage work plan creation support system 2...Calculation device 2a...Input unit 2b...Training data generation unit 2c...Trained model generation unit 2d...Evaluation unit 2e...Output unit 3...Storage device 4...Substation database 5...Weather information database 5a...Actual data 5b...Forecast data 6...Input database 6a...Distribution line power flow data 6b...Photovoltaic power generation amount data 7...Training data database 7a, 7b...Distribution line power flow training data 7c, 7d...Photovoltaic power generation amount training data 8...Trained model database 8a...Distribution line power flow prediction model 8b...Photovoltaic power generation amount prediction model 9...Output database 9a...Power outage work plan evaluation data 50, 51...Substation 50a, 51a...Bank 1 transformer 50b, 51b...Bank 2 transformer 52a to 52d...Primary side circuit breaker (primary side CB) 53a~53d...Secondary side circuit breaker (secondary side CB) 54a, 54b...Electric wire 55a~55d...Transmission line 56a~56d...Bus bar 57a, 57b...Bus bar connection circuit breaker (bus link CB) 58a~58d...Breaker 59a~59d...Distribution line 60a~60d...Disconnector 61a, 61b...breaker 62a, 62b...disconnector 63a...consumer 63b...solar power generation equipment

Claims

1. A trained distribution line power flow prediction model trained by inputting distribution line power flow training data in order to evaluate the feasibility of transformer outage work at a substation; a computing device that performs a calculation to determine the distribution line power flow based on the forecast data of the scheduled work timing of the power outage work and the weather information input into the power distribution line power flow prediction model, and outputs the calculation result as a predicted power distribution line power flow in the power outage work plan evaluation data; A trained solar power generation prediction model trained by inputting solar power generation training data, The calculation device performs a calculation to determine the amount of solar power generation based on the forecast data of the weather information for the scheduled time of the power outage work input into the solar power generation prediction model, and the sum of the predicted value of the solar power generation and the predicted value of the distribution line flow is set as the predicted distribution line flow.

2. A trained distribution line power flow prediction model trained by inputting distribution line power flow training data in order to evaluate the feasibility of transformer power outage work at a substation; a computing device that performs a calculation to determine the distribution line power flow based on the forecast data of the scheduled work timing of the power outage work and weather information input into the power distribution line power flow prediction model, and outputs the calculation result as a predicted power distribution line power flow in the power outage work plan evaluation data; The calculation device has an evaluation unit that evaluates the feasibility of the power outage work by comparing the difference between the predicted distribution line flow and the capacity of the remaining transformer at the substation as the transfer load amount, with the capacity of other transformers installed at substations nearby the substation, and outputs the evaluation results as part of the power outage work plan evaluation data.

3. The power outage work plan creation support system of claim 1, characterized in that the calculation device has an evaluation unit that evaluates the feasibility of the power outage work by comparing the difference between the predicted distribution line current and the capacity of the remaining transformer at the substation as the transfer load amount and the transfer load amount with the capacity of other transformers installed at substations nearby the substation, and outputs the evaluation results as part of the power outage work plan evaluation data.

4. A program executed by a computer to evaluate the feasibility of performing outage work on a transformer at a substation, comprising: A trained distribution line power flow prediction model trained by inputting distribution line power flow training data, A trained solar power generation prediction model trained by inputting solar power generation training data, a power outage work plan creation support program that performs a calculation to determine the distribution line current based on the forecast data for the planned work timing of the power outage work and weather information that have been input into the distribution line current flow prediction model, and outputs the calculation result as the predicted distribution line current by including it in the power outage work plan evaluation data; and performs a calculation to determine the amount of solar power generation based on the forecast data for the weather information for the planned work timing of the power outage work that have been input into the solar power generation amount prediction model, and sets the sum of the predicted value of the solar power generation amount and the predicted value of the distribution line current as the predicted distribution line current.

5. A program executed by a computer to evaluate the feasibility of power outage work on a transformer at a substation, comprising: It is equipped with a trained distribution line power flow prediction model that has been trained by inputting distribution line power flow teacher data, A calculation is performed to determine the distribution line power flow based on the planned work timing of the power outage work and the weather information forecast data input into this distribution line power flow prediction model, and the calculation result is included in the power outage work plan evaluation data as the distribution line predicted power flow and is output; A power outage work plan creation support program that evaluates the feasibility of the power outage work by comparing the difference between the predicted distribution line current and the capacity of the remaining transformer at the substation as the transfer load amount and comparing this transfer load amount with the capacity of other transformers installed at substations nearby the substation, and includes the evaluation results in the power outage work plan evaluation data and outputs it.

6. The power outage work plan creation support program of claim 4, characterized in that the difference between the predicted distribution line current and the capacity of the remaining transformer at the substation is used as the transfer load, and this transfer load is compared with the capacity of other transformers installed at substations nearby the substation to evaluate the feasibility of the power outage work, and the evaluation results are included in the power outage work plan evaluation data and output.

Images