Unmatching classification device and unmatching classification method

Abstract

To provide an unmatching classification device capable of reducing the burden required for responding to unmatching, and an unmatching classification method.SOLUTION: An unmatching classification device 1 comprises: a model generation section 22 for generating an unmatching classification model 14 by performing machine learning while defining, as teacher data, unmatching in which data registered on a first ledger 101 do not match data registered on a second ledger 110, and a response to the unmatching; an unmatching classification section 24 for determining whether or not there is unmatching by using the unmatching classification model 14 and for classifying at least one of a content of the unmatching and the response to the unmatching; and an output section 32 which outputs a classification result.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to a mismatch classification device and a mismatch classification method for ledgers related to power distribution (power distribution related ledgers).

Background Art

[0002] For power flow calculation in a distribution automation system, various data such as data related to power distribution and data related to customers are required. These data may be registered redundantly in multiple ledgers. To enhance the soundness of the distribution automation system, the consistency of data between ledgers may be confirmed.

[0003] Patent Document 1 discloses a technique in a system for monitoring a power system in which a state estimation method and a power flow calculation method are combined. Regarding the inconsistency between the estimated state of the power system and the state of the power system to be monitored and controlled, an abnormality detection unit determines whether it is abnormal using a determination formula.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] By the way, regarding the result of determining the consistency of data between power distribution related ledgers, the responses to matters determined to be inconsistent are diverse. Hereinafter, being inconsistent is called a mismatch. Also, matters determined to be a mismatch may simply be called a mismatch. Examples of responses to a mismatch include ledger correction, confirmation of the current status of power distribution facilities, and inquiries to other departments. The response to a mismatch places a large human burden. Therefore, an object of the present invention is to provide a mismatch classification device and a mismatch classification method that can reduce the burden required for responding to a mismatch.

Means for Solving the Problem

[0006] (1) The mismatch classification device of the present invention is a mismatch classification device used in a system that performs power flow calculation based on data registered in a first ledger and operates a power system based on the result of the power flow calculation. The device includes a model generation unit that performs machine learning using, as teacher data, a mismatch in which data registered in the first ledger does not match data registered in a second ledger having the same item data as the data registered in the first ledger, and a correspondence to the mismatch, and generates a mismatch classification model; a mismatch classification unit that uses the mismatch classification model to determine whether there is a mismatch between the data registered in the first ledger and the data registered in the second ledger, and classifies at least one of the content of the mismatch and the correspondence to the mismatch when there is a mismatch; and an output unit that outputs the classification result of the mismatch classification unit.

[0007] (2) In the mismatch classification device, the mismatch classification unit may classify the mismatch according to whether human correspondence is required.

[0008] (3) In the mismatch classification device, the output from the output unit may include content prompting correspondence to the mismatch to a predetermined department or a predetermined person.

[0009] (4) The unmatching classification method of the present invention is an unmatching classification method used in a system that performs power flow calculation based on the data registered in the first ledger and operates the power system based on the result of the power flow calculation. The method includes a model generation step of machine learning an unmatching where the data registered in the first ledger does not match the data registered in the second ledger in which data of the same item as the data registered in the first ledger is registered, and the correspondence to the unmatching as teacher data to generate an unmatching classification model, an unmatching classification step of using the unmatching classification model to determine whether there is an unmatching between the data registered in the first ledger and the data registered in the second ledger, and classifying at least one of the content of the unmatching and the correspondence to the unmatching when there is an unmatching, and an output step of outputting the classification result of the unmatching classification step.

Effects of the Invention

[0010] According to the present invention, it is possible to provide an unmatching classification device and an unmatching classification method capable of reducing the burden required for the correspondence to the unmatching.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Embodiments for Carrying Out the Invention

[0012] Embodiments for carrying out the invention will be described with reference to the drawings. FIG. 1 is a diagram showing the relationship between a distribution automation system 102 and a distribution-related ledger.

[0013] (Distribution Automation System) The power distribution automation system 102 performs power flow calculations. The power distribution automation system 102 determines the actions necessary for the proper operation of the power system based on the calculation results. Examples of actions are changes to the settings related to power distribution. The power distribution automation system 102 determines, for example, where and how much power should be supplied in response to power demand forecasting and changes in customer contract details. Also, the power distribution automation system 102 determines what changes to the settings related to power distribution are necessary to achieve this.

[0014] As shown by the arrow L12 in FIG. 1, section data is registered in the power distribution automation system 102 from the high-voltage system ledger 101. The power distribution automation system 102 performs power flow calculations based on the data in the registered high-voltage system ledger 101. In the high-voltage system ledger 101, data used for power flow calculations, such as pole numbers, data related to transformers and switches, is registered.

[0015] (Power distribution related ledger) The overall picture of the power distribution related ledger is described. As shown in FIG. 1, the power distribution related ledger includes the high-voltage system ledger 101 as the first ledger and the individual ledger 110 as the second ledger. The individual ledger 110 includes the bank ledger 111, the equipment ledger 112, the incoming pole ledger 113, and the customer ledger 114. Note that each ledger shown in FIG. 1 is an example. The power distribution related ledger may include ledgers other than those shown in FIG. 1. Also, the power distribution related ledger may not include all of the ledgers shown in FIG. 1.

[0016] (Power distribution map data system) The registration of data to the ledger when there are works related to power distribution, etc. is described. When there are works related to power distribution, etc., the details are registered in the power distribution map data system 131. The registration of data to the power distribution map data system 131 is performed without delay from the work.

[0017] The data registered in the power distribution map data system 131 is registered in each ledger in two flows. One flow is defined as the first flow L1 and is indicated by the arrow L1 in FIG. 1. The other flow is defined as the second flow L2 and is indicated by the arrow L2 in FIG. 1. Also, the flow of change data is indicated by the arrows L11 to L12 and L21 to L23 in FIG. 1.

[0018] (First Flow) The first flow L1 is a flow in which, when data is registered in the power distribution map data system 131, the data is registered in the high-voltage system ledger 101 as soon as possible without delay. In the first flow L1, when work or the like is carried out, the data is timely registered in the high-voltage system ledger 101. In order to perform the power flow calculation correctly, the data registered in the high-voltage system ledger 101, which is the basis of the power flow calculation, needs to reflect the actual situation. Therefore, the registration of data in the high-voltage system ledger 101 is carried out without delay as much as possible.

[0019] The arrow L11 in FIG. 1 indicates the flow of change data from the power distribution map data system 131 to the high-voltage system ledger 101. The transmission of change data from the power distribution map data system 131 to the high-voltage system ledger 101 is carried out via a transfer slip.

[0020] The data registered in the high-voltage system ledger 101 is registered in the distribution automation system 102. The arrow L12 in FIG. 1 indicates the flow of change data from the high-voltage system ledger 101 to the distribution automation system 102. The transmission of change data from the high-voltage system ledger 101 to the distribution automation system 102 is carried out as section data.

[0021] (Second Flow) The second flow L2 is a flow in which, after the data is registered in the power distribution map data system 131, it is registered in the individual ledger 110 after a certain number of days. In the second flow L2, when work or the like is carried out, the data is registered in the individual ledger 110 with a time lag. The registration of data in the individual ledger 110 is carried out, for example, in the case of work, after inspection and passing the inspection. Therefore, for example, the results of the work are inspected after the end of the month of the work completion date, and after passing the inspection, the data is registered in the ledger.

[0022] (Individual ledger 110) The individual ledger 110 includes a plurality of ledgers in which at least some of the items of data to be registered are different. In the example shown in FIG. 1, the individual ledger 110 includes a bank ledger 111, a facility ledger 112, a lead-in pole ledger 113, and a customer ledger 114.

[0023] (Bank ledger) The pole number is registered in the bank ledger 111. The arrow L21 in FIG. 1 indicates the flow of change data from the distribution map data system 131 to the bank ledger 111. The transmission of change data from the distribution map data system 131 to the bank ledger 111 is performed via the outside line design document.

[0024] (Facility ledger) The pole number, transformer, and switch are registered in the facility ledger 112. The arrow L22 indicates the flow of change data from the distribution map data system 131 to the facility ledger 112. The transmission of change data from the distribution map data system 131 to the facility ledger 112 is performed via the outside line design document.

[0025] (Lead-in pole ledger) The pole number and contract number are registered in the lead-in pole ledger 113. The arrow L23 indicates the flow of change data from the distribution map data system 131 to the lead-in pole ledger 113. The transmission of change data from the distribution map data system 131 to the lead-in pole ledger 113 is performed via the outside line design document (high-voltage contract).

[0026] In addition to the registration of data based on the outside line design document for the high-voltage contract, data registration based on, for example, the lead-in wire and instrument construction design document 121 for the low-voltage contract is also performed in the lead-in pole ledger 113. The arrow L24 indicates the flow of change data based on the lead-in wire and instrument construction design document 121 for the low-voltage contract.

[0027] (Customer ledger) In the customer ledger 114, the customer's contract number, customer name, and customer address are registered. Different from other individual ledgers 110, the data in the customer ledger 114 is not registered based on the change data from the distribution map data system 131. The data in the customer ledger 114 is registered based on a new contract with the customer and changes in the customer's contract details, etc.

[0028] (Mismatch list) In order to maintain the high integrity of the distribution automation system 102, the data registered in the high-voltage system ledger 101 needs to be accurate. To verify that the data registered in the high-voltage system ledger 101 is accurate, it is verified whether the data registered in the high-voltage system ledger 101 matches the data registered in the individual ledger 110. This verification can be performed at predetermined intervals. The predetermined intervals can be, for example, once a month.

[0029] As a result of verifying whether they match, if the data registered in the high-voltage system ledger 101 does not match the data registered in the individual ledger 110, it is called a mismatch. A list in which the mismatched items are listed is called a mismatch list.

[0030] An example of the mismatch list will be described. L31 in Figure 1 shows the verification of whether the high-voltage system ledger 101 and the equipment ledger 112 match. In this verification, it is verified whether the data of the pole number, transformer, and switchgear registered in the high-voltage system ledger 101 matches the data of the pole number, transformer, and switchgear registered in the equipment ledger 112. As a result of this verification, if there is mismatched data, a high-voltage system - equipment ledger mismatch list is created. The high-voltage system - equipment ledger mismatch list describes the mismatched contents such as the mismatched items.

[0031] L32 indicates the verification of whether the high-voltage system ledger 101 and the incoming pole ledger 113 are consistent. In this verification, it is verified whether the data of the contract number (high-voltage) registered in the high-voltage system ledger 101 is consistent with the data of the contract number (high-voltage) registered in the incoming pole ledger 113. As a result of this verification, if there is inconsistent data, a high-voltage customer verification list (high-voltage system ledger 101 - incoming pole ledger 113 unmatched list) is created. The high-voltage customer verification list describes the inconsistent content, such as the inconsistent items.

[0032] Note that the verification of whether the data registered in the high-voltage system ledger 101 is consistent with the data registered in the individual ledger 110 is not limited to the verifications of L31 and L32 described above. Verifications can also be performed on whether the data registered in the high-voltage system ledger 101 and other ledgers included in the individual ledger 110 are consistent.

[0033] (Corresponding to Unmatched) Explain the corresponding measures for the items described in the unmatched list. The content of the corresponding measures varies depending on the content of the unmatched. Therefore, the content of the corresponding measures for the unmatched described in the unmatched list is diverse. Hereinafter, examples of the corresponding measures for the unmatched will be explained.

[0034] (Corresponding to the High-Voltage System Ledger - Equipment Ledger Unmatched List) Explain an example of the corresponding measures for the high-voltage system ledger - equipment ledger unmatched list. The condition for outputting to the high-voltage system ledger - equipment ledger unmatched list is that there is no match when comparing the equipment (electric poles, transformers, and switches) in the high-voltage system ledger 101 and the equipment ledger 112. The creation date of the unmatched list, in other words, the date for verifying the data consistency, can be, for example, a fixed date every month.

[0035] (1) When the unmatched is about an electric pole. Compare the utility poles in the high-voltage system ledger 101 and the equipment ledger 112. If a utility pole is registered in only one of the ledgers, an error is displayed on the side of the ledger where the utility pole is registered. Examples of how to handle this error are as follows. Use construction photos and the like registered in the distribution map data system 131 to confirm the existence of the equipment, that is, the utility pole. Then, correct the ledger in which the incorrect data is registered. When correcting the high-voltage system ledger 101, for example, the high-voltage system ledger 101 can be corrected via the high-voltage system management system. The high-voltage system management system is a system that manages the high-voltage system ledger 101.

[0036] (2) If the mismatch is about the transformer. Compare the transformers in the high-voltage system ledger 101 and the equipment ledger 112. If a transformer is registered in only one of the ledgers, an error is displayed on the side of the ledger where the transformer is registered. Examples of how to handle this error are as follows. Use construction photos and the like registered in the distribution map data system 131 to confirm the existence of the equipment, that is, the transformer. Then, correct the ledger in which the incorrect data is registered.

[0037] (3) If the mismatch is about the switch. Compare the switches in the high-voltage system ledger 101 and the equipment ledger 112. If a switch is registered in only one of the ledgers, an error is displayed on the side of the ledger where the switch is registered. Examples of how to handle this error are as follows. Use construction photos and the like registered in the distribution map data system 131 to confirm the existence of the equipment, that is, the switch. Then, correct the ledger in which the incorrect data is registered.

[0038] (Correspondence to the high-voltage system ledger - incoming pole ledger (customer ledger) mismatch list) An example of how to handle the non - matching list of the high - voltage system ledger - incoming pillar ledger (customer ledger) will be described. In the following example, the case where the customer ledger is included in the incoming pillar ledger will be described. The condition for output to the non - matching list of the high - voltage system ledger - incoming pillar ledger (customer ledger) is that the high - voltage customer data between the high - voltage system ledger 101 and the incoming pillar ledger 113 does not match. However, the contract in the incoming pillar ledger 113 is determined to be registered in the incoming pillar ledger 113 only when it exists in both the incoming pillar ledger 113 and the customer ledger 114. For contracts in which other ledgers such as the distribution ledger are being automatically updated due to contract changes, etc., they are excluded from the non - matching targets. Also, the creation date of the non - matching list, in other words, the date for verifying data consistency, can be set as the equipment regular update date from the high - voltage system management system (high - voltage system ledger 101) to the distribution automation system 102.

[0039] (1) Case - 1 where the non - match is about high - voltage customer data. If high - voltage customer data exists in the incoming pillar ledger 113 (customer ledger 114) but does not exist in the high - voltage system ledger 101, an error is displayed in the incoming pillar ledger 113. An example of how to handle this error is to register the high - voltage customer data in the high - voltage system ledger 101 by the high - voltage system management system.

[0040] (2) Case - 2 where the non - match is about high - voltage customer data. If high - voltage customer data exists in the high - voltage system ledger 101 but does not exist in the incoming pillar ledger 113 (customer ledger 114), an error is displayed in the high - voltage system ledger 101. An example of how to handle this error is as follows. Check with the sales department whether the contract exists. If data registration is not required, delete the high - voltage customer data in the high - voltage system ledger 101 by the high - voltage system management system.

[0041] (3) Case where the non - match is about duplicate registration. If the contract number is registered repeatedly in the high-voltage system ledger 101, an error will be displayed in the high-voltage system ledger 101. An example of dealing with this error is to delete incorrect high-voltage customer data, such as the contract number, from the high-voltage system ledger 101 by the high-voltage system management system.

[0042] (4) If the mismatch is about the capacity mismatch. When the capacity confirmed from the contract number in the incoming pillar ledger 113 to the contract capacity in the customer ledger 114 does not match the contract capacity registered in the high-voltage system ledger 101, an error will be displayed in the high-voltage system ledger 101. An example of dealing with this error is to correct the contract capacity of the high-voltage customer in the high-voltage system ledger 101 by the high-voltage system management system.

[0043] The above is an example of the content of the mismatch and the response to the mismatch. The content of the mismatch and the response to the mismatch are diverse other than the above examples.

[0044] (Mismatch classification device) Based on Figure 2, the mismatch classification device 1 of this embodiment will be described. Figure 2 is a functional block diagram showing the outline of the mismatch classification device 1. The mismatch classification device 1 of this embodiment can determine the presence or absence of a mismatch. Also, when the mismatch classification device 1 determines that there is a mismatch, it can classify the mismatch.

[0045] (Memory unit and processing unit) The mismatch classification device 1 includes a memory unit 10, a processing unit 20, an input unit 31, and an output unit 32. The memory unit 10 is a part that stores at least one of data, programs, etc. The processing unit 20 is a part that reads at least one of data, programs, etc. from the memory unit 10 and processes the read data. The input unit 31 is a part into which the data necessary to verify whether the data registered in the high-voltage system ledger 101 is accurate is input. The output unit 32 is a part that outputs the result classified by the mismatch classification unit 24, etc.

[0046] In the example shown in FIG. 2, the storage unit 10 stores the teacher data 12 and the unmatched classification model 14. The processing unit 20 includes a model generation unit 22 and an unmatched classification unit 24.

[0047] (Teacher data) The teacher data 12 is learning data used when generating the unmatched classification model 14. The teacher data 12 includes the content of the unmatched and the corresponding to the unmatched. The teacher data 12 may include at least one of the classification of the content of the unmatched and the classification of the corresponding to the unmatched, which will be described later. The unmatched is not limited to the unmatched exemplified above, and can include various unmatched that actually occur or may occur. In addition, the corresponding to the unmatched can include various corresponding, not limited to confirmation of data and correction of data, but also inquiries to other departments and requests for corresponding to other departments.

[0048] (Model generation unit) The model generation unit 22 is a part that reads the teacher data 12 from the storage unit 10, learns the teacher data 12, and generates the unmatched classification model 14. The model generation unit 22 generates the unmatched classification model 14 by, for example, machine learning of the teacher data 12. The generated unmatched classification model 14 is stored in the storage unit 10.

[0049] (Unmatched classification model) The unmatched classification model is a model that can determine the presence or absence of an unmatched and classify the unmatched when it is determined that there is an unmatched. The unmatched classification model is read and used by the unmatched classification unit 24.

[0050] The unmatched classification model can classify unmatched items from various perspectives. The classification perspectives include, for example, the necessity of response, the urgency of response, the content of response, etc. Classification from the perspective of the urgency of response means, for example, that immediate response should be made and it should be postponed until the next verification, etc. Classification from the perspective of the content of response means, for example, the types of specific response content such as the change or deletion of registered data, the necessity of cooperation with other departments, the necessity of human-based response, etc. The case where human-based response is necessary means, for example, when it is necessary to actually check the site, when it is necessary to consult with other departments, when it is necessary for a person to check the ledger due to reasons such as the complex way of checking the ledger, when it is necessary for a person to change data, etc. due to reasons such as the complex way of changing or deleting data.

[0051] (Unmatched Classification Section) The unmatched classification section 24 is a part that determines whether the data registered in the high-voltage system ledger 101 matches the data registered in the individual ledger 110, and classifies the unmatched items when they do not match. The unmatched classification section 24 performs this determination and classification using the unmatched classification model 14. The unmatched classification section 24 reads the unmatched classification model 14 from the storage section 10. Also, data necessary for verifying whether the data registered in the high-voltage system ledger 101 is accurate is input to the unmatched classification section 24. This input is performed via the input section 31.

[0052] In the unmatched classification section 24, the input data is input to the unmatched classification model 14. The unmatched classification model 14 determines whether the data matches, and classifies and outputs the unmatched data for the data that does not match. The classification in the unmatched classification section 24 may be performed from multiple perspectives. The classification may be a comprehensive classification including, for example, the content of response, urgency, the necessity of human-based response, and the necessity of cooperation or communication with other departments.

[0053] The result output from the unmatched classification model 14 is output from the unmatched classification device 1 via the output unit 32. The output may also include communication with other departments and instructions for corresponding to other departments. That is, the output from the output unit 32 may include content that prompts a response to the unmatched to a predetermined department or a predetermined person. Here, "predetermined" means suitable for responding to the unmatched.

[0054] (Flow of classification, etc.) The processing flow in the unmatched classification device 1 will be described. FIG. 3 is a flowchart showing the processing flow of the unmatched classification method. In the figure and the following description, S1 means step 1. The same applies to other steps.

[0055] (S1) S1 is a step of acquiring teacher data. The teacher data is data used when generating the unmatched classification model. The teacher data includes the content of the unmatched and the corresponding measures for the unmatched.

[0056] (S2) S2 is a step of generating the unmatched classification model. The unmatched classification model is generated by subjecting the teacher data acquired in S1 to machine learning or the like. S2 corresponds to the model generation step in the unmatched classification method of the present embodiment.

[0057] (S11) S11 is a step in which ledger data is input. The ledger data means data necessary for verifying whether the data registered in the high-voltage system ledger is accurate.

[0058] (S12) S12 is a step of determining whether there is an unmatched. The determination of whether there is an unmatched is made using the unmatched classification model generated in S2. If there is no unmatched, the process ends. If there is an unmatched, the process proceeds to S13.

[0059] (S13) S13 is the step where the classification of mismatches is performed. The classification of mismatches is carried out using the mismatch classification model generated in S2. S13 corresponds to the mismatch classification step in the mismatch classification method of this embodiment.

[0060] (S14) S14 is the step where the classification result is output. After the output in S14, the process ends. S14 corresponds to the output step in the mismatch classification method of this embodiment.

[0061] The example of the flow shown in FIG. 3 includes the generation of a mismatch classification model and the classification of mismatches. The process flow can be divided into S1 and S2 and S11 to S14. When only generating the mismatch classification model, only S1 and S2 can be executed. When the mismatch classification model has already been generated and only the determination of the presence or absence of a mismatch and the classification of the mismatch are performed, only A11 to A14 can be executed.

[0062] As described above, an example of an embodiment of the present invention has been explained. The present invention is not limited to the above-described embodiments, and various changes, modifications, and combinations are possible.

[0063] In the above description, the mismatch classification model was a model for determining whether there is a mismatch and classifying the mismatch when there is a mismatch. Also, the mismatch classification unit was a part for determining whether there is a mismatch and classifying the mismatch when there is a mismatch. The mismatch classification model may be a model that classifies mismatches without determining whether there is a mismatch. Also, the mismatch classification unit may be a part that classifies mismatches without determining whether there is a mismatch. In this case, for example, the content of the mismatch is input from the input unit to the mismatch classification unit. The mismatch classification unit classifies the input mismatch using the mismatch classification model. The classified result is output from the output unit.

Explanation of Reference Numerals

[0064] 1 Unmatched Classification Device 10 Memory Unit 12 Teacher Data 14 Unmatched Classification Model 20 Processing Unit 22 Model Generation Unit 24 Unmatched Classification Unit 31 Input Unit 32 Output Unit 101 High-Voltage System Ledger (First Ledger) 102 Distribution Automation System 110 Individual Ledger (Second Ledger) 111 Bank Ledger 112 Equipment Ledger 113 Incoming Pole Ledger 114 Customer Ledger 121 Incoming Line and Metering Construction Design Document 131 Distribution Map Data System

Claims

1. An unmatched classification device used in a system that performs power flow calculations based on data registered in a first ledger and operates a power system based on the results of the power flow calculations, comprising: a model generation unit that performs machine learning using, as teacher data, an unmatched situation where data registered in the first ledger does not match data registered in a second ledger in which data of the same items as the data registered in the first ledger is registered, and a correspondence to the unmatched situation, and generates an unmatched classification model; an unmatched classification unit that uses the unmatched classification model to determine whether there is an unmatched situation between the data registered in the first ledger and the data registered in the second ledger, and classifies at least one of the content of the unmatched situation and the correspondence to the unmatched situation when there is an unmatched situation; an output unit that outputs the classification result of the unmatched classification unit.

2. The unmatched classification device according to claim 1, wherein the unmatched classification unit classifies the unmatched situation based on whether human correspondence is required.

3. The unmatched classification device according to claim 1, wherein the output from the output unit includes content prompting correspondence to the unmatched situation to a predetermined department or a predetermined person.

4. An unmatched classification method used in a system that performs power flow calculations based on data registered in a first ledger and operates a power system based on the results of the power flow calculations, comprising: a model generation step of performing machine learning using, as teacher data, an unmatched situation where data registered in the first ledger does not match data registered in a second ledger in which data of the same items as the data registered in the first ledger is registered, and a correspondence to the unmatched situation, and generating an unmatched classification model; an unmatched classification step of using the unmatched classification model to determine whether there is an unmatched situation between the data registered in the first ledger and the data registered in the second ledger, and classifying at least one of the content of the unmatched situation and the correspondence to the unmatched situation when there is an unmatched situation; an output step of outputting the classification result of the unmatched classification step.

Images