Repair result determination device and repair result determination method
The repair result determination device and method address the inefficiencies in evaluating air conditioner repairs by using data analysis and machine learning to assess repair correctness and improve fault diagnosis accuracy.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- DAIKIN INDUSTRIES LTD
- Filing Date
- 2025-12-23
- Publication Date
- 2026-07-02
Smart Images
Figure JP2025045104_02072026_PF_FP_ABST
Abstract
Description
Repair result determination device and repair result determination method
[0009] ,
[0008] ,
[0001] The present disclosure relates to a repair result determination device and a repair result determination method.
[0002] For example, a technique related to the failure diagnosis of an air conditioner to be diagnosed has been conventionally known (see Patent Document 1).
[0003] In Patent Document 1, multivariate analysis is applied to a plurality of state quantities of an air conditioner to be diagnosed, and a diagnosis regarding an abnormality of the diagnosis target is performed. More specifically, in Patent Document 1, the Mahalanobis distance of a plurality of state quantities related to the diagnosis target is calculated, and the presence or absence of an abnormality is diagnosed by comparing the Mahalanobis distance with a threshold value.
[0004] Japanese Patent No. 4265982
[0005] An air conditioner in which a failure is detected by failure diagnosis is repaired by an operator such as a service engineer. In addition, the operator records repair history information including the cause of the failure after the repair is completed. The cause of the failure included in the repair history information is the cause of the failure specified by the operator.
[0006] However, the cause of the failure recorded in the repair history information may be incorrect. Also, it was not easy to determine whether the repair performed by the operator was appropriate (correct repair) even by referring to the repair history information.
[0007] An object of the present disclosure is to provide a repair result determination device and a repair result determination method that enable an operator to more easily evaluate the repair performed on an air conditioner.
[0008] A first aspect of the present disclosure is a repair result determination device having a control unit, wherein the control unit acquires an actual measurement value of operation data of an air conditioner repaired by an operator, a normal value of the operation data, and information regarding the repair time, determines the timing of the repair from the information regarding the repair time, evaluates the correctness of the repair based on the actual measurement value and the normal value before and after the timing of the repair, and outputs the result of the evaluation.
[0009] According to a first aspect of this disclosure, a repair result determination device can be provided that makes it easier to evaluate repairs performed by an operator on an air conditioner.
[0010] A second aspect of this disclosure is a repair result determination device according to the first aspect, wherein the control unit obtains a notification from the worker's terminal that the repair has been completed.
[0011] According to a second aspect of this disclosure, the timing of the repair can be determined by obtaining a notification from the worker's terminal indicating that the repair has been completed.
[0012] A third aspect of the present disclosure is a repair result determination device according to the first or second aspect, wherein the control unit acquires identification information of an air conditioner repaired by the worker from the worker's terminal, and acquires measured values of the operating data of the air conditioner identified by the identification information and normal values of the operating data of the air conditioner.
[0013] According to a third aspect of this disclosure, by obtaining identification information of the air conditioner that the worker has repaired from the worker's terminal, it is possible to obtain measured values and normal values of the operating data of the air conditioner that the worker has repaired.
[0014] A fourth aspect of this disclosure is a repair result determination device according to any one of the first to third aspects, wherein the normal value of the operating data of the air conditioner is the initial value of the operating data of the air conditioner.
[0015] According to a fourth aspect of this disclosure, the initial value of the air conditioner's operating data can be set to the normal value of the air conditioner's operating data.
[0016] A fifth aspect of this disclosure is a repair result determination device according to any one of the first to fourth aspects, wherein the normal values of the operating data of the air conditioner are normal sensor values and control values predicted using a machine learning model of the operating data of the air conditioner.
[0017] According to a fifth aspect of this disclosure, normal sensor values and control values predicted using a machine learning model of air conditioner operating data can be used as normal values for air conditioner operating data.
[0018] A sixth aspect of this disclosure is a repair result determination device according to any one of the first to fifth aspects, wherein the control unit outputs the evaluation result linked with information about the worker who performed the repair.
[0019] According to a sixth aspect of this disclosure, the evaluation results can be linked and output with information about the worker who performed the repair.
[0020] A seventh aspect of this disclosure is a repair result determination device according to any one of the first to sixth aspects, wherein the control unit associates the repair history information of the repair determined to have been performed correctly by the operator with the measured values prior to the timing of the repair, and outputs this as training data for a machine learning model that performs fault diagnosis from the measured values of the operating data of the air conditioner.
[0021] According to the seventh aspect of this disclosure, the repair history information of repairs that the operator determined to have been performed correctly can be associated with the measured values prior to the repair, and output as training data for a machine learning model that performs fault diagnosis from the measured values of the air conditioner's operating data.
[0022] An eighth aspect of this disclosure is a repair result determination device according to any one of the first to sixth aspects, wherein the control unit associates the repair history information of the repair determined to have been performed correctly by the operator with the measured values prior to the timing of the repair, and outputs it as rule-based evaluation data for a fault diagnosis logic that performs fault diagnosis from the measured values of the operating data of the air conditioner.
[0023] According to the eighth aspect of this disclosure, the repair history information of repairs that the operator determined to have been performed correctly can be associated with the measured values prior to the timing of the repair, and output as rule-based evaluation data for a fault diagnosis logic that performs fault diagnosis from the measured values of the air conditioner's operating data.
[0024] A ninth aspect of the present disclosure is a repair result determination method performed by a repair result determination device having a control unit, wherein the control unit acquires measured values of operating data of an air conditioner repaired by a worker, normal values of the operating data, and information regarding the timing of the repair, determines the timing of the repair from the information regarding the timing of the repair, evaluates the correctness of the repair based on the measured values and normal values before and after the timing of the repair, and outputs the result of the evaluation.
[0025] According to the ninth aspect of this disclosure, a repair result determination method can be provided that makes it easier to evaluate repairs performed by an operator on an air conditioner.
[0026] This is a diagram showing the configuration of an example of the repair result determination system 1 according to this embodiment. This is a hardware configuration diagram of an example of the computer 500 according to this embodiment. This is a graph showing an example of the behavior of sensor values of an air conditioner 10 that has been repaired by a worker. This is a graph showing an example of the behavior of sensor values of an air conditioner 10 that has been repaired by a worker. This is a flowchart showing an example of the processing of the repair result determination device 30 according to this embodiment. This is a flowchart showing an example of the processing of step S14. This is a flowchart showing an example of the processing for quantitatively visualizing the worker's skills. This is an explanatory diagram of an example of the fault diagnosis process. This is a flowchart showing an example of the fault diagnosis process.
[0027] Next, embodiments of this disclosure will be described in detail.
[0028] <System Configuration> Figure 1 is a configuration diagram of an example of the repair result determination system 1 according to this embodiment. The repair result determination system 1 in Figure 1 includes an air conditioner 10, a worker terminal 20, and a repair result determination device 30. In the repair result determination system 1 in Figure 1, one air conditioner 10 and one worker terminal 20 are shown, but there may be multiple units.
[0029] The air conditioner 10, the worker terminal 20, and the repair result determination device 30 in Figure 1 are connected to each other via a communication network N. The communication network N is, for example, the Internet or a LAN (Local Area Network).
[0030] The air conditioner 10 is installed in a building. The air conditioner 10 is a multi-split air conditioner for buildings, an air conditioner for shops, an air conditioner for offices, or a room air conditioner. The air conditioner 10 is an example of a refrigeration cycle system.
[0031] The air conditioner 10 in Figure 1 is configured to have one or more sensors 12 and actuators 14. The sensors 12 are, for example, temperature sensors or pressure sensors, and output sensor values. The actuators 14 are, for example, solenoid valves or motors. The amount of operation of the actuators 14 is output as actuator values. For example, the sensor values and actuator values of the air conditioner 10 are an example of the operating data of the air conditioner 10. The air conditioner 10 is also repaired by an operator who operates the operator terminal 20.
[0032] The worker terminal 20 is an information processing terminal that receives operations from workers repairing the air conditioner 10. Workers repairing the air conditioner 10 possess the worker terminal 20. For example, the worker terminal 20 may be a smartphone or a tablet device.
[0033] The repair result determination device 30 evaluates the correctness of the repair performed by the worker based on the measured and normal values of the operating data of the air conditioner 10 before and after the repair performed by the worker. The repair result determination device 30 acquires the operating data of the air conditioner 10 used to evaluate the correctness of the repair performed by the worker from the air conditioner 10. The repair result determination device 30 may also acquire the operating data of the air conditioner 10 used to evaluate the correctness of the repair performed by the worker from a device other than the air conditioner 10. Further details of the repair result determination device 30 will be described later.
[0034] The repair result determination device 30 has a control unit 32. The control unit 32 is a hardware configuration that executes a program. The control unit 32 can be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field Programmable Gate Array), etc. For example, the repair result determination device 30 can perform various processes described later by having a CPU, one example of the control unit 32, execute a program.
[0035] The configuration of the repair result determination system 1 shown in Figure 1 is an example. The repair result determination device 30 may be implemented using one or more information processing devices. Alternatively, the repair result determination device 30 may be implemented as a cloud computing service.
[0036] In Figure 1, the control unit 32 of the repair result determination device 30 is shown as an example, but the air conditioner 10 may also have a control unit. The various processes that the control unit 32 of the repair result determination device 30 performs, as described later, may also be performed by the control unit of the air conditioner 10. The various processes that the control unit 32 of the repair result determination device 30 performs may also be performed in cooperation with the control unit of the air conditioner 10.
[0037] When the control unit of the air conditioner 10 performs the various processes described later, which are performed by the control unit 32 of the repair result determination device 30, it is possible to acquire and utilize more types of operating data at shorter intervals than when the processes are performed by the control unit 32 of the repair result determination device 30.
[0038] It goes without saying that the configuration of the repair result determination system 1 shown in Figure 1 can vary depending on the application and purpose.
[0039] <Hardware Configuration> The repair result determination device 30 in Figure 1 can be implemented, for example, by a computer 500 with the hardware configuration shown in Figure 2. The worker terminal 20 in Figure 1 may also be implemented by a computer 500 with the hardware configuration shown in Figure 2.
[0040] Figure 2 is a hardware configuration diagram of an example of a computer 500 according to this embodiment. The computer 500 includes an input device 501, a display device 502, an external interface 503, RAM (Random Access Memory) 504, ROM (Read Only Memory) 505, a CPU 506, a communication interface 507, and an HDD (Hard Disk Drive) 508, and these are all interconnected via bus B. The input device 501 and the display device 502 may be connected and used only when necessary.
[0041] The input device 501 includes a touch panel, operation keys and buttons, a keyboard and mouse, etc., used by the user to input various signals. The display device 502 consists of a display such as a liquid crystal or organic EL that displays the screen, and a speaker that outputs sound data such as voice and music. The communication interface 507 is an interface for the computer 500 to communicate data via a communication network.
[0042] Furthermore, the HDD 508 is an example of a non-volatile storage device that stores programs and data. The programs and data stored include the OS (Operating System), which is the basic software that controls the entire computer 500, and applications (hereinafter simply referred to as "apps") that provide various functions on the OS. Note that the computer 500 may use a drive device that uses flash memory as a storage medium (for example, a solid-state drive: SSD) instead of the HDD 508.
[0043] The external interface 503 is an interface to an external device. The external device is a recording medium 503a, for example. The computer 500 reads from and writes to the recording medium 503a via the external interface 503.
[0044] The recording medium 503a includes flexible disks, CDs (Compact Discs), DVDs (Digital Versatile Discs), SD (Secure Digital) memory cards, and USB (Universal Serial Bus) memory.
[0045] The ROM 505 is an example of a non-volatile semiconductor memory (storage device) that can retain programs and data even when the power is turned off. The ROM 505 stores programs and data such as the BIOS (Basic Input Output System), OS settings, and network settings that are executed when the computer 500 is started up. The RAM 504 is an example of a volatile semiconductor memory (storage device) that temporarily holds programs and data.
[0046] The CPU 506 is an arithmetic unit that realizes the control and functions of the entire computer 500 by reading programs and data from a storage device such as the ROM 505 or HDD 508 onto the RAM 504 and executing processing, and is an example of the control unit 32.
[0047] <Evaluation of the correctness of repair based on the measured value and normal value of the operating data of the air conditioner> Here, an example of evaluating the correctness of repair using the sensor values included in the operating data of the air conditioner 10 will be described. FIGS. 3A and 3B are graph diagrams showing an example of the behavior of the sensor values of the air conditioner 10 repaired by an operator. In the graph diagrams of FIGS. 3A and 3B, the sensor values are shown on the vertical axis and the passage of time is shown on the horizontal axis.
[0048] FIG. 3A shows the behavior of the sensor values when the repair of the air conditioner 10 performed by the operator is correct. FIG. 3B shows the behavior of the sensor values when the repair of the air conditioner 10 performed by the operator is incorrect. In FIGS. 3A and 3B, the behavior of the measured value and predicted value of the sensor value is shown. The predicted value of the sensor value is a prediction of the sensor value output by the sensor 12 during normal operation. The prediction of the sensor value output by the sensor 12 during normal operation is performed using a normal prediction model described later.
[0049] The initial stage shown in FIGS. 3A and 3B is a period when the measured value of the sensor value of the air conditioner 10 approximates the normal value, for example, the period when the air conditioner 10 is installed. The sensor value at the time when the air conditioner 10 is installed is an example of the initial value of the operating data of the air conditioner 10.
[0050] Furthermore, as shown in Figures 3A and 3B, the measured and predicted sensor values of the air conditioner 10 diverge over time. Figures 3A and 3B show an example of a repair performed in 2020.
[0051] Figure 3A shows an example where the air conditioner 10 was properly repaired by a worker. The behavior of the air conditioner 10 during a malfunction, where the measured and predicted sensor values were diverging before the repair, changed to normal behavior after the repair. On the other hand, Figure 3B shows an example where the air conditioner 10 was not properly repaired by a worker. The behavior of the air conditioner 10 during a malfunction, where the measured and predicted sensor values were diverging before the repair, did not change to normal behavior after the repair, but remained the same as the malfunction behavior before the repair.
[0052] Therefore, in this embodiment, the characteristics of the behavior of the measured and predicted sensor values when the air conditioner 10 is repaired correctly, as shown in Figure 3A, and the characteristics of the behavior of the sensor values when the air conditioner 10 is repaired incorrectly, as shown in Figure 3B, are used to evaluate whether the repair performed by the worker is correct or incorrect.
[0053] For example, based on the behavioral characteristics shown in Figure 3A, in this embodiment, a correct repair is evaluated when the initial sensor value and the measured sensor value after repair are similar, and the measured sensor value before repair and the measured sensor value after repair are different. The reason for confirming that the measured sensor value before repair and the measured sensor value after repair are different is that if the measured sensor value before repair and the measured sensor value after repair are not different, it does not mean that the initial sensor value and the measured sensor value after repair have changed from a different state to an approximate state due to the repair performed by the worker.
[0054] Furthermore, based on the behavioral characteristics shown in Figure 3B, in this embodiment, if there is a discrepancy between the initial sensor value and the measured sensor value after repair, it is evaluated that an incorrect repair was performed (or that a proper repair was not performed).
[0055] <Processing> The repair result determination device 30 according to this embodiment evaluates the repairs performed by the worker on the air conditioner 10 by the process shown in Figure 4. Figure 4 is a flowchart of an example of the processing of the repair result determination device 30 according to this embodiment.
[0056] In step S10, the control unit 32 of the repair result determination device 30 acquires the operating data of the air conditioner 10 that has been repaired by the worker, and information regarding the timing of the repair. The operating data of the air conditioner 10 that has been repaired by the worker includes both measured and normal operating data. The normal operating data is either a predicted value based on the normal prediction model described later, or the initial value (measured value) of the operating data of the air conditioner 10. Since the operating data is constantly changing depending on how the air conditioner 10 is used (state), it is desirable to use the predicted value of the operating data for the same state using the normal prediction model described later.
[0057] The process in step S10 may involve obtaining identification information of the air conditioner 10 that the worker repaired from the worker's terminal 20, thereby obtaining the measured values of the operating data of the air conditioner 10 identified by the identification information and the normal values of the operating data of the air conditioner 10. The information regarding the repair timing is information indicating when the worker repaired the air conditioner 10. The information regarding the repair timing may also be a notification from the worker's terminal 20 indicating that the repair has been completed.
[0058] In step S12, the control unit 32 determines the timing of the repair performed by the worker on the air conditioner 10 based on the information regarding the timing of the repair acquired in step S10.
[0059] In step S14, the control unit 32 evaluates whether the repair is correct or incorrect based on the measured and normal values of the operating data before and after the timing of the repair. The process in step S14 is carried out as shown in Figure 5, for example.
[0060] Figure 5 is a flowchart of an example of the process in step S14.
[0061] In step S30, the control unit 32 compares the initial operating data of the air conditioner 10 that has been repaired by the operator with the measured operating data after the repair. In step S30, the degree of deviation between the actual and predicted values of the initial operating data may be compared with the degree of deviation between the measured and predicted values of the operating data after the repair.
[0062] In step S32, the control unit 32 determines, based on the comparison results from step S30, whether the initial values of the operating data of the air conditioner 10 repaired by the worker are similar to the measured values of the operating data after the repair. The initial values of the operating data of the air conditioner 10 may be, for example, the measured values of the operating data for the first year after installation, which should be normal. The measured values of the operating data of the air conditioner 10 after the repair may be, for example, the measured values of the operating data for one month after the repair. An existing index indicating similarity may be used as the index used to determine whether or not they are similar. If the initial values of the operating data of the air conditioner 10 repaired by the worker are similar to the measured values of the operating data after the repair, the process proceeds to step S34.
[0063] In step S34, the control unit 32 compares the measured operating data of the air conditioner 10 before the repair with the measured operating data after the repair. The measured operating data of the air conditioner 10 before the repair may be, for example, the measured operating data from one month before the customer reported the malfunction of the air conditioner 10 until the repair. In step S32, the degree of deviation between the actual and predicted operating data before the repair may be compared with the degree of deviation between the measured and predicted operating data after the repair.
[0064] In step S36, the control unit 32 determines, based on the comparison in step S34, whether there is a discrepancy between the measured operating data before the repair of the air conditioner 10 repaired by the worker and the measured operating data after the repair. The indicator used to determine whether there is a discrepancy can be an existing indicator that shows the degree of discrepancy. If there is a discrepancy between the measured operating data before the repair of the air conditioner 10 repaired by the worker and the measured operating data after the repair, the process proceeds to step S38.
[0065] In step S38, the control unit 32 evaluates that the correct repair has been performed by the worker (that the repair has been done properly). As shown in the flowchart in Figure 5, the control unit 32 evaluates that the correct repair has been performed if the initial value of the operating data of the air conditioner 10 that has been repaired by the worker is similar to the measured value of the operating data after the repair, and there is a discrepancy between the measured value of the operating data of the air conditioner 10 before the repair and the measured value of the operating data after the repair.
[0066] Furthermore, if it is determined in step S32 that the initial operating data of the air conditioner 10 repaired by the worker and the measured operating data after the repair are not similar, the process proceeds to step S40. Also, if it is determined in step S36 that the measured operating data of the air conditioner 10 before the repair and the measured operating data after the repair are not significantly different, the process proceeds to step S40.
[0067] In step S40, the control unit 32 evaluates that an incorrect repair was performed by the worker (i.e., the proper repair was not performed). The control unit 32 evaluates that an incorrect repair was performed if the initial value of the operating data of the air conditioner 10 that was repaired by the worker is not similar to the measured value of the operating data after the repair, or if the measured value of the operating data of the air conditioner 10 before the repair is not significantly different from the measured value of the operating data after the repair.
[0068] Returning to Figure 4, in step S16, the control unit 32 outputs the evaluation of whether the repair was correct or incorrect in step S14, linked with information about the worker who performed the repair. Note that linking the evaluation of whether the repair was correct or incorrect in step S14 with information about the worker who performed the repair is a process that is performed as needed, but is not mandatory. If linking the evaluation of whether the repair was correct or incorrect in step S14 with information about the worker who performed the repair is not necessary, the control unit 32 outputs the evaluation of whether the repair was correct or incorrect in step S14. Information about the worker who performed the repair may be, for example, the worker's identification information or the identification information of the organization to which the worker belongs. The output in step S16 may be performed using existing technology that allows notification to the worker (for example, email function, application notification function, or printing function). Alternatively, the output in step S16 may be registered in a database.
[0069] According to the flowcharts shown in Figures 4 and 5, the correctness of repairs performed by the worker can be automatically evaluated. Therefore, according to this embodiment, it becomes easier to select correct repair history information from the repair history information recorded by the worker.
[0070] Operational data from correct repairs performed by operators can be used as training data for the normal prediction model described later. Furthermore, the history and operational data of correct repairs performed by operators can be used as training data for the machine learning model of the fault diagnosis logic described later. Additionally, the history and operational data of correct repairs performed by operators can be used as rule-based evaluation data for the fault diagnosis logic described later.
[0071] Furthermore, incorrect repair history information performed by operators may contain incorrect failure causes recorded by the operators, and therefore may not be suitable as training data for the machine learning model of the fault diagnosis logic described later, or as rule-based evaluation data for the fault diagnosis logic described later. In order to improve the diagnostic accuracy of the fault diagnosis logic described later, it is necessary to remove incorrect repair history information and operating data performed by operators from the training data or evaluation data. In this embodiment, correct repair history information and operating data performed by operators can be automatically selected, so that repair history information and operating data suitable as training data for the machine learning model of the fault diagnosis logic described later, or as rule-based evaluation data for the fault diagnosis logic described later, can be selected with less effort than if a human were to select them.
[0072] Furthermore, by using the results of the evaluation of the repairs performed by the worker, the worker's skills can be quantitatively visualized in this embodiment through the process shown in Figure 6. The worker's skills may be, for example, the success rate or failure rate of repairs. The worker's skills may also be the success rate or failure rate for each cause of failure.
[0073] Figure 6 is a flowchart of an example of a process for quantitatively visualizing the skills of a worker. The control unit 32 may perform the process shown in the flowchart of Figure 6 after the process shown in the flowcharts of Figures 4 and 5, or it may perform the process shown in the flowchart of Figure 6 according to the actions of the worker or other personnel.
[0074] In step S50, the control unit 32 identifies the worker who will output the results of the repair evaluation performed according to the flowcharts shown in Figures 4 and 5.
[0075] In step S52, the control unit 32 aggregates the evaluation of the correctness of repairs associated with the worker identified in step S50. The control unit 32 aggregates, for example, the correctness rate or failure rate of repairs as a skill. Alternatively, the control unit 32 may aggregate, for example, the correctness rate or failure rate for each cause of failure as a skill.
[0076] In step S53, the control unit 32 visualizes the success rate or failure rate of repairs as an indicator of the worker's skill, based on the aggregated results from step S52. The control unit 32 may also visualize the success rate or failure rate for each cause of failure as an indicator of the worker's skill, based on the aggregated results from step S52. The visualization of the worker's skill can be done using existing technologies that allow the worker to visually perceive the results.
[0077] While Figure 6 illustrates an example of visualizing a worker's skills, the skills of the organization to which the worker belongs can also be visualized. Organizational skills can be visualized, for example, by aggregating the evaluations of the accuracy of repairs performed by multiple workers belonging to that organization.
[0078] According to the flowchart in Figure 6, the control unit 32 can visualize the quality of work performed by the worker repairing the air conditioner 10. Similarly, the control unit 32 can visualize the quality of work performed by the organization to which the worker belongs.
[0079] The correct repair history information and operating data selected as described above can be used as training data for the normal prediction model 100 used in the fault diagnosis process shown in Figures 7 and 8, as training data for the machine learning model of the fault diagnosis logic 102, or as rule-based evaluation data for the fault diagnosis logic 102.
[0080] Figure 7 is an explanatory diagram of an example of the fault diagnosis process. Figure 8 is a flowchart of an example of the fault diagnosis process.
[0081] In this embodiment, the selection of correct repair history information and operating data performed by the operator becomes easier, making the selection of training data for the normal prediction model 100 simpler and more accurate. The normal prediction model 100 is a machine learning model that outputs predicted values of operating data for a normal air conditioner 10.
[0082] The normal prediction model 100 is machine-trained to output predicted values of the sensor value of sensor 12 or the actuator value of actuator 14 under normal conditions as predicted values of the operating data, using the history information of correct repairs performed by the operator and the operating data.
[0083] Furthermore, in this embodiment, the selection of correct repair history information and operating data performed by the operator becomes easier, making the selection of training data for the machine learning model of the fault diagnosis logic 102 simpler and more accurate. The machine learning model of the fault diagnosis logic 102 is trained to diagnose the cause of a fault based on the trend of the deviation between the predicted value, which is the normal value of the operating data, and the measured value of the operating data, using correct repair history information and operating data performed by the operator.
[0084] The judgment logic 104 in Figure 7 is a function to prevent false detection of faults. The judgment logic 104 may determine that a fault detection is not a false detection if the fault detection continues for a predetermined period (for example, 2 to 3 days).
[0085] In step S70, the control unit 32 acquires actual operating data of the air conditioner 10 to be diagnosed. The control unit 32 may acquire the operating data of the air conditioner 10 to be diagnosed in real time, or it may acquire data for a predetermined period all at once.
[0086] In step S72, the control unit 32 predicts the normal values of the operating data of the air conditioner 10, which is to be fault diagnosed, using the normal prediction model 100, and outputs them as predicted values.
[0087] In step S74, the control unit 32 calculates the degree of deviation between the predicted and actual operating data of the air conditioner 10 to be fault diagnosed. The degree of deviation is an index that represents the difference between the actual and predicted operating data. The degree of deviation is, for example, the difference between the actual and predicted operating data. The control unit 32 calculates multiple degree of deviations for each item, such as the degree of deviation of the subcooling degree or the degree of deviation of the discharge superheating degree (DSH).
[0088] In step S76, the control unit 32 uses a machine learning model or rule-based fault diagnosis logic 102 to diagnose the cause of the failure from the trends of multiple deviations. For example, the control unit 32 compares the multiple deviations calculated in step S74 with a threshold and diagnoses the cause of the failure from the trends of the multiple deviations that can be seen from the comparison result. The threshold should be set to a value that does not misdiagnose the cause of the failure. For example, the control unit 32 diagnoses the cause of the failure by combinations of items that show a deviation of a threshold or higher.
[0089] In step S78, the control unit 32 notifies a service technician or other worker of the cause of the fault diagnosed in step S76. The notification in step S78 can be made using existing technologies that enable notification to workers, such as email, application notification functions, or printing functions.
[0090] As described above, the repair result determination system 1 according to this embodiment makes it easy to select correct repair history information and operating data that can be used as training data for the normal prediction model 100 used in fault diagnosis processing, training data for the machine learning model of the fault diagnosis logic 102, or rule-based evaluation data for the fault diagnosis logic 102. Therefore, it becomes easier to construct a fault diagnosis system that performs fault diagnosis processing.
[0091] According to the repair result determination system 1 of this embodiment, it is possible to provide a repair result determination device and a repair result determination method that make it easier to evaluate the repairs performed on an air conditioner by a worker.
[0092] [Operation] This embodiment is a repair result determination device 30 having a control unit 32. The control unit 32 acquires the measured values of the operation data of the air conditioner 10 that was repaired by the worker, the normal values of the operation data, and information regarding the timing of the repair. The control unit 32 determines the timing of the repair from the information regarding the timing of the repair, evaluates the correctness of the repair based on the measured values and normal values before and after the timing of the repair, and outputs the evaluation result.
[0093] In this embodiment, the timing of repairs is determined from information regarding the timing of repairs, and the correctness of the repairs is evaluated based on the measured values before and after the timing of the repairs and the normal values. This makes it easier to evaluate and output the repairs performed by the worker on the air conditioner.
[0094] For example, the control unit 32 evaluates that a correct repair has been performed if the normal operating data and the measured operating data after the repair are similar, and there is a discrepancy between the measured operating data before the repair and the measured operating data after the repair. Conversely, the control unit 32 evaluates that an incorrect repair has been performed if there is a discrepancy between the normal operating data and the measured operating data after the repair.
[0095] Furthermore, the control unit 32 receives notification from the worker's terminal 20 that the repair has been completed.
[0096] In this embodiment, the timing of the repair can be determined by obtaining a notification from the worker's terminal 20 indicating that the repair has been completed.
[0097] Furthermore, the control unit 32 obtains identification information of the air conditioner 10 that the worker has repaired from the worker's terminal 20, and obtains the measured values and normal values of the operating data of the air conditioner 10 identified by the identification information.
[0098] In this embodiment, by obtaining identification information of the air conditioner 10 that the worker has repaired from the worker terminal 20, the measured and normal operating data of the air conditioner 10 that the worker has repaired can be obtained.
[0099] Furthermore, the normal operating data for the air conditioner 10 is the initial operating data for the air conditioner 10.
[0100] In this embodiment, the initial value of the operating data of the air conditioner 10 can be set to the normal value of the operating data of the air conditioner 10. The initial value of the operating data of the air conditioner 10 is the measured or predicted value of the operating data of the air conditioner 10 at a time when the measured value approximates the normal value, for example, the measured or predicted value of the operating data at the time the air conditioner 10 was installed.
[0101] Furthermore, the normal values for the operating data of the air conditioner 10 are the normal sensor values and control values predicted using a machine learning model of the operating data of the air conditioner 10.
[0102] In this embodiment, normal sensor values and control values predicted using a machine learning model of the operating data of the air conditioner 10 can be used as normal values for the operating data of the air conditioner 10.
[0103] Furthermore, the control unit 32 outputs the evaluation results linked with information about the worker who performed the repair.
[0104] In this embodiment, the correctness of the repair can be evaluated, and the evaluation results can be linked with information about the worker who performed the repair and output.
[0105] Furthermore, the control unit 32 associates the repair history information, which indicates that the repair was performed correctly by the operator, with the measured values prior to the repair, and outputs this as training data for a machine learning model that performs fault diagnosis from the measured values of the air conditioner 10's operating data.
[0106] In this embodiment, training data that associates repair history information, which indicates that a repair was performed correctly by an operator, with measured values prior to the repair, can be output as training data for a machine learning model that performs fault diagnosis from measured values of the air conditioner 10's operating data. Since training data that excludes repair history information, which indicates that an incorrect repair was performed, can be used, the machine learning model that performs fault diagnosis from measured values of the air conditioner 10's operating data can be expected to improve the inference accuracy of fault diagnosis.
[0107] Furthermore, the control unit 32 associates the repair history information, which indicates that the repair was performed correctly by the operator, with the measured values prior to the repair, and outputs this as rule-based evaluation data for the fault diagnosis logic that performs fault diagnosis from the measured values of the operating data of the air conditioner 10.
[0108] In this embodiment, evaluation data that associates repair history information of repairs determined to have been performed correctly by an operator with measured values prior to the timing of the repair can be output as rule-based evaluation data for fault diagnosis from measured values of the operating data of the air conditioner 10. Since evaluation data that excludes repair history information of repairs determined to have been performed incorrectly can be used, the rule-based fault diagnosis from measured values of the operating data of the air conditioner 10 can be expected to improve the inference accuracy of the fault diagnosis.
[0109] Furthermore, this embodiment is a repair result determination method executed by a repair result determination device 30 having a control unit 32. The control unit 32 acquires the measured values of the operation data of the air conditioner 10 that was repaired by the worker, the normal values of the operation data, and information regarding the timing of the repair. The control unit 32 determines the timing of the repair from the information regarding the timing of the repair, evaluates whether the repair was successful or not based on the measured values and normal values before and after the timing of the repair, and outputs the evaluation result.
[0110] In this embodiment, the timing of the repair is determined from information regarding the repair timing, and the correctness of the repair is evaluated based on the measured values before and after the repair timing and the normal values, thereby making it easier to evaluate and output the repairs performed by the worker on the air conditioner 10.
[0111] Although this embodiment has been described above, it will be understood that various modifications to the form and details are possible without departing from the spirit and scope of the claims. Although the present invention has been described above based on examples, the present invention is not limited to the above examples, and various modifications are possible within the scope described in the claims. This application claims priority to Basic Application No. 2024-231598 filed with the Japan Patent Office on December 27, 2024, the entire contents of which are incorporated herein by reference.
[0112] 1 Repair result determination system 10 Air conditioner 20 Worker terminal 30 Repair result determination device 32 Control unit
Claims
1. A repair result determination device having a control unit, wherein the control unit acquires measured values of operating data of an air conditioner repaired by a worker, normal values of the operating data, and information regarding the timing of the repair; determines the timing of the repair from the information regarding the timing of the repair; evaluates the correctness of the repair based on the measured values and normal values before and after the timing of the repair; and outputs the result of the evaluation.
2. The repair result determination device according to claim 1, wherein the control unit obtains a notification from the worker's terminal that the repair has been completed.
3. The repair result determination device according to claim 1 or 2, wherein the control unit obtains identification information of the air conditioner that the worker has repaired from the worker's terminal, and obtains measured values of the operating data of the air conditioner identified by the identification information and normal values of the operating data of the air conditioner.
4. The repair result determination device according to any one of claims 1 to 3, wherein the normal value of the operating data of the air conditioner is the initial value of the operating data of the air conditioner.
5. The repair result determination device according to any one of claims 1 to 4, wherein the normal values of the operating data of the air conditioner are the normal sensor values and control values predicted using a machine learning model of the operating data of the air conditioner.
6. The repair result determination device according to any one of claims 1 to 5, wherein the control unit outputs the evaluation result linked with information about the worker who performed the repair.
7. The repair result determination device according to any one of claims 1 to 6, wherein the control unit associates the repair history information, which the operator has determined to have performed the repair correctly, with the measured values prior to the timing of the repair, and outputs this as training data for a machine learning model that performs fault diagnosis from the measured values of the operating data of the air conditioner.
8. The repair result determination device according to any one of claims 1 to 6, wherein the control unit associates the repair history information, which the operator has determined to have performed the repair correctly, with the measured values prior to the timing of the repair, and outputs it as rule-based evaluation data for a fault diagnosis logic that performs fault diagnosis from the measured values of the operating data of the air conditioner.
9. A repair result determination method performed by a repair result determination device having a control unit, wherein the control unit acquires measured values of the operation data of an air conditioner repaired by a worker, normal values of the operation data, and information regarding the timing of the repair; determines the timing of the repair from the information regarding the timing of the repair; evaluates the correctness of the repair based on the measured values and normal values before and after the timing of the repair; and outputs the result of the evaluation.