Method and device for diagnosing manhole pumps

Abstract

To provide a diagnostic method and a diagnostic device for a manhole pump, which can reduce a burden imposed on a manager in a case of erroneous diagnosis while automatically diagnosing.SOLUTION: A manhole pump diagnostic method includes: a diagnosis step of automatically diagnosing an operating status at a predetermined diagnostic interval and outputting a diagnostic result; a notification step of sending the diagnosis result to a specified notification destination when an abnormality is diagnosed; and a determination step of determining whether or not to send the diagnosis result diagnosed as abnormal to the notification destination. The diagnosis step outputs, when it is diagnosed as abnormal, notification determination values of at least two high and low stages to the determination step according to the level of the abnormality, and the determination step executes filter processing to control the frequency of transmission by the notification step based on the notification determination values output from the diagnosis step at a predetermined number of diagnosis intervals in the recent past every time the notification determination values are output.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a diagnostic method and a diagnostic device for two manhole pumps installed in a manhole and repeatedly starting and stopping.

Background Art

[0002] Manhole pump equipment includes a water storage section for storing sewage flowing in from an inflow pipe, a plurality of pumps for draining the sewage stored in the water storage section to an outflow pipe, a water level gauge for measuring the water level of the sewage stored in the water storage section, and a control device that, when the water level measured by the water level gauge reaches the pump start-up water level, starts any one of the pumps to drain the sewage to the outflow pipe, and when the water level reaches the pump stop water level, stops the pump, and includes a control panel equipped with a sewage conveyance control device that executes the control.

[0003] Such manhole pump equipment usually has two pumps installed, and the control device is configured to operate these pumps alternately each time sewage is conveyed.

[0004] Patent Document 1 proposes a monitoring system for a pump station, which includes means for managing, as time-series data, the value obtained by dividing the integrated inflow volume per predetermined time to be discharged by a plurality of pumps in the pump station by the operation time within the predetermined time of each pump as the average drainage capacity of each pump, means for determining whether or not the average drainage capacity of each pump managed as the time-series data exceeds a preset drainage capacity range, and means for issuing an alarm when the average drainage capacity of each of the plurality of pumps exceeds the set drainage capacity range.

[0005] Patent Document 2 discloses a control device for a submersible pump that includes a pressure control unit that activates an electromagnetic switch to drive a submersible pump when the stored water level detected by a water level sensor reaches the pump activation water level, an abnormality determination unit that determines whether there is an abnormality based on the operating state of the electromagnetic switch that drives the electric motor, the detection state of a current sensor that detects the current of a power supply line connected to the armature winding via the electromagnetic switch, and the stored water level, wherein the abnormality determination unit determines that there is an overheating abnormality of the electric motor with auto-cut activated if no current is detected by the current sensor while the electromagnetic switch is operating and no decrease in the stored water level is detected by the water level sensor.

[0006] The control devices installed in conventional manhole pump systems, as described above, determined whether each pump was malfunctioning based on whether physical quantities such as the pump's drive current, the pump's operating time from the starting water level to the stopping water level, and the pump's temperature exceeded predetermined thresholds. However, measuring the appropriate physical quantities for each anticipated type of malfunction required using various sensors, resulting in a highly cumbersome process.

[0007] Furthermore, setting a uniform threshold for detecting abnormalities was difficult because it depended on the environment of the manhole where the pump was installed. For example, in areas with high and low water inflow per unit time, there would be biases in the pump's startup frequency and operating time, making it difficult to make accurate judgments using a fixed threshold.

[0008] In particular, in a remote monitoring system where each manhole pump facility's control panel is equipped with a communication device, and a server manages the pump operation data transmitted from each communication device, allowing administrators to access the server from their terminals to monitor the operating status, a very large amount of physical data needs to be transmitted to individually determine various abnormalities in the pumps installed in each manhole pump facility. This leads to an increase in the number of sensors and the volume of transmitted data, while simultaneously making it difficult to set appropriate thresholds.

[0009] Therefore, Patent Document 3 proposes a method for diagnosing manhole pumps that can accurately diagnose abnormalities caused by various factors even with small physical quantities. This manhole pump diagnostic method is for diagnosing two manhole pumps installed in a manhole that are repeatedly started and stopped, and comprises: a sampling step of sampling the operating time per start of each manhole pump within a predetermined diagnostic target period and calculating the average operating time of each as a pair of characteristic values; a normalization step of normalizing the pair of characteristic values ​​calculated in the sampling step; and a diagnostic step of plotting the normalized pair of characteristic values ​​on a two-dimensional coordinate system with one characteristic value as the x-axis and the other as the y-axis, and automatically diagnosing whether each manhole pump is normal or abnormal based on which side of a boundary threshold predetermined in the two-dimensional coordinate system the plotted characteristic value points lie on using a machine learning device. [Prior art documents] [Patent Documents]

[0010] [Patent Document 1] Japanese Patent Publication No. 2001-34338 [Patent Document 2] Japanese Patent Publication No. 2010-236191 [Patent Document 3] Japanese Patent Publication No. 2020-107073 [Overview of the Initiative] [Problems that the invention aims to solve]

[0011] The manhole pump diagnostic method disclosed in Patent Document 3 is configured to automatically diagnose a manhole pump as normal if the average operating time of each manhole pump during a predetermined diagnostic period is normalized to a pair of characteristic values, and a boundary threshold indicating the normal range is predetermined by a machine learning device and plotted on a two-dimensional coordinate system. If the pair of characteristic values ​​fall within the boundary threshold, it is considered normal, and if they deviate from the boundary threshold, it is considered abnormal.

[0012] The boundary thresholds determined by the machine learning device are obtained by learning the normal characteristic values ​​of many manhole pumps installed in various environments as training data. By normalizing the characteristic values ​​of the manhole pump to be diagnosed based only on statistical data derived from the past characteristic values ​​of that manhole pump, it is possible to achieve appropriate diagnosis using a limited number of characteristic values ​​while reducing the influence of differences in installation environment and other factors.

[0013] However, even with such manhole pump diagnostic methods, accuracy is not perfect, and there may be slight discrepancies between the actual operating condition of the manhole pump and the diagnostic results. For example, a manhole pump that is actually operating normally may be diagnosed as abnormal, or conversely, a manhole pump that is actually abnormal may be diagnosed as normal.

[0014] In such cases, if the system is configured to automatically send the diagnostic results uniformly to a designated recipient, such as the administrator's mobile device, there is a risk that the burden on administrators may not be sufficiently reduced. For example, an administrator who receives an abnormality notification might go to the actual installation site of the manhole pump device and find that it is functioning normally, resulting in a false alarm. Conversely, an abnormality might actually occur in a manhole pump device that did not receive an abnormality notification, leading to a missed detection.

[0015] In view of the above-mentioned problems, the object of the present invention is to provide a manhole pump diagnostic method and diagnostic device that can perform automatic diagnosis while reducing the burden on the administrator in the event of a misdiagnosis. [Means for solving the problem]

[0016] To achieve the above objective, the first characteristic configuration of the manhole pump diagnostic method according to the present invention is a manhole pump diagnostic method comprising: a diagnostic step that automatically diagnoses whether the operating state of a manhole pump installed in a manhole is normal or abnormal at predetermined diagnostic intervals and outputs a diagnostic result; and a notification step that, if the diagnostic step diagnoses an abnormality, transmits the diagnostic result to a predetermined notification destination, further comprising a determination step that determines whether or not to transmit the diagnostic result, which was diagnosed as abnormal in the diagnostic step, to the notification destination by the notification step, and when the diagnostic step diagnoses an abnormality, outputs a notification determination value of at least two levels, high and low, according to the level of the abnormality to the determination step, and each time the notification determination value is output, the determination step checks the output from the diagnostic step for the most recent predetermined number of diagnostic intervals in the past The aforementioned level is set The aforementioned notification determination value Probability of occurrence Based on this, a filtering process is performed to control the transmission frequency of the notification step, and the notification step determines whether or not to transmit the diagnostic result according to the result of the filtering process.

[0017] When an abnormality is diagnosed in the diagnostic step, the determination step determines in advance whether or not to send a notification to the recipient in the notification step. When an abnormality is diagnosed in the diagnostic step, a notification determination value of at least two levels, high and low, is output to the determination step, depending on the level of the abnormality. Each time a notification determination value is output in the determination step, a filtering process is performed to control the frequency of transmission by the notification step based on the notification determination values ​​output from the diagnostic step in the most recent predetermined number of diagnostic intervals, thereby reducing the frequency of "false" notifications.

[0018] The second characteristic configuration is, in addition to the first characteristic configuration described above, that the filtering process is based on whether the occurrence probability of the notification determination value output by the diagnosis step at the first predetermined number of times (N1) of diagnosis intervals in the immediate past has reached the first reference probability (P1), and whether the occurrence probability of the high level of the notification determination value output by the diagnosis step at the second predetermined number of times (N2) of diagnosis intervals in the immediate past has reached the second reference probability (P2), to control the transmission frequency in the notification step.

[0019] When an abnormality is diagnosed in the diagnosis step, the transmission frequency is appropriately controlled based on the occurrence probability of the notification determination value at any level (high or low) in the immediate past and the occurrence probability of the high level, and as a result, the lower the level of the abnormality, the lower the transmission frequency to the notification destination, and the frequency of "idle" notifications is reduced.

[0020] The third characteristic configuration is, in addition to the second characteristic configuration described above, that the filtering process allows the transmission of the diagnosis result when the occurrence probability of the notification determination value output by the diagnosis step at at least the first predetermined number of times (N1) of diagnosis intervals is higher than the first reference probability (P1).

[0021] When the occurrence probability of the notification determination value at any level (high or low) output by the diagnosis step at the first predetermined number of times (N1) of diagnosis intervals is higher than the first reference probability (P1), the transmission of the diagnosis result is allowed on the assumption that it is a highly reliable diagnosis result.

[0022] The fourth characteristic configuration is, in addition to the second characteristic configuration described above, that the filtering process allows the transmission of the diagnosis result when the occurrence probability of the high level of the notification determination value output by the diagnosis step at at least the second predetermined number of times (N2) of diagnosis intervals is higher than the second reference probability (P2).

[0023] When the occurrence probability of the high level of the notification determination value output by the diagnosis step at the second predetermined number of times (N2) of diagnosis intervals is higher than the second reference probability (P2), the transmission of the diagnosis result is allowed on the assumption that it is a highly urgent abnormality.

[0024] The fifth characteristic configuration is that, in addition to any of the second to fourth characteristic configurations described above, in the filter processing, when the occurrence probability of the notification determination value output by the diagnosis step at the first predetermined number of times (N1) of diagnosis intervals is not more than the first reference probability (P1), and the occurrence probability of the high level of the notification determination value output by the diagnosis step at the second predetermined number of times (N2) of diagnosis intervals is not more than the second reference probability (P2), the transmission of the diagnosis result is prohibited.

[0025] By prohibiting the transmission of the diagnosis result when the occurrence probability of the notification determination value output by the diagnosis step at the first predetermined number of times (N1) of diagnosis intervals is not more than the first reference probability (P1), and the occurrence probability of the high level of the notification determination value output by the diagnosis step at the second predetermined number of times (N2) of diagnosis intervals is not more than the second reference probability (P2), the occurrence of "false swing" notifications is effectively reduced.

[0026] The sixth characteristic configuration is that, in addition to any of the second to four ... characteristic configurations described above, each value of the first predetermined number of times (N1), the first reference probability (P1), the second predetermined number of times (N2), and the second reference probability (P2) is configured to be variably settable.

[0027] By configuring these values to be variably settable, it is possible to more effectively reduce the frequency of "false swing" notifications according to the characteristics of individual manhole devices. For example, when an administrator dislikes "oversight", it is also possible to adjust to reduce the frequency of "oversight".

[0028] The seventh characteristic configuration is that, in addition to the first to fourIn addition to any of the above feature configurations, the diagnostic step is configured to diagnose the state of the manhole pump during the diagnostic period based on diagnostic indicators calculated from the operating information of the manhole pump collected during the diagnostic period. The diagnostic indicators are normalized based on statistical data calculated only from each diagnostic indicator calculated in the most recent predetermined period including the diagnostic period, plotted in a predetermined coordinate system, and the normality or abnormality of the manhole pump is diagnosed based on which side of the boundary threshold predetermined in the coordinate system the plotted diagnostic indicator points lie on. The boundary threshold is automatically generated by a machine learning device that learns using only diagnostic indicators for normal operation obtained from multiple manholes, including the manhole pump installed in the manhole subject to diagnosis, and indicates the boundary of the coordinate system, which is a feature space containing normal diagnostic indicators.

[0029] For example, even when the values ​​of diagnostic indicators fluctuate significantly depending on the manhole installation environment, such as areas with high and low water inflow, the influence of external factors such as the installation environment can be eliminated by normalizing the diagnostic indicators during the diagnostic period based on statistical data calculated only from individual diagnostic indicators calculated during the most recent predetermined period, including the diagnostic period. Furthermore, by plotting the normalized diagnostic indicators in a predetermined coordinate system and adopting a boundary threshold for diagnosing normal or abnormal conditions, obtained by machine learning using only diagnostic indicators obtained during normal operation from manhole pumps installed in multiple manholes with different installation environments as training data, it becomes possible to perform appropriate diagnoses that are not affected by the installation environment, without having to prepare diagnostic indicators for abnormal conditions in advance.

[0030] The eighth characteristic configuration is that, in addition to the seventh characteristic configuration described above, the diagnostic indicator includes the ratio of the number of times one manhole pump was operated to the number of times the other manhole pump was operated during the diagnostic period, and the ratio of the operating time of one manhole pump to the operating time of the other manhole pump during the diagnostic period.

[0031] By using the ratio of the number of times each manhole pump is operated and the ratio of the operating time each manhole pump is operated as diagnostic indicators, it is possible to effectively diagnose abnormalities with a limited number of diagnostic indicators.

[0032] The first characteristic configuration of the manhole pump diagnostic device according to the present invention is a manhole pump diagnostic device comprising: a diagnostic processing unit that automatically diagnoses whether the operating state of a manhole pump installed in a manhole is normal or abnormal at predetermined diagnostic intervals and outputs a diagnostic result; and a notification processing unit that transmits the diagnostic result to a predetermined notification destination when the diagnostic processing unit diagnoses an abnormality, further comprising a determination processing unit that determines whether or not to transmit the diagnostic result diagnosed as abnormal by the diagnostic processing unit to the notification destination, the diagnostic processing unit outputs a notification determination value of at least two levels, high and low, according to the level of the abnormality when it diagnoses an abnormality, the determination processing unit, each time the notification determination value is output, checks the output from the diagnostic processing unit for the most recent predetermined number of diagnostic intervals in the past. The aforementioned level is set The aforementioned notification determination value Probability of occurrence Based on this, the notification processing unit performs filtering to control the transmission frequency, and the notification processing unit decides whether or not to transmit the diagnostic result according to the result of the filtering.

[0033] The second characteristic configuration is that, in addition to the first characteristic configuration described above, the filtering process controls the transmission frequency by the notification processing unit based on whether the probability of occurrence of the notification judgment value output by the diagnostic processing unit in the most recent past first predetermined number of diagnostic intervals (N1) has reached a first reference probability (P1), and whether the probability of occurrence of a high level of the notification judgment value output by the diagnostic processing unit in the most recent past second predetermined number of diagnostic intervals (N2) has reached a second reference probability (P2).

[0034] The third characteristic configuration is that, in addition to the second characteristic configuration described above, the filtering process allows the transmission of the diagnostic result if the probability of the notification judgment value output by the diagnostic processing unit at least one predetermined number of diagnostic intervals (N1) is higher than the first reference probability (P1).

[0035] The fourth characteristic configuration is that, in addition to the second characteristic configuration described above, the filtering process allows the transmission of the diagnostic result if the probability of a high level occurring in the notification judgment value output by the diagnostic processing unit at least for the second predetermined number of diagnostic intervals (N2) is higher than the second reference probability (P2).

[0036] The fifth characteristic configuration is that, in addition to any of the second to fourth characteristic configurations described above, the filtering process prohibits the transmission of the diagnostic results when the probability of the notification judgment value output by the diagnostic processing unit at a first predetermined number of diagnostic intervals (N1) is less than or equal to the first reference probability (P1), and the probability of a high level of the notification judgment value output by the diagnostic processing unit at a second predetermined number of diagnostic intervals (N2) is less than or equal to the second reference probability (P2).

[0037] The sixth characteristic configuration is as described above, from the second to the third four In addition to any of the above characteristic configurations, the first predetermined number of times (N1), the first reference probability (P1), the second predetermined number of times (N2), and the second reference probability (P2) are configured to be variably set.

[0038] The seventh characteristic configuration is as described above from the first to the second fourIn addition to any of the above feature configurations, the diagnostic processing unit is configured to diagnose the state of the manhole pump during the diagnostic period based on diagnostic indicators calculated from the operating information of the manhole pump collected during the diagnostic period. It plots normalized diagnostic indicators, calculated based on statistical data obtained only from each diagnostic indicator calculated in the most recent predetermined period including the diagnostic period, in a predetermined coordinate system, and diagnoses whether the manhole pump is normal or abnormal based on which side of the boundary threshold predetermined in the coordinate system the plotted diagnostic indicator points lie on. The boundary threshold is automatically generated by a machine learning device that learns using only diagnostic indicators for normal operation obtained from multiple manholes, including the manhole pump installed in the manhole subject to diagnosis, and indicates the boundary of the coordinate system, which is a feature space containing normal diagnostic indicators.

[0039] The eighth characteristic configuration is that, in addition to the seventh characteristic configuration described above, the diagnostic indicator includes the ratio of the number of times one manhole pump was operated to the number of times the other manhole pump was operated during the diagnostic period, and the ratio of the operating time of one manhole pump to the operating time of the other manhole pump during the diagnostic period. [Effects of the Invention]

[0040] As described above, the present invention provides a manhole pump diagnostic method and diagnostic device that can perform automatic diagnosis while reducing the burden on administrators due to misdiagnosis. [Brief explanation of the drawing]

[0041] [Figure 1] Diagram illustrating a manhole pump. [Figure 2] Diagram illustrating the diagnostic device for manhole pumps. [Figure 3] Diagram illustrating the procedure for diagnosing a manhole pump. [Figure 4] (a) is an explanatory diagram of the measurement data, and (b) is an enlarged explanatory diagram of the key parts of the measurement data. [Figure 5] Diagram illustrating the normalization process. [Figure 6] (a) is an explanatory diagram of the diagnosis, and (b) is an explanatory diagram showing the values ​​of the diagnostic indicators before and after the correction process when the difference in the number of driving cycles is 1. [Figure 7] Diagram explaining the diagnostic map [Figure 8] (a) is an explanatory diagram of the simulation results of the characteristics of missed and missed detections that occur when the first predetermined number of detections (N1), first reference probability (P1), second predetermined number of detections (N2), and second reference probability (P2) are changed based on the diagnostic results of a certain manhole device, and (b) is a graph showing the simulation results. [Figure 9] Diagram illustrating other steps in diagnosing a manhole pump. [Figure 10] Diagram illustrating an alternative embodiment of the notification process. [Modes for carrying out the invention]

[0042] The following describes the diagnostic method and diagnostic apparatus for manhole pumps according to the present invention.

[0043] Figure 1 shows the manhole pump device 10. The manhole pump device 10 includes a manhole 12 which serves as a reservoir for storing wastewater flowing in from the upstream wastewater inlet pipe 11, two pumps PA and PB which pump the wastewater stored in the manhole 12 to the downstream wastewater outlet pipe 13, and water level gauges 18 and 19 which measure the water level of the wastewater stored in the manhole 12.

[0044] The discharge bend 15a of the first pump PA is flange-connected to the first pumping pipe 15b, the first bend 15c, and the first horizontal pipe 15d, respectively, and the first horizontal pipe 15d is flange-connected to the sewage outflow pipe 13 via the header pipe 13a. A check valve 15e is provided between the first pumping pipe 15b and the first bend 15c.

[0045] The discharge bend 17a of the second pump PB is flange-connected to the second pumping pipe 17b and the second bend 17c, respectively, and the second bend 17c is flange-connected to the sewage outflow pipe 13 via the header pipe 13a. A check valve 17e is provided between the second pumping pipe 17b and the second bend 17c.

[0046] A submersible pressure type or bubble type water level gauge 18 is installed at the bottom of the manhole 12. This water level gauge 18 continuously detects the water level of the sewage stored in the manhole 12. In addition, a float type water level gauge 19 is installed as a backup water level gauge to detect abnormally high water levels (HHWL).

[0047] Near the manhole 12, a control panel device 200 is installed, which houses a control panel 20 containing a control unit 21 that performs wastewater transport control, which controls pumps PA and PB to pump the wastewater accumulated in the manhole 12 to the wastewater outflow pipe 13.

[0048] The control panel 20 is equipped with a control unit 21, a storage unit 22, and a communication unit 24. The storage unit 22 stores control information from the control unit 21, water level information from water level gauges 18 and 19, and so on. The communication unit 24 includes a transmission unit that transmits various information stored in the storage unit 22 to a remote monitoring device 40, and a reception unit that receives control commands from the monitoring device 40.

[0049] A wireless communication medium, such as a mobile phone network, is preferably used as the communication medium connecting the communication unit 24 and the monitoring device 40. The monitoring device 40 and the communication unit 24 are connected to the Internet via such a communication medium, and the monitoring device 40 is also configured to be able to connect to the Internet via the wireless communication medium to a mobile communication terminal 30 owned by the manager of the manhole pump device 10.

[0050] The control panel 20 and each pump PA and PB are connected by AC power supply lines L1 and L2, and the control panel 20 and water level gauges 18 and 19 are connected by signal line S.

[0051] When the control unit 21 detects that the water level measured by the water level gauge 18 has reached a predetermined pump start water level HWL, it controls the power supply from the power supply line L1 to start one of the pumps PA and PB, pump PA. When it detects that the water level has reached a pump stop water level LWL which is lower than the pump start water level HWL, it stops the power supply and stops the pump PA.

[0052] The control unit 21 then detects that the water level has reached the pump start water level HWL again and controls the power supply from the power line L2 to start the other pump PB. When it detects that the water level has reached the pump stop water level LWL, it stops the power supply and stops the pump PB. In other words, the control unit 21 alternately controls the operation of pumps PA and PB.

[0053] Furthermore, if the control unit 21 cannot detect the pump activation water level HWL due to a malfunction of the water level gauge 18, or if it detects that a large amount of rainwater exceeding the drainage capacity of one pump has flowed into the manhole 12 due to torrential rain and has reached an abnormally high water level HHWL as measured by the water level gauge 19, the control unit 21 will operate both pumps PA and PB simultaneously.

[0054] The control unit 21 samples water level information detected by water level gauges 18 and 19 in a time series, for example at 1-minute intervals, and stores it in the storage unit 22. It also stores in the storage unit 22 time series operating information such as the start and stop times of each pump PA and PB, and the operating time from start to stop.

[0055] As shown in Figure 2, the communication unit 24 provided in the control panel 20 of each manhole pump device 10 is configured to transmit water level information and operating information stored in the storage unit 22 to the monitoring device 40 at predetermined intervals.

[0056] The monitoring device 40 functions as a diagnostic device for multiple manhole pump devices 10 and can be configured, for example, as a server computer. The monitoring device 40 includes a communication unit 41 that communicates with the communication unit 24 of each manhole pump device 10 and the administrator's mobile communication terminal 30, a database DB that stores water level information and operation information transmitted from the communication unit 24 of each manhole pump device 10, a data processing unit 42 that exchanges data with the database DB, and a diagnostic unit 44 that diagnoses whether each manhole pump device 10 is operating normally based on the water level information and / or operation information stored in the database DB.

[0057] The diagnostic unit 44 includes a diagnostic index calculation processing unit 45, a diagnostic processing unit 46, a determination processing unit 47, and a notification processing unit 48. The diagnostic index calculation processing unit 45 calculates diagnostic indexes for each manhole pump device 10 from the operating information of the manhole pumps PA and PB collected during the diagnostic period. The diagnostic processing unit 46 automatically diagnoses whether the operating status of the manhole pumps PA and PB during the diagnostic period is normal or abnormal based on the diagnostic index calculated by the diagnostic index calculation processing unit 45, at predetermined diagnostic intervals, and outputs the diagnostic results.

[0058] When the diagnostic processing unit 46 detects an abnormality, the notification processing unit 48 sends the diagnostic result to a designated notification recipient, such as a pre-configured administrator. The determination processing unit 47 determines whether or not to send the diagnostic result, which the diagnostic processing unit 46 has identified as an abnormality, to the notification processing unit 48.

[0059] As diagnostic indicators, the ratio n1 / n2 of the number of times one manhole pump PA was operated n1 to the number of times the other manhole pump PB was operated n2 during the diagnostic period, and the ratio t1 / t2 of the operating time t1 of one manhole pump PA to the operating time t2 of the other manhole pump PB during the diagnostic period are used.

[0060] Before calculating the ratio of operating times n1 / n2, the diagnostic index calculation processing unit 45 adds 1 to the operating time of the manhole pump with fewer operating times, or subtracts 1 from the operating time of the manhole pump with more operating times, if the difference between the operating time n1 of one manhole pump PA and the operating time n2 of the other manhole pump PB during the diagnostic period is 1.

[0061] Under normal conditions, the two manhole pumps PA and PB operate alternately. As the number of times each manhole pump PA and PB operate during the diagnostic period increases, the ratio of the number of operations n1 / n2 converges to 1. However, if the number of times each manhole pump PA and PB operate during the diagnostic period is low, the ratio of the number of operations n1 / n2 will deviate significantly from 1, potentially leading to a misdiagnosis.

[0062] Therefore, if the difference between the number of operations n1 of one manhole pump PA and the number of operations n2 of the other manhole pump PB is 1, it is presumed that both manhole pumps PA and PB are operating alternately appropriately. In this case, 1 is added to the number of operations of the manhole pump with fewer operations, or 1 is subtracted from the number of operations of the manhole pump with more operations, to correct the ratio of operations to 1. This correction prevents misdiagnosis even if the number of operations of manhole pumps PA and PB is low during the diagnostic period.

[0063] The diagnostic index calculation processing unit 45 adds 1 to the number of operations for manhole pumps with few operations, or subtracts 1 from the number of operations for manhole pumps with many operations, and adopts the average operating time per start during the diagnostic period as the operating time per start for the manhole pump corresponding to the addition or subtraction.

[0064] To maintain a balance with the corrected number of operations, the operating time per start of the manhole pump, which has been corrected by increasing or decreasing the number of operations, is increased or decreased. By adopting the average operating time per start during the diagnostic period as the correction value, an appropriate diagnosis can be made.

[0065] Furthermore, the diagnostic index calculation processing unit 45 includes a normalization processing unit that normalizes the diagnostic index during the diagnostic period based on statistical data calculated only from each diagnostic index calculated during the most recent predetermined period including the diagnostic period.

[0066] For example, even if the values ​​of the diagnostic indicators vary greatly depending on the installation environment of the manhole pump device 10, such as whether the volume of inflowing water is high or low, the influence of external factors such as the installation environment can be eliminated by normalizing the diagnostic indicators during the diagnostic period based on statistical data calculated only from individual diagnostic indicators calculated during the most recent predetermined period, including the diagnostic period.

[0067] The diagnostic processing unit 46 plots normalized diagnostic indicators in a predetermined coordinate system, in this embodiment a two-dimensional coordinate system of x and y, and automatically diagnoses whether the manhole pumps PA and PB are normal or abnormal based on which side of the boundary thresholds pre-set in the two-dimensional coordinate system the plotted diagnostic indicator points lie on.

[0068] The boundary threshold is automatically generated by a machine learning device that learns using only diagnostic indicators for normal operation obtained from multiple manhole pumps PA and PB installed in multiple manholes 12, including the manhole pumps PA and PB installed in the manhole 12 being diagnosed, as training data. It indicates the boundary of a dimensional coordinate system that becomes a feature space containing normal diagnostic indicators. This machine learning device is built into the diagnostic processing unit 46 and consists of functional blocks that execute a one-class support vector machine algorithm.

[0069] By plotting normalized diagnostic indicators in a predetermined coordinate system and adopting a boundary threshold for diagnosing normal or abnormal conditions, obtained by machine learning using only diagnostic indicators obtained during normal operation from manhole pumps installed in multiple manholes with different installation environments as training data, it becomes possible to perform appropriate diagnoses that are not affected by the installation environment, without having to prepare diagnostic indicators for abnormal conditions in advance.

[0070] The area outside the boundary threshold set in the two-dimensional coordinate system is divided into multiple regions, and each region has a diagnostic map associated with one of the causes of anomalies. The diagnostic processing unit 46 is configured to diagnose the cause of the anomaly based on the region where each characteristic value point is plotted.

[0071] Figure 3 shows the flow of a series of diagnostic processes performed by the monitoring device 40. When the data processing unit 42 has finished receiving measurement data, which is the operating information for one day set as the diagnostic period, from each manhole pump device 10 and storing it in the database DB (SA1), the diagnostic index calculation processing unit 45 extracts the number of operations n1 and n2 for pumps PA and PB respectively as measurement data for each pump of each manhole pump device 10 to calculate diagnostic indexes used for abnormality determination, and also extracts the total operating time t1 and t2 for the pumps for one day, and calculates the ratio of the number of operations n1 / n2 and the ratio of the operating time t1 / t2 as diagnostic indexes (SA2).

[0072] In step SA2, before calculating the ratio of operating times n1 / n2, if the difference between the operating times n1 of one manhole pump and the operating times n2 of the other manhole pump during the diagnostic period is 1, then 1 is added to the operating times of the manhole pump with fewer operating times, or 1 is subtracted from the operating times of the manhole pump with more operating times. If the difference between the operating times n1 and n2 is not 1, that is, if they are equal or 2 or more, then this correction process is not performed.

[0073] Furthermore, when adding 1 to the operating count of manhole pumps with few operating counts, or subtracting 1 from the operating count of manhole pumps with many operating counts, the average operating time per start during the diagnostic period is calculated, and this average value is added or subtracted as the operating time per start for the manhole pump corresponding to the addition or subtraction.

[0074] The diagnostic indicator calculation processing unit 45 normalizes the diagnostic indicators for the diagnostic period (1 day) based on statistical data calculated from only the diagnostic indicators of the manhole pump device 10 that were calculated during the most recent predetermined period including the diagnostic period (1 day) stored in the database DB, for example, several months (3 months in this embodiment). (SA3)

[0075] The diagnostic processing unit 46 inputs the normalized diagnostic index calculated by the diagnostic index calculation processing unit 45 to the machine learning device and performs an anomaly determination based on a pre-set boundary threshold (SA5). If an anomaly is determined, it sets a notification determination value that represents the degree of the anomaly in two levels, high and low, based on the distance of the diagnostic index from the boundary threshold, and outputs the notification determination value to the determination processing unit 47 (SA6).

[0076] Each time a notification judgment value is output, the judgment processing unit 47 performs filtering (SA7) to control the frequency of sending an anomaly to the notification recipient by the notification processing unit 48, based on the notification judgment values ​​output from the diagnostic processing unit 46 in the most recent predetermined number of past diagnostic intervals. If the filtering process determines that it is necessary to send an anomaly to the notification recipient (SA8, Y), the notification processing unit 48 identifies the cause of the anomaly based on a diagnostic map that shows a correlation with the cause of the anomaly in advance (SA9), and sends an alarm indicating the cause of the anomaly and the abnormal state to the administrator's mobile terminal or other device (SA10). The alarm notification is sent as an email via the mailer provided in the communication unit 41.

[0077] The diagnostic unit 44 will be described in detail below. Figure 4(a) shows the operating data of the manhole pump system 10 for the 24-hour period from 0:00 AM to 0:00 AM the following day, which is the period to be diagnosed. From top to bottom, the data shows the water level fluctuations, the operating timing and duration of pumps PA and PB, the current value of pump PA, and the current value of pump PB, respectively. In other words, in this embodiment, the diagnostic interval is 24 hours.

[0078] Figure 4(b) is an enlarged view to facilitate understanding of the relationship between the water level fluctuations shown in Figure 4(a) and the operating timing and operating time of pumps PA and PB. When the water level in the manhole reaches HWL, pump PA is started and stopped when the water level drops to LWL. Next, when the water level reaches HWL, pump PB is started and stopped when the water level drops to LWL. One of the pumps is running while the water level is dropping from HWL to LWL. The operating time of the pumps will be longer if the pump's transport volume is low or if the inflow of sewage into the manhole is high.

[0079] Such water level information and operating information stored in the memory unit 22 of each manhole pump device 10 are transmitted to the monitoring device 40 via the communication unit 24 and stored in the database DB via the data processing unit 42.

[0080] The data processing unit 42 passes this operational information for the diagnostic period to the diagnostic index calculation processing unit 45, and the diagnostic index calculation processing unit 45 calculates the "number of operations n1, n2" and the "total pump operating time t1, t2" for each pump PA, PB from the operational information passed to it as diagnostic indices.

[0081] As shown in Figure 5, the diagnostic index calculation processing unit 45 uses past diagnostic indexes obtained from driving information accumulated in the database DB over the past three months as the population, and normalizes the data by setting the mean μ and variance σ. 2 The average operating time of each feature is calculated and normalized using the formula (x-μ) / σ for the feature x.

[0082] While not limited to the most recent three months, it is preferable that the statistical data (mean and variance) required for normalization be calculated based on pump operation information including diagnostic indicators for the most recent predetermined period, including the diagnostic period at the time of normalization. This eliminates influences caused by the passage of time, such as seasonal variations, enabling a more reliable diagnosis.

[0083] When a pair of characteristic values, consisting of the normalized diagnostic indicators "ratio of operation counts n1 / n2" and "ratio of total pump operating time t1 / t2," are input to the diagnostic processing unit 46, one of the characteristic values ​​is plotted in a two-dimensional coordinate system with the x-axis and the y-axis. Based on which side of the pre-set boundary threshold the plotted characteristic value points lie on in the two-dimensional coordinate system, the normal or abnormal status of each manhole pump is diagnosed.

[0084] Figure 6(a) shows how a roughly circular closed curve (shown by a thick line) with a boundary threshold is used in a two-dimensional coordinate system where the vertical axis (y-axis) is the ratio of the total pump operating time (t1 / t2) and the horizontal axis (x-axis) is the ratio of the number of operating cycles (n1 / n2). The origin of the curve is the average value of each value. If the plotted characteristic value points are located inside the boundary threshold, it is determined to be normal, and if the plotted characteristic value points are located outside the boundary threshold, it is determined to be abnormal.

[0085] As described above, the machine learning device provided in the diagnostic processing unit 46 automatically generates boundary thresholds by learning using only diagnostic indicators for normal operation obtained from multiple manhole pumps installed in multiple manhole pump devices 10, including the manhole pump installed in the manhole pump device 10 that is the target of the diagnosis, as training data. The generated boundary thresholds indicate the boundaries of the coordinate system that becomes the feature space containing the normal diagnostic indicators. As learning algorithms executed by the machine learning device, one-class support vector machines, LOF (local outlier factor) methods, IF (Isolation Forest) methods, RC (Robust Covariance) methods, etc., can be used.

[0086] By performing machine learning, a normal data space, i.e., an internal space for boundary thresholds, is generated by mapping normal measurement data (training data) to the mapping space (feature space) shown in Figure 6(a). In the example in Figure 6(a), boundary thresholds obtained by training using one year's worth of paired characteristic values ​​in a normal state obtained from 100 manhole pump devices 10, each equipped with two pumps, are shown.

[0087] Figure 6(b) illustrates the case where the difference between the number of operating cycles n1 and n2 for each pump PA and PB is 1, showing uncorrected data plotted for the diagnostic indicators and corrected data plotted for the corrected diagnostic indicators. In the uncorrected diagnostic indicators, the values ​​deviate from the boundary threshold and are diagnosed as abnormal, while in the corrected diagnostic indicators, the values ​​lie within the boundary threshold and are diagnosed as normal.

[0088] When the diagnostic processing unit 46 diagnoses an abnormality because a characteristic value point is located outside the boundary threshold shown in Figure 6(a), it outputs a notification judgment value of at least two levels, high and low, to the judgment processing unit 47 according to the level of the abnormality, and stores the diagnostic result in the storage unit 22. Specifically, an abnormality degree judgment threshold (shown by a dashed line in Figure 6(a)) is set at a predetermined distance outside the boundary threshold, and when a plotted characteristic value point (shown by a black circle indicated by the symbol d1 in Figure 6(a)) is located between the boundary threshold and the abnormality degree judgment threshold, the notification judgment value is set to the abnormality level "low". When a plotted characteristic value point (shown by a black circle indicated by the symbol d2 in Figure 6(a)) is located outside the abnormality degree judgment threshold, the notification judgment value is set to the abnormality level "high".

[0089] Each time a notification judgment value is output, the judgment processing unit 47 performs filtering to control the transmission frequency by the notification processing unit 48, based on the notification judgment values ​​output from the diagnosis processing unit 46 during the most recent predetermined number of past diagnosis intervals stored in the storage unit 22. The notification processing unit 48 decides whether or not to transmit the diagnosis result according to the result of the filtering.

[0090] The filtering process controls the transmission frequency by the notification processing unit 48 based on whether the probability of occurrence of the notification judgment value output by the diagnostic processing unit 47 in the most recent past first predetermined number of diagnostic intervals (N1) has reached the first reference probability (P1), and whether the probability of occurrence of a high level of the notification judgment value output by the diagnostic processing unit 47 in the most recent past second predetermined number of diagnostic intervals (N2) has reached the second reference probability (P2).

[0091] The first predetermined number of repetitions (N1), the first reference probability (P1), the second predetermined number of repetitions (N2), and the second reference probability (P2) are parameters used in the filtering process, and are configured to be configurable via an input device such as a keyboard connected to the monitoring device 40, or via an external device such as a mobile terminal 30 through the communication unit 41.

[0092] Specifically, the filtering process allows the transmission of the diagnostic results if the probability of occurrence of the notification judgment value output by the diagnostic processing unit 47 within at least a first predetermined number of diagnostic intervals (N1) is higher than the first reference probability (P1). Furthermore, the filtering process allows the transmission of the diagnostic results if the probability of occurrence of a high level of the notification judgment value output by the diagnostic processing unit 47 within at least a second predetermined number of diagnostic intervals (N2) is higher than the second reference probability (P2).

[0093] The filtering process prohibits the transmission of diagnostic results if the probability of occurrence of the notification judgment value output by the diagnostic processing unit 47 within a first predetermined number of diagnostic intervals (N1) is less than or equal to the first reference probability (P1), and the probability of occurrence of a high level of the notification judgment value output by the diagnostic processing unit 47 within a second predetermined number of diagnostic intervals (N2) is less than or equal to the second reference probability (P2).

[0094] Furthermore, the values ​​of the first predetermined number of trials (N1), the first reference probability (P1), the second predetermined number of trials (N2), and the second reference probability (P2) are configured to be variable.

[0095] Figure 7 illustrates a diagnostic map. The area outside the boundary threshold shown in Figure 6(a) is divided into eight regions, and each region is associated with one of the causes of anomalies. The diagnostic processing unit 46 diagnoses the cause of the anomaly based on the region where each feature data is plotted.

[0096] For example, in the example shown in Figure 7, if characteristic value points are plotted in regions 1, 2, and 4, it is diagnosed that an abnormality, a decrease in drainage capacity, has occurred in pump PA, which is in long-term operation. If characteristic value points are plotted in regions 5, 7, and 8, it is diagnosed that an abnormality, a decrease in drainage capacity, has occurred in pump PB, which is in long-term operation. If characteristic value points are plotted in regions 3 and 6, it is diagnosed that an abnormality has occurred in the circuit components installed in the control panel 20, as the operating time and number of pumps are skewed to one side or the other.

[0097] Figure 8(a) shows the total number of notifications, missed cases, and false alarms when the first predetermined number of times (N1), first reference probability (P1), second predetermined number of times (N2), and second reference probability (P2) are specifically set as the parameters described above, for a manhole pump device 10 diagnosed by the diagnostic processing unit described above.

[0098] Figure 8(b) shows the characteristics of missed notifications and false alarms in relation to the total number of notifications. As the total number of notifications increases, the number of missed notifications decreases, but the number of false alarms increases. As the total number of notifications decreases, the number of missed notifications tends to increase, but the number of false alarms tends to decrease. Therefore, by setting the first predetermined number of notifications (N1), first baseline probability (P1), second predetermined number of notifications (N2), and second baseline probability (P2) to a level that provides a good balance between missed notifications and false alarms, it becomes possible to achieve good notification processing for administrators and other recipients. Furthermore, if administrators dislike "missed notifications," it becomes possible to prioritize avoiding "missed notifications" over increasing "false alarms," ​​and conversely, if administrators dislike "false alarms," ​​it becomes possible to prioritize avoiding "false alarms" over increasing "missed notifications."

[0099] In the above embodiment, the diagnostic processing unit 46 automatically diagnoses whether the operating status of the manhole pump is normal or abnormal based on the results of the diagnosis by the machine learning device at 24-hour diagnostic intervals. If it is diagnosed as abnormal, the determination processing unit 47 performs filtering to control the transmission frequency by the notification processing unit 48 based on the notification determination value output from the diagnostic processing unit 46 in the most recent predetermined number of past diagnostic intervals.

[0100] In this case, the diagnostic interval is not limited to 24 hours but can be set to an appropriate value. The notification judgment value is set to two levels, high and low, but it may be set to more than two levels. The value for the predetermined number of times can also be set to an appropriate value.

[0101] Furthermore, the diagnostic processing unit 46 may be configured to process the results of the machine learning device's diagnosis as a temporary diagnosis and output a notification judgment value of two levels, high and low, to the judgment processing unit 47 based on the cumulative results of multiple temporary diagnoses.

[0102] This configuration is shown in Figure 9. The process from step SB1 to step SB4 in Figure 9 is the same as the process from step SA1 to SA4 described in Figure 3. The diagnostic processing unit 46 may, each time a characteristic value point is initially diagnosed as abnormal by a machine learning device (SB5, "abnormal" judgment), multiply a predetermined abnormality criterion value by a weight coefficient based on the distance between the characteristic value point and the boundary threshold and add it (SB6). It may also, each time a characteristic value point is diagnosed as normal in the initial diagnosis (SB5, "normal" judgment), subtract a predetermined normal return evaluation value (SB7). If, as a result, the cumulative evaluation value is determined to be abnormal, it may execute the processes from steps SB8 to SB13. The processes from steps SB8 to SB13 are the same as the processes from steps SA5 to SA10 described in Figure 3.

[0103] The processing performed in steps SB5 to SB7 by the diagnostic processing unit 46 is as follows: Each time the diagnostic processing unit 46 makes a primary diagnosis that a characteristic value point is abnormal, it calculates a cumulative evaluation value V by multiplying a predetermined abnormality reference value Vnb by a weighting coefficient W based on the distance between the position of each characteristic value point and the boundary threshold, based on the following formula, and each time a characteristic value point is diagnosed as normal in the primary diagnosis, it subtracts a predetermined normal recovery evaluation value Vpb, and makes a final diagnosis of whether the manhole pump device is normal or abnormal based on the cumulative evaluation value. The distance between the position of a characteristic value point and the boundary threshold refers to the length of the normal from the boundary threshold that passes through the characteristic value point. V = Vnb × W - Vpb In this embodiment, Vnb = 1, and Vnb < Vpb < Wmax × Vnb is set.

[0104] Even if an abnormality is diagnosed in the primary diagnosis, there are mild abnormalities that will be diagnosed as normal later, and there are also abnormalities that will continue to be diagnosed as abnormal and eventually lead to a serious failure. In such a case, every time an abnormality is diagnosed in the primary diagnosis, a value obtained by multiplying a weight coefficient based on the distance between the characteristic value point and the boundary threshold by a predetermined abnormality reference value is added, and every time a normal diagnosis is made, a predetermined normal recovery evaluation value is subtracted to calculate a cumulative evaluation value, and it can be configured to determine the feasibility of the notification process based on the cumulative evaluation value.

[0105] FIG. 10 shows the transition of the cumulative evaluation value V. The cumulative evaluation value V is updated every time a characteristic value point is plotted. When an abnormality is determined, a value of Vnb × W is added to the initial value 0, and when a normal determination is made, the value of Vpb is subtracted.

[0106] When the cumulative evaluation value V exceeds the first threshold value Vth1, the notification determination value is set to the abnormal level "low", and when the cumulative evaluation value V exceeds the second threshold value Vth2 (Vth1 < Vth2), the notification determination value is set to the abnormal level "high".

[0107] In the above-described example, an example of adopting the ratio n1 / n2 of the number of pump operations and the ratio t1 / t2 of the operation time as diagnostic indicators has been described. However, the diagnostic indicators applicable to the present invention are not limited to these, and any indicators that can indicate the operating state of the pump can be appropriately adopted.

[0108] For example, the average operation time per startup of one manhole pump and the ratio of the average operation time per startup of both manhole pumps can be selected as a pair of diagnostic indicators. Also, the water level drop rate during the operation of the manhole pump and the pump drive current value can be selected as a pair of diagnostic indicators.

[0109] As described above, the manhole pump diagnostic method according to the present invention is a manhole pump diagnostic method that diagnoses the condition of two manhole pumps installed in a manhole that are repeatedly started and stopped, and is performed by the diagnostic device described above.

[0110] Specifically, a manhole pump diagnostic method comprising: a diagnostic step that automatically diagnoses whether the operating state of a manhole pump installed in a manhole is normal or abnormal at predetermined diagnostic intervals and outputs a diagnostic result; and a notification step that, if the diagnostic step diagnoses an abnormality, transmits the diagnostic result to a predetermined recipient, further comprising a determination step that determines whether or not to transmit the diagnostic result, which has been diagnosed as abnormal in the diagnostic step, to the recipient by the notification step, wherein the diagnostic step, upon diagnosing an abnormality, outputs a notification determination value of at least two levels, high and low, to the determination step according to the level of the abnormality, and each time the notification determination value is output, the determination step performs a filtering process to control the transmission frequency by the notification step based on the notification determination values ​​output from the diagnostic step in the most recent predetermined number of past diagnostic intervals, and the notification step is configured to determine whether or not to transmit the diagnostic result according to the result of the filtering process.

[0111] The filtering process is configured to control the transmission frequency of the notification step based on whether the probability of the notification judgment value output by the diagnostic step in the most recent past first predetermined number of diagnostic intervals (N1) has reached a first reference probability (P1), and whether the probability of a high level of the notification judgment value output by the diagnostic step in the most recent past second predetermined number of diagnostic intervals (N2) has reached a second reference probability (P2).

[0112] The filtering process is configured to allow the transmission of the diagnostic result if the probability of the notification judgment value output by the diagnostic step occurring at least one predetermined number of times (N1) within the diagnostic interval is higher than the first reference probability (P1).

[0113] The filtering process is configured to allow the transmission of the diagnostic result if the probability of a high level occurring in the notification judgment value output by the diagnostic step is higher than the second reference probability (P2) at least for the second predetermined number of diagnostic intervals (N2).

[0114] The filtering process is configured to prohibit the transmission of the diagnostic results if the probability of the notification judgment value output by the diagnostic step within a first predetermined number of diagnostic intervals (N1) is less than or equal to the first reference probability (P1), and the probability of a high level of the notification judgment value output by the diagnostic step within a second predetermined number of diagnostic intervals (N2) is less than or equal to the second reference probability (P2).

[0115] The values ​​of the first predetermined number of times (N1), the first reference probability (P1), the second predetermined number of times (N2), and the second reference probability (P2) are configured to be variable.

[0116] The diagnostic step is configured to diagnose the state of the manhole pump during the diagnostic period based on diagnostic indicators calculated from the operating information of the manhole pump collected during the diagnostic period. The diagnostic indicators are normalized based on statistical data calculated only from each diagnostic indicator calculated in the most recent predetermined period including the diagnostic period, plotted in a predetermined coordinate system, and the normality or abnormality of the manhole pump is diagnosed based on which side of a predetermined boundary threshold in the coordinate system the plotted diagnostic indicator points lie on. The boundary threshold is automatically generated by a machine learning device that learns using only diagnostic indicators for normal operation obtained from multiple manholes, including the manhole pump installed in the manhole subject to diagnosis, and indicates the boundary of the coordinate system, which is a feature space containing normal diagnostic indicators.

[0117] The diagnostic indicators include the ratio of the number of times one manhole pump was operated to the number of times the other manhole pump was operated during the diagnostic period, and the ratio of the operating time of one manhole pump to the operating time of the other manhole pump during the diagnostic period.

[0118] The embodiments described above are merely examples of the present invention, and the technical scope of the present invention is not limited by this description. It goes without saying that the specific configuration of each part, including pumps and water level gauges, the threshold values ​​set for abnormality detection, the diagnostic period, the predetermined period for which diagnostic indicators used in normalization processing are accumulated, the sampling period for diagnostic indicators used in learning processing, and the number of manhole pump devices can be appropriately modified within the scope in which the effects of the present invention are achieved. [Explanation of symbols]

[0119] 10: Manhole pump equipment PA, PB: Pump 18,19: Water level gauge 21: Control Unit 22: Storage section 24: Communications Department 30: Mobile devices 40: Monitoring device 41: Communications Department 42: Data Processing Unit 44: Diagnostic Department 45: Diagnostic Index Calculation Processing Unit 46: Diagnostic Processing Unit 47: Determination Processing Unit 48: Notification Processing Unit

Claims

1. A manhole pump diagnostic method comprising: a diagnostic step that automatically diagnoses whether the operating status of a manhole pump installed in a manhole is normal or abnormal at predetermined diagnostic intervals and outputs a diagnostic result; and a notification step that, if an abnormality is diagnosed in the diagnostic step, transmits the diagnostic result to a predetermined recipient, The system further comprises a determination step that determines whether or not to send the diagnosis result, which was diagnosed as abnormal in the diagnosis step, to the notification recipient in the notification step, If the diagnostic step diagnoses an abnormality, it outputs a notification judgment value to the judgment step at least in two levels, high and low, depending on the level of the abnormality. The determination step, each time the notification determination value is output, performs a filtering process to control the transmission frequency by the notification step, based on the probability of the notification determination value with the set level being output from the diagnosis step in the most recent predetermined number of past diagnostic intervals. The notification step is a method for diagnosing a manhole pump, which determines whether or not to transmit the diagnostic result according to the result of the filtering process.

2. The method for diagnosing a manhole pump according to claim 1, wherein the filtering process controls the transmission frequency by the notification step based on whether the probability of occurrence of the notification judgment value output by the diagnosis step in the most recent past first predetermined number of diagnostic intervals (N1) has reached a first reference probability (P1), and whether the probability of occurrence of a high level of the notification judgment value output by the diagnosis step in the most recent past second predetermined number of diagnostic intervals (N2) has reached a second reference probability (P2).

3. The method for diagnosing a manhole pump according to claim 2, wherein the filtering process allows transmission of the diagnosis result when the probability of occurrence of the notification judgment value output by the diagnosis step is higher than the first reference probability (P1) at least one predetermined number of diagnosis intervals (N1).

4. The method for diagnosing a manhole pump according to claim 2, wherein the filtering process allows transmission of the diagnosis result when the probability of a high level occurring in the notification judgment value output by the diagnosis step is higher than the second reference probability (P2) at least two predetermined diagnostic intervals (N2).

5. The method for diagnosing a manhole pump according to any one of claims 2 to 4, wherein the filtering process prohibits the transmission of the diagnosis results when the probability of the notification judgment value output by the diagnosis step in a first predetermined number of diagnosis intervals (N1) is less than or equal to the first reference probability (P1), and the probability of a high level of the notification judgment value output by the diagnosis step in a second predetermined number of diagnosis intervals (N2) is less than or equal to the second reference probability (P2).

6. A method for diagnosing a manhole pump according to any one of claims 2 to 4, wherein the values ​​of the first predetermined number of times (N1), the first reference probability (P1), the second predetermined number of times (N2), and the second reference probability (P2) are configured to be variable.

7. The diagnostic step is configured to diagnose the condition of the manhole pump during the diagnostic period based on diagnostic indicators calculated from the operating information of the manhole pump collected during the diagnostic period. The normal or abnormal state of the manhole pump is diagnosed based on the normalized diagnostic indicators calculated using statistical data derived only from each diagnostic indicator calculated during the most recent predetermined period including the aforementioned diagnostic target period, plotted in a predetermined coordinate system, and determined based on which side of the predetermined boundary threshold the plotted diagnostic indicator points lie on in the coordinate system. The method for diagnosing a manhole pump according to any one of claims 1 to 4, wherein the boundary threshold is automatically generated by a machine learning device that learns using only diagnostic indicators for normal operation obtained from a plurality of manholes, including the manhole pump installed in the manhole to be diagnosed, and the boundary of the coordinate system is a feature space that contains normal diagnostic indicators.

8. The method for diagnosing a manhole pump according to claim 7, wherein the diagnostic indicators include the ratio of the number of times one manhole pump is operated to the number of times the other manhole pump is operated during the diagnostic period, and the ratio of the operating time of one manhole pump to the operating time of the other manhole pump during the diagnostic period.

9. A manhole pump diagnostic device comprising: a diagnostic processing unit that automatically diagnoses whether the operating status of a manhole pump installed in a manhole is normal or abnormal at predetermined diagnostic intervals and outputs a diagnostic result; and a notification processing unit that transmits the diagnostic result to a predetermined recipient when the diagnostic processing unit diagnoses an abnormality, The system further includes a determination processing unit that determines whether or not to send the diagnosis result, which has been diagnosed as abnormal by the diagnosis processing unit, to the notification recipient via the notification processing unit. When the diagnostic processing unit diagnoses an abnormality, it outputs a notification judgment value to the judgment processing unit at least two levels, high and low, according to the level of the abnormality. Each time the notification judgment value is output, the judgment processing unit executes a filter process to control the transmission frequency by the notification processing unit, based on the probability of occurrence of the notification judgment value with the set level, which was output from the diagnostic processing unit in the most recent predetermined number of past diagnostic intervals. The notification processing unit is a diagnostic device for a manhole pump that determines whether or not to transmit the diagnostic result according to the result of the filtering process.

10. The diagnostic device for a manhole pump according to claim 9, wherein the filtering process controls the transmission frequency by the notification processing unit based on whether the probability of occurrence of the notification judgment value output by the diagnostic processing unit in the most recent past first predetermined number of diagnostic intervals (N1) has reached a first reference probability (P1), and whether the probability of occurrence of a high level of the notification judgment value output by the diagnostic processing unit in the most recent past second predetermined number of diagnostic intervals (N2) has reached a second reference probability (P2).

11. The manhole pump diagnostic device according to claim 10, wherein the filtering process allows transmission of the diagnostic result when the probability of occurrence of the notification judgment value output by the diagnostic processing unit at least one predetermined number of diagnostic intervals (N1) is higher than the first reference probability (P1).

12. The manhole pump diagnostic device according to claim 10, wherein the filtering process allows transmission of the diagnostic result when the probability of a high level occurring in the notification judgment value output by the diagnostic processing unit at least the second predetermined number of diagnostic intervals (N2) is higher than the second reference probability (P2).

13. The manhole pump diagnostic device according to any one of claims 10 to 12, wherein the filtering process prohibits the transmission of the diagnostic results when the probability of occurrence of the notification judgment value output by the diagnostic processing unit at a first predetermined number of diagnostic intervals (N1) is less than or equal to the first reference probability (P1), and the probability of occurrence of a high level of the notification judgment value output by the diagnostic processing unit at a second predetermined number of diagnostic intervals (N2) is less than or equal to the second reference probability (P2).

14. A diagnostic device for a manhole pump according to any one of claims 10 to 12, wherein the values ​​of the first predetermined number of times (N1), the first reference probability (P1), the second predetermined number of times (N2), and the second reference probability (P2) are configured to be variably set.

15. The diagnostic processing unit is configured to diagnose the state of the manhole pump during the diagnostic period based on diagnostic indicators calculated from the operating information of the manhole pump collected during the diagnostic period. The normal or abnormal state of the manhole pump is diagnosed based on the normalized diagnostic indicators calculated using statistical data derived only from each diagnostic indicator calculated during the most recent predetermined period including the aforementioned diagnostic target period, plotted in a predetermined coordinate system, and determined based on which side of the predetermined boundary threshold the plotted diagnostic indicator points lie on in the coordinate system. The manhole pump diagnostic device according to any one of claims 9 to 12, wherein the boundary threshold is automatically generated by a machine learning device that learns using only diagnostic indicators for normal operating conditions obtained from a plurality of manholes, including the manhole pump installed in the manhole to be diagnosed, and indicates the boundary of the coordinate system which is a feature space containing normal diagnostic indicators.

16. The manhole pump diagnostic device according to claim 15, wherein the diagnostic indicator includes the ratio of the number of times one manhole pump is operated to the number of times the other manhole pump is operated during the diagnostic period, and the ratio of the operating time of one manhole pump to the operating time of the other manhole pump during the diagnostic period.

Images