Manhole pump diagnostic method and diagnostic device

Abstract

To provide a checkup method for a manhole pump that is able to give an accurate checkup for anomaly caused by various reasons even with a small physical quantity and is able to appropriately eliminate an influence caused by a lapse of time.SOLUTION: A checkup method for a manhole pump comprises: a sampling step of chronologically sampling measurement data indicating characteristics of the manhole pump to obtain characteristic data for checkup; a normalization step of, based on predetermined statistical data, normalizing the characteristic data obtained in the sampling step; a checkup step of giving a checkup to determine whether the manhole pump is normal or anomalous based on the characteristic data normalized in the normalization step and a preset boundary threshold; and a statistical data update step of, when a predetermined update condition is satisfied, updating the predetermined statistical data based on characteristic data before normalization which has been checked up as normal in the checkup step in an immediately preceding predetermined period.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a method and device for diagnosing a manhole pump that is installed in a manhole and is repeatedly started and stopped. [Background technology]

[0002] The manhole pump equipment includes a storage section for storing sewage flowing in from the inlet pipe, a number of pumps for discharging the sewage stored in the storage section into an outlet pipe, a water level gauge for measuring the level of the sewage stored in the storage section, and a control panel equipped with a control device that performs sewage transport control by starting one of the pumps to drain the sewage into the outlet pipe when the water level measured by the water level gauge reaches the pump start-up level, and stopping the pump when the water level reaches the pump stop level.

[0003] Such manhole pump facilities are usually equipped with two pumps, and the control device is configured to operate the pumps alternately each time sewage is transported.

[0004] Patent document 1 proposes a pumping station monitoring system that includes a means for managing the cumulative inflow volume to be discharged by multiple pumps at a pumping station per given time period divided by the operating time of each pump within a given time period as the average drainage capacity of each pump as time series data, a means for determining whether the average drainage capacity of each pump managed as time series data has exceeded a predetermined drainage capacity range, and a means for issuing an alarm when the average drainage capacity of each of the multiple pumps exceeds the set drainage capacity range.

[0005] Patent document 2 discloses a submersible pump control device that includes a pressure control unit that activates an electromagnetic contactor to drive the submersible pump when the stored water level detected by a water level sensor reaches a pump start-up water level, and an abnormality determination unit that determines the presence or absence of an abnormality based on the operating state of the electromagnetic contactor that drives the electric motor, the detection state of a current sensor that detects the current in the power supply line connected to the armature winding via the electromagnetic contactor, and the stored water level, where the abnormality determination unit determines that an overheating abnormality has occurred in the electric motor with auto-cut activated if no current is detected by the current sensor while the electromagnetic contactor is operating and no drop in the stored water level is detected by the water level sensor.

[0006] The control device installed in such manhole pump facilities judges whether each pump is abnormal based on whether physical quantities such as the pump's drive current value, the pump's operation time from the start water level to the stop water level, and the pump's temperature exceed predetermined thresholds. In order to measure physical quantities appropriate to the type of abnormality expected, it is necessary to perform measurement processing using various sensors, which causes a problem of being very complicated.

[0007] In addition, it was difficult to set a uniform threshold value for judging an abnormality because it depended on the environment of the manhole the pump was installed in. For example, in areas with a high amount of water inflow per unit time and areas with a low amount of water inflow per unit time, there was a bias in the frequency of pump startup and the length of operation, making it difficult to make an accurate judgment if a fixed threshold value was used.

[0008] Therefore, Patent Document 3 proposes a method for diagnosing machinery and equipment that can accurately diagnose abnormalities caused by various factors even with a small physical quantity. The mechanical equipment diagnosis method includes a sampling step of chronologically sampling n types (n is an integer greater than or equal to 2) of measurement data groups indicating characteristics of the mechanical equipment, a normalization step of normalizing the sampled measurement data groups, and a diagnosis step of plotting each of the normalized measurement data groups in an n-dimensional coordinate system and making a primary diagnosis of whether the mechanical equipment is normal or abnormal based on on which side of a boundary threshold preset in the n-dimensional coordinate system the plotted measurement data groups are located.

[0009] Moreover, Patent Document 4 proposes a method for diagnosing a manhole pump with a similar purpose. The manhole pump diagnostic method includes a sampling step of sampling the operating time per start of each manhole pump within a predetermined period and calculating the respective average operating times as a pair of characteristic values, a normalization step of normalizing the pair of characteristic values ​​calculated in the sampling step, and a diagnosis step of plotting the pair of normalized characteristic values ​​in a two-dimensional coordinate system in which one of the pair of characteristic values ​​is on the x-axis and the other is on the y-axis, and performing a primary diagnosis of whether each manhole pump is normal or abnormal based on on which side of a boundary threshold preset in the two-dimensional coordinate system the plotted characteristic value point is located. [Prior art documents] [Patent documents]

[0010] [Patent Document 1] JP 2001-34338 A [Patent Document 2] JP 2010-236191 A [Patent Document 3] JP 2020-107072 A [Patent Document 4] JP 2020-107073 A Summary of the Invention [Problem to be solved by the invention]

[0011] According to the diagnostic methods described in Patent Documents 3 and 4, it is possible to judge an abnormality using an appropriate threshold value, regardless of the environment of the manhole in which the pump is installed.

[0012] Furthermore, the statistical data required for the normalization process used in the normalization step is calculated based on multiple measurement data or each characteristic value for a specified period immediately preceding the time the normalization process is performed, thereby eliminating effects due to the passage of time, such as seasonal fluctuations, and enabling highly reliable diagnosis.

[0013] However, the measurement data or characteristic values ​​for the most recent specified period used to calculate the statistical data include data that was diagnosed as abnormal in the diagnostic step, and if all of the measurement data or characteristic values ​​including such data are used, it may become difficult to properly eliminate the effects caused by the passage of time.

[0014] In view of the above-mentioned problems, an object of the present invention is to provide a manhole pump diagnostic method and diagnostic device that can accurately diagnose abnormalities due to various causes even with small physical quantities and can appropriately eliminate effects caused by the passage of time. [Means for solving the problem]

[0015] In order to achieve the above-mentioned object, a first characteristic configuration of a manhole pump diagnostic method according to the present invention is a method for diagnosing a manhole pump that is installed in a manhole and repeatedly starts and stops, comprising: a sampling step of obtaining diagnostic characteristic data by chronologically sampling measurement data showing characteristics of the manhole pump; a normalization step of normalizing the characteristic data obtained in the sampling step based on predetermined statistical data; and a diagnosis step of diagnosing whether the manhole pump is normal or abnormal based on the characteristic data normalized in the normalization step and a preset boundary threshold value. and, a statistical data update step for updating the predetermined statistical data based on characteristic data before normalization that was diagnosed as normal in the diagnosis step during the immediately preceding predetermined period when a predetermined update condition is satisfied; and an external trigger is set as the predetermined update condition, the external trigger being generated in response to a predetermined time lapse corresponding to a seasonal variation in the amount of wastewater inflow into the manhole, or a change in the surrounding environment that causes an increase or decrease in the amount of wastewater generation source. It's at the point.

[0016] By normalizing the diagnostic characteristic data obtained from the measurement data sampled in a time-series manner in the sampling step in the normalization step, it is possible to extract characteristic data as feature quantities that have been free of the effects of, for example, inherent machine differences in the manhole pump and the effects of the installation environment, and the diagnostic step appropriately diagnoses whether the manhole pump is normal or abnormal based on the characteristic data and a preset boundary threshold. When a predetermined update condition is met, the predetermined statistical data for normalization is updated based on the characteristic data before normalization that was sampled in the immediately preceding predetermined period and diagnosed as normal in the diagnostic step, so that there is no risk of errors occurring over time. Due to fluctuations in the amount of inflow sewage The influence is appropriately eliminated.

[0017] The predetermined time lapse is a period required to eliminate influences caused by the passage of time, such as the time elapsed after the execution of a statistical data update step. In addition, the occurrence of an external trigger can be used as a trigger to eliminate influences other than those caused by the passage of time. This can be the case when the amount of sewage flowing into the sewage conveyance system including the target manhole fluctuates greatly, such as when a new sewage source is created or an existing sewage source is eliminated due to the construction or demolition of a house, condominium, factory, etc., or long-term bad weather.

[0018] The second characteristic configuration is a method for diagnosing a manhole pump that is installed in a manhole and repeatedly starts and stops, comprising: a sampling step of obtaining diagnostic characteristic data by sampling measurement data showing characteristics of the manhole pump in a time series manner; a normalization step of normalizing the characteristic data obtained in the sampling step based on predetermined statistical data; and a diagnosis step of diagnosing whether the manhole pump is normal or abnormal based on the characteristic data normalized in the normalization step and a preset boundary threshold value. and, a statistical data update step for updating the predetermined statistical data based on a predetermined number or more of pre-normalized characteristic data diagnosed as normal in the diagnosis step during the immediately preceding period when a predetermined update condition is satisfied; and an external trigger is set as the predetermined update condition, the external trigger being generated in response to a predetermined time lapse corresponding to a seasonal variation in the amount of wastewater inflow into the manhole, or a change in the surrounding environment that causes an increase or decrease in the amount of wastewater generation source. It's at the point.

[0019] When a predetermined update condition is satisfied, the predetermined statistical data for normalization is updated based on a predetermined number or more of characteristic data before normalization that have been sampled in the immediately preceding period and diagnosed as normal in the diagnosis step. Due to fluctuations in the amount of inflow sewage The influence is appropriately eliminated.

[0020] The predetermined time lapse is a period required to eliminate influences caused by the passage of time, such as the time elapsed after the execution of a statistical data update step. In addition, the occurrence of an external trigger can be used as a trigger to eliminate influences other than those caused by the passage of time. This can be the case when the amount of sewage flowing into the sewage conveyance system including the target manhole fluctuates greatly, such as when a new sewage source is created or an existing sewage source is eliminated due to the construction or demolition of a house, condominium, factory, etc., or long-term bad weather.

[0021] The third characteristic configuration is, in addition to the first or second characteristic configuration described above, that the measurement data includes the operating time and the number of times the manhole pump is operated.

[0022] By including the operating time and number of times the manhole pump is operated in the measurement data and using the simple characteristic data obtained from this, it becomes possible to accurately diagnose abnormalities due to various causes.

[0023] A fourth characteristic configuration of the same is that, in addition to the first or second characteristic configuration described above, the measurement data includes the operating time of the manhole pump, a current value, and a water level in the manhole.

[0024] The measurement data includes the operating time of the manhole pump, the current value, and the water level in the manhole, and by using the characteristic data obtained from these, it becomes possible to accurately diagnose abnormalities caused by various factors. same Five The characteristic configuration of the above is the first to second four In addition to any one of the characteristic configurations above, the boundary threshold is maintained at an initially set value regardless of whether or not the statistical data update step is performed.

[0025] The boundary threshold required to determine abnormalities in manhole pumps installed in individual manholes is not set individually for each manhole, but is set to a common value for all manhole pumps, making it possible to perform accurate abnormality diagnosis without having to respond individually.

[0026] A first characteristic configuration of a manhole pump diagnostic device according to the present invention is a diagnostic device for a manhole pump that is installed in a manhole and repeatedly starts and stops, the diagnostic device comprising: a normalization processing unit that normalizes diagnostic characteristic data obtained from measurement data that indicates characteristics of the manhole pump sampled within a predetermined period, based on predetermined statistical data; and a diagnostic processing unit that diagnoses whether the manhole pump is normal or abnormal, based on the characteristic data normalized by the normalization processing unit and a preset boundary threshold value. and,a statistical data update processing unit that updates the predetermined statistical data based on pre-normalization characteristic data diagnosed as normal by the diagnosis processing unit during a previous predetermined period when a predetermined update condition is satisfied; and an external trigger is set as the predetermined update condition, the external trigger being generated in response to a predetermined time lapse corresponding to a seasonal variation in the amount of wastewater inflow into the manhole, or a change in the surrounding environment that causes an increase or decrease in the amount of wastewater generation source. It's at the point.

[0027] The second characteristic configuration is a diagnostic device for a manhole pump that is installed in a manhole and repeatedly starts and stops, the diagnostic device including a normalization processing unit that normalizes diagnostic characteristic data obtained from measurement data that indicates characteristics of the manhole pump sampled within a predetermined period based on predetermined statistical data, a diagnostic processing unit that diagnoses whether the manhole pump is normal or abnormal based on the characteristic data normalized by the normalization processing unit and a preset boundary threshold value, and a statistical data update unit that updates the predetermined statistical data when a predetermined update condition is established based on a predetermined number or more of characteristic data before normalization that have been diagnosed as normal by the diagnostic processing unit within the immediately preceding period. Processing section and an external trigger is set as the specified update condition, the external trigger being generated in response to a specified time period corresponding to seasonal fluctuations in the amount of sewage inflow into the manhole, or a change in the surrounding environment that causes an increase or decrease in the number of sewage sources. Effect of the Invention

[0028] As described above, according to the present invention, it is possible to provide a manhole pump diagnostic method and diagnostic device that can accurately diagnose abnormalities due to various causes even with small physical quantities and appropriately eliminate effects caused by the passage of time. [Brief description of the drawings]

[0029] [Figure 1] Illustration of a manhole pump [Diagram 2] Diagram of manhole pump abnormality diagnosis device [Diagram 3] An explanatory diagram showing the procedure for diagnosing an abnormality in a manhole pump [Figure 4] (a) is an explanatory diagram of the measurement data, and (b) is an enlarged explanatory diagram of the main part of the measurement data. [Diagram 5]Normalization process diagram [Figure 6] Illustration of primary diagnosis [Figure 7] Cumulative evaluation score and final diagnosis [Figure 8] Diagram of diagnostic map [Figure 9] FIG. 11 is an explanatory diagram showing another embodiment and illustrating the procedure of a method for diagnosing an abnormality in a manhole pump. [Figure 10] FIG. 13 is an explanatory diagram of a primary diagnosis according to another embodiment. [Figure 11] FIG. 13 is an explanatory diagram of a diagnostic map according to another embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0030] The manhole pump diagnostic method and diagnostic device according to the present invention will be described below. [Manhole pump system configuration] Figure 1 shows a manhole pump device 10. The manhole pump device 10 includes a manhole 12 as a water storage section for storing sewage that flows in from an upstream sewage inlet pipe 11, two pumps PA and PB for pumping the sewage stored in the manhole 12 to a downstream sewage outlet pipe 13, and water level gauges 18 and 19 for measuring the water level of the sewage stored in the manhole 12.

[0031] A first lift pipe 15b, a first curved pipe 15c, and a first horizontal pipe 15d are each flange-connected to the discharge curved pipe 15a of the first pump PA, 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 lift pipe 15b and the first curved pipe 15c.

[0032] A second lift pipe 17b and a second bent pipe 17c are flange-connected to the discharge bent pipe 17a of the second pump PB, and the second bent pipe 17c is flange-connected to the sewage outflow pipe 13 via the header pipe 13a. A check valve 17e is provided between the second lift pipe 17b and the second bent pipe 17c.

[0033] A water level gauge 18 of the plunge pressure type or air bubble type is installed at the bottom of the manhole 12. The 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 for detecting the abnormally high water level HHWL.

[0034] A control panel device 200 is installed near the manhole 12, and includes a control panel 20 including a control unit 21 that controls pumps PA, PB to perform sewage transport control to pressure-feed the sewage accumulated in the manhole 12 into the sewage outflow pipe 13.

[0035] The control panel 20 is provided 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 the water level gauges 18 and 19, and the like. The communication unit 24 includes a transmission unit that transmits the 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.

[0036] 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, and the monitoring device 40 and the communication unit 24 are connected to the Internet via such a communication medium. Furthermore, the mobile communication terminal 30 owned by the manager of the manhole pump device 10 and the monitoring device 40 are configured to be able to connect to the Internet via the wireless communication medium.

[0037] The control panel 20 and each of the pumps PA, PB are connected by AC power feed lines L1, L2, and the control panel 20 and the water level gauges 18, 19 are connected by a signal line S.

[0038] When the control unit 21 detects that the water level measured by the water level gauge 18 has reached a predetermined pump start-up water level HWL, it controls the power supply from the power supply line L1 to start one of the pumps PA, PB, and when it detects that the water level has reached a pump stop water level LWL, which is lower than the pump start-up water level HWL, it stops the power supply and stops the other pump PA.

[0039] When the control unit 21 detects that the water level has reached the pump start water level HWL again, it controls the power supply from the power supply line L2 to start the other pump PB, and 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 controls the operation of the pumps PA and PB alternately.

[0040] Furthermore, when the control unit 21 cannot detect the pump start-up water level HWL due to a malfunction of the water level gauge 18, or when it detects that a large amount of rainwater has flowed into the manhole 12 due to heavy rain, exceeding the drainage capacity of a single pump, and the water level gauge 19 measures that the water level has reached the abnormally high water level HHWL, it operates the two pumps PA and PB simultaneously.

[0041] The control unit 21 samples the water level information detected by the water level gauges 18, 19 in chronological order, for example at one-minute intervals, and stores it in the memory unit 22, and also stores in the memory unit 22 chronological operation information such as the start-up and stop-up times of each pump PA, PB, and the operating time from start-up to stop.

[0042] As shown in FIG. 2, the communication unit 24 provided on the control panel 20 of each manhole pump device 10 is configured to transmit the water level information and operating information stored in the memory unit 22 to the monitoring device 40 at predetermined intervals.

[0043] [Configuration of manhole pump diagnostic device] The monitoring device 40 functions as a diagnostic device for the manhole pump apparatus 10, and is equipped with a communication unit 41 that communicates with the communication unit 24 of each manhole pump apparatus 10 and the manager's mobile communication terminal 30, a database DB that stores water level information and operating information transmitted from the communication unit 24 of each manhole pump apparatus 10, a data processing unit 42 that exchanges data with the database DB, and a diagnostic unit 44 that diagnoses whether each manhole pump apparatus 10 is operating normally based on the water level information and operating information stored in the database DB.

[0044] The diagnosis unit 44 includes a normalization processing unit 46, a statistical data update processing unit 47, and a diagnosis processing unit . The normalization processing unit 46 is configured to normalize diagnostic characteristic data obtained from measurement data indicating the characteristics of the manhole pump sampled within a specified period, for example, the last 24 hours, based on specified statistical data generated by the statistical data update processing unit 47.

[0045] The diagnosis processing unit 48 diagnoses whether the manhole pump is normal or abnormal based on the characteristic data normalized by the normalization processing unit 46 and a preset boundary threshold value.

[0046] In detail, the diagnostic processing unit 48 is configured with a machine learning device that automatically generates boundary thresholds for diagnosing whether something is normal or abnormal based on multiple past characteristic data input as learning data, and the machine learning device diagnoses whether the normalized characteristic data is in a normal state where it falls within the boundary threshold, or whether the characteristic data is in an abnormal state where it deviates from the boundary threshold.

[0047] The machine learning device is composed of a computer that executes a one-class support vector machine algorithm, and the boundary threshold is set based on the results of learning using, as learning data, one year's worth of characteristic data obtained from multiple manhole pump devices, for example, up to 100 locations. The boundary threshold is a fixed value that is shared for the diagnosis of all manhole pump devices and is maintained thereafter after being initially set.

[0048] The statistical data update processing unit 47 is configured to update and process predetermined statistical data based on a predetermined number or more of pre-normalized characteristic data sampled during a predetermined immediately preceding period and diagnosed as normal by the diagnosis processing unit when a predetermined update condition is met.

[0049] The predetermined update condition is the passage of a predetermined time or the occurrence of an external trigger. The predetermined time is set in units of several tens of days to several months to correspond to seasonal fluctuations, and is set in response to an increase or decrease in the amount of wastewater generated, for example, at the start of the rainy season or the start of the dry season. The elapsed time is counted based on a timer function provided in the monitoring device 40, and when the preset time has elapsed, the update process of the statistical data is started. In addition, the number of times the manhole pump is operated may be counted, and if the number of times the manhole pump is operated within a preset time does not reach a preset number, the preset time may be extended until the preset number is reached.

[0050] In addition, the occurrence of an external trigger can be used as a trigger to eliminate influences other than those caused by the passage of time. Examples of external triggers include an update command from the mobile communication terminal 30 owned by the manager of the manhole pump device 10, and the statistical data update process is started when the amount of sewage flowing into the sewage conveyance system including the target manhole changes significantly, such as when a new sewage source is created or an existing sewage source disappears due to the construction or demolition of a house, condominium, factory, etc., or long-term bad weather.

[0051] The specific values ​​of the immediately preceding specified period and the specified number of times diagnosed as normal by the diagnosis processing unit are not particularly limited, but in this embodiment, the specified period is set to 10 to 100 days, and the specified number is set to 20 to 200. The specified period and the specified elapsed time, which is an update condition, may be the same value or may be different values.

[0052] In the initial stage when no statistical data is set, the statistical data update processing unit 47 generates statistical data based on characteristic data sampled during a preset period. For example, the statistical data is generated based on the mean value μ and variance σ for normalizing the data, using the feature amounts accumulated in the database DB during the most recent three months as a population. 2 Then, the characteristic data of each feature is normalized by the formula (x-μ) / σ for the feature x.

[0053] In the above example, the statistical data update processing unit 47 is configured to update the specified statistical data based on a specified number or more of pre-normalization characteristic data diagnosed as normal by the diagnostic processing unit in the immediately preceding specified period when a specified update condition is met. However, the statistical data update processing unit 47 may be configured to update the specified statistical data based on pre-normalization characteristic data diagnosed as normal by the diagnostic processing unit in the specified period without setting a limit on the number of pre-normalization characteristic data diagnosed as normal, or may be configured to update the specified statistical data based on a specified number or more of pre-normalization characteristic data diagnosed as normal by the diagnostic processing unit in the immediately preceding period without setting a limit on the specified period.

[0054] [Example 1 of a manhole pump diagnostic method] A first specific example will be described below. The normalization processing unit 46 is configured to normalize each pair of characteristic data obtained from the operation times of the pumps PA and PB over a predetermined period of time, which are sampled in a time series manner. The average operation times of the pumps can be calculated as a pair of characteristic data, and the average operation time of one pump and the ratio of the average operation time of one pump to the average operation time of the other pump can also be calculated as a pair of characteristic data.

[0055] The machine learning device provided in the diagnostic processing unit 48 is configured to plot one of a pair of normalized characteristic data on a two-dimensional coordinate system with the x-axis and the other on the y-axis, and to perform a primary diagnosis of the normality or abnormality of each manhole pump based on on which side of a boundary threshold preset in the two-dimensional coordinate system the plotted characteristic data point is located.

[0056] Furthermore, the diagnostic processing unit 48 is configured to multiply a predetermined abnormality reference value by a weighting coefficient based on the distance between the characteristic data point and the boundary threshold value and add the result each time the characteristic data point is primarily diagnosed as abnormal, and to make a final diagnosis of whether the manhole pump is normal or abnormal based on a cumulative evaluation value obtained by subtracting a predetermined normal return evaluation value each time the characteristic data point is primarily diagnosed as normal.

[0057] The outside of the boundary threshold set in the two-dimensional coordinate system is divided into a plurality of regions, and a diagnostic map is provided in which each region is associated with one of the causes of anomaly. The diagnostic processing unit 48 is configured to diagnose the cause of anomaly based on the region in which each characteristic value point is plotted.

[0058] 3 shows a flow of a series of diagnostic processes executed by the monitoring device 40. When one day's worth of measurement data has been received from each manhole pump apparatus 10 and stored in the database DB (SA1), the operation time of each of the pumps PA, PB is extracted as characteristic data to be used for abnormality determination for each pump of each manhole pump apparatus 10, and the average operation time of each of the pumps PA, PB is calculated by dividing the total operation time of the pumps in one day by the number of operations (SA2).

[0059] As the feature amount, the average operation time of each pump may be selected as a pair of characteristic data, or the average operation time of one pump and the ratio of the average operation time of one pump to the average operation time of the other pump may be selected as a pair of characteristic data, or the natural logarithm of the ratio of the operation times of both pumps and the natural logarithm of the ratio of the operation counts may be selected as a pair of characteristic data.

[0060] It is determined whether a predetermined update condition is satisfied, and if it is satisfied (SA3, Y), the statistical data update processing unit 47 executes update processing of the statistics (SA4), and normalizes the feature quantity of the day to be judged based on the updated statistics, i.e., the average value and variance value (SA5).If the update condition is not satisfied (SA3, N), the previous statistics are used to normalize the feature quantity of the day to be judged (SA5).

[0061] The normalized features are input into a machine learning device and a primary judgment is made based on a set boundary threshold (SA7). If an abnormality is detected, the cumulative evaluation value is increased taking into account the degree of abnormality (SA8), and if a normality is detected, the cumulative evaluation value is decreased (SA9).

[0062] This process is repeated every day to determine whether the cumulative evaluation value calculated exceeds a preset cumulative abnormality threshold value (SA10), and if it does exceed the cumulative abnormality threshold value, the cause of the abnormality is identified based on a diagnostic map that indicates a correlation with the cause of the abnormality (SA11), and an alarm indicating an abnormal state together with the cause of the abnormality is sent to a mobile terminal or the like owned by the manager (SA12). The alarm notification is sent as an e-mail via a mailer provided in the communication unit 41.

[0063] Figure 4(a) shows the operation data of the manhole pump equipment for 24 hours from 00:00 to 00:00 the next day. From the top to the bottom, the data shows the fluctuation of the water level, the operation timing and operation time of pumps PA and PB, the current value of pump PA, and the current value of pump PB.

[0064] Figure 4(b) is an enlarged view to make it easier to understand the relationship between the fluctuations in the water level shown in Figure 4(a) and the operation timing and operation time of pumps PA and PB. When the stored water level in the manhole reaches HWL, pump PA is started, and when the water level drops to LWL, it is stopped. Next, when the stored water level reaches HWL, pump PB is started, and when the water level drops to LWL, it is stopped. One of the pumps is running until the water level drops from HWL to LWL. The pump operation time will be longer if the pump's transport rate is low or if there is a large amount of sewage flowing into the manhole.

[0065] Such water level information and operation information stored in the memory unit 22 of each manhole pump device 10 is transmitted to the monitoring device 40 via the communication unit 24 and stored in the database DB via the data processing unit 42.

[0066] The data processing unit 42 extracts from such data the “total pump operation time” and the “number of operations” for each pump PA, PB when it is started, and based on the average daily operation time obtained by dividing the “total pump operation time” by the “number of operations”, calculates the average operation time of each pump as a pair of characteristic data, or calculates the average operation time of one pump and the ratio of the average operation time of one pump to the average operation time of the other pump as a pair of characteristic data, and passes them to the normalization processing unit 46.

[0067] As shown in FIG. 5, the mean value μ and the variance σ , which are statistics for normalization generated by the statistical data update processing unit 47, 2 Based on this, the normalization processing unit 46 normalizes the average driving time, which is the feature amount, for the feature amount x using the formula (x-μ) / σ.

[0068] When a pair of characteristic data consisting of the feature "average operating time of each pump" normalized by the normalization processing unit 46, or a pair of characteristic data consisting of "the average operating time of one pump and the ratio of the average operating time of one pump to the average operating time of the other pump," or a pair of characteristic data consisting of the natural logarithm of the ratio of the operating times of both pumps and the natural logarithm of the ratio of the number of times they are operated, is input to the diagnostic processing unit 48, the normalized pair of characteristic data is plotted in a two-dimensional coordinate system with one of the pair of characteristic data on the x-axis and the other on the y-axis, and a primary diagnosis of the normality or abnormality of each manhole pump is made based on on which side of a boundary threshold preset in the two-dimensional coordinate system the plotted characteristic data point is located.

[0069] By normalizing the characteristic data calculated based on the measurement data sampled in a time series, it is possible to extract characteristic values ​​that eliminate the effects of, for example, mechanical differences in the manhole pump and the effects of the installation environment as feature values. Through diagnosis, these feature values, i.e., the points representing the two types of measurement data groups, are plotted as multiple points in a time series in a two-dimensional coordinate system, and an appropriate initial diagnosis is made as to whether the manhole pump equipment is normal or abnormal using a preset boundary threshold as an indicator.

[0070] 6 shows how the plotted characteristic data points are determined to be normal when they are inside the boundary threshold, and abnormal when they are outside the boundary threshold, with a substantially circular closed curve (shown by a thick line) having the origin (average value of each value) of a two-dimensional coordinate system with the vertical axis (y-axis) being the ratio of the average operation times of the pumps (T(PA) / T(PB)) and the horizontal axis (x-axis) being the average operation time of pump PB (T(PB)). Note that the characteristic data points may be plotted in a two-dimensional coordinate system with the vertical axis (y-axis) being the average operation time of one pump PA and the horizontal axis (x-axis) being the average operation time of the other pump PB.

[0071] As a machine learning device provided in the diagnostic processing unit 48, in addition to a one-class support vector machine, a computer that executes algorithms such as the LOF (local outlier factor) method, the IF (Isolation Forest) method, and the RC (Robust Covariance) method can be used.

[0072] By carrying out machine learning, a normal data space in which normal measurement data (training data) is mapped, that is, an internal space of boundary thresholds, is generated in the mapping space (feature space) shown in Fig. 6. The example in Fig. 6 shows boundary thresholds obtained by learning past data of pairs of characteristic values ​​obtained from manhole pump equipment in multiple locations where two pumps are installed, as learning data.

[0073] Each time a characteristic data point is primarily diagnosed as abnormal, the diagnostic processing unit 48 multiplies a predetermined abnormality reference value Vnb by a weighting coefficient W based on the distance between the position of each characteristic data point and the boundary threshold value based on the following formula, and adds the result, and each time each characteristic data point is diagnosed as normal in the primary diagnosis, calculates a cumulative evaluation value V by subtracting a predetermined normal return evaluation value Vpb, and makes a final diagnosis of the normality or abnormality of the manhole pump apparatus based on the cumulative evaluation value. The distance between the position of the characteristic data point and the boundary threshold value refers to the length of the normal line from the boundary threshold value that passes through the characteristic data point. V=Vnb×W-Vpb In this embodiment, Vnb = 1, and it is set that Vnb < Vpb < Wmax × Vnb.

[0074] Even if an abnormality is diagnosed in the primary diagnosis, there are mild abnormalities that are later diagnosed as normal, and there are also abnormalities that are continuously diagnosed as abnormal and eventually lead to a serious failure. Even in such cases, each time an abnormality is diagnosed in the primary diagnosis, a value obtained by multiplying the distance between the characteristic data point and the boundary threshold by a weight coefficient based on a predetermined abnormality reference value is added, and each time a normal diagnosis is made, a predetermined normal recovery evaluation value is subtracted to calculate a cumulative evaluation value, and based on the cumulative evaluation value, a final diagnosis of the normality or abnormality of the machinery and equipment is made. For example, a diagnosis taking into account the degree of abnormality such as an abnormality that requires maintenance in the near future becomes possible.

[0075] FIG. 7 shows the transition of the cumulative evaluation value. The cumulative evaluation value is primarily evaluated every predetermined period in which the characteristic data points are plotted. When an abnormality is determined, a value of Vnb × W is added to the initial value of 0, and when a normal determination is made, a value of Vpb is subtracted. When the cumulative evaluation value exceeds a predetermined threshold value, which is "10" in this example, a final abnormality determination is made, and this fact is transmitted to a mobile terminal or the like owned by the administrator. Note that the primary determination and the final determination are continuously made even after an abnormality is determined.

[0076] FIG. 8 illustrates a diagnostic map. The outside of the boundary threshold shown in the feature space shown in FIG. 6 is divided into eight regions, and each region is associated with one of the causes of abnormality. The diagnostic processing unit 48 diagnoses the cause of abnormality based on the region in which each characteristic data is plotted.

[0077] For example, in the example of FIG. 8, when a characteristic data point is plotted in region 2, only the pump PA that operates for a long time is abnormal; in region 4, an abnormality in the external environment such as an abnormality in the water level gauge; in region 5, abnormalities in both pumps that operate for a long time, an increase in the inflow amount of sewage, an abnormality in the external environment such as a blockage at the confluence of the discharge pipes; in region 8, only the pump PB due to long-time operation is diagnosed as abnormal.

[0078] In the above example, the average operation time of each pump, or the ratio of the average operation time of one pump to the average operation time of the other pump, is used as a pair of characteristic values. However, the pair of characteristic values ​​is not limited to these data, and can be set as appropriate to the operation time per operation of each pump, the ratio of the operation time per operation of one pump to the operation time per operation of the other pump, the natural logarithm of the ratio of the operation times of both pumps and the natural logarithm of the ratio of the number of operations, the current value, average current value, and ratio of the average current values ​​of both pumps during operation as shown in Figure 4(a).

[0079] [Specific example 2 of manhole pump diagnostic method] A second specific example will now be described. The normalization processing unit 46 is configured to normalize a group of n types (n is an integer greater than or equal to 2) of measurement data that indicate the characteristics of the manhole pump apparatus 10 sampled in a time series manner.

[0080] The machine learning device provided in the diagnostic processing unit 48 is configured to plot each of the normalized measurement data groups in an n-dimensional coordinate system, and perform a primary diagnosis of whether the manhole pump apparatus 10 is normal or abnormal based on on which side of a boundary threshold preset in the n-dimensional coordinate system the plotted measurement data group exists.

[0081] The diagnostic processing unit 48 is configured to multiply a predetermined abnormality standard value by a weighting coefficient based on the distance between each measurement data group and the boundary threshold value and add the result each time a primary diagnosis is performed, and to make a final diagnosis of whether the manhole pump apparatus 10 is normal or abnormal based on a cumulative evaluation value obtained by subtracting the predetermined abnormality standard value each time each measurement data group is diagnosed as normal in the primary diagnosis.

[0082] The outside of the boundary threshold set in the n-dimensional coordinate system is divided into a plurality of regions, and a diagnostic map is provided in which each region is associated with one of the causes of anomaly. The diagnostic processing unit 48 is configured to diagnose the cause of anomaly based on the region in which each measurement data group is plotted.

[0083] 9 shows a flow of a series of diagnostic processes executed by the monitoring device 40. When one day's worth of measurement data is received from each manhole pump apparatus 10 and stored in the database DB (SB1), feature amounts, which are a group of measurement data used to determine an abnormality, are extracted for each pump of each manhole pump apparatus 10.

[0084] It is determined whether a predetermined update condition is satisfied, and if it is satisfied (SB3, Y), the statistical data update processing unit 47 executes update processing of the statistics (SB4), and normalizes the feature quantity of the day to be judged based on the updated statistics, i.e., the average value and variance value (SB5).If the update condition is not satisfied (SB3, N), the previous statistics are used to normalize the feature quantity of the day to be judged (SA5).

[0085] Each normalized feature is input in turn into a machine learning device and a primary judgment is made based on the set boundary threshold (SB6, SB7). If an abnormality is detected, the cumulative evaluation value is increased taking into account the degree of abnormality (SB8), and if a normality is detected, the cumulative evaluation value is decreased (SB9).

[0086] A final abnormality determination is made as to whether the cumulative evaluation value calculated in this way exceeds a preset cumulative abnormality threshold value (SB10), and if it exceeds the cumulative abnormality threshold value, the cause of the abnormality is identified based on a diagnostic map that indicates a correlation with the abnormality cause in advance (SB11), and an alarm indicating an abnormal state together with the cause of the abnormality is sent to a mobile terminal or the like owned by the manager (SB12). The alarm notification is sent as an e-mail via a mailer provided in communication unit 41.

[0087] As shown in Figure 4(b), when the pump's conveying rate is low or when the amount of wastewater flowing into the manhole is large, the pump operation time is longer and the rate at which the water level falls while the pump is operating is smaller. For simplicity, the rate at which the water level falls while the pump is operating is referred to below simply as the "water level slope."

[0088] Such water level information and operation information stored in the memory unit 22 of each manhole pump device 10 is transmitted to the monitoring device 40 via the communication unit 24 and stored in the database DB via the data processing unit 42.

[0089] From such data, the data processing unit 42 extracts the "water level gradient" (=(HWL-LWL) / operating time) and the "current value" of each pump PA, PB when it is started as n types (n is an integer greater than or equal to 2, but in this embodiment, n=2) of measurement data groups (also called characteristic data groups) indicating the characteristics of the manhole pump device 10 sampled in a time series manner, that is, feature values, on a daily basis, and passes them on to the normalization processing unit 46. The pump current values ​​and manhole water level gradient, which are the measurement data groups, are correlated with each other when the pump is normal, and are measurement data that converge to a predetermined range.

[0090] The mean value μ and the variance σ, which are statistics for normalization generated by the statistical data update processing unit 47, 2 Based on this, the normalization processing unit 46 normalizes the feature quantities "gradient of water level" and "current value" using the formula (x-μ) / σ for the feature quantity x.

[0091] When the feature quantities “water level gradient” and “current value” normalized by the normalization processing unit 46 are input to the diagnostic processing unit 48, the normalized feature quantities are plotted on a two-dimensional coordinate system showing the current and water level gradient, respectively, and a primary diagnosis of normality or abnormality is made based on which side of a boundary threshold preset in the two-dimensional coordinate system the plotted measurement data group is on.

[0092] By normalizing the n types of measurement data sampled in a time series manner in a normalization step, the measurement data group that has been freed of the effects of the inherent bias of the manhole pump and the effects of the installation environment can be extracted as features. In the diagnosis, these features, i.e., the points representing the two types of measurement data groups, are plotted as multiple points in a time series in a two-dimensional coordinate system, and an appropriate initial diagnosis is made as to whether the manhole pump equipment is normal or abnormal using a preset boundary threshold as an indicator.

[0093] In FIG. 10, with the origin (average value of current and average value of water level gradient) in a two-dimensional coordinate system where the vertical axis is current and the horizontal axis is the water level gradient as the center, a circle with a predetermined radius (indicated by a thick line) is used as a boundary threshold. If the plotted feature quantity is located inside the boundary threshold, it is determined to be normal, and if the plotted feature quantity is located outside the boundary threshold, it is determined to be abnormal. This state is shown.

[0094] Similar to the above, by the machine learning device provided in the diagnosis processing unit 48, in the mapping space (feature space) shown in FIG. 10, a normal data space, that is, the internal space of the boundary threshold, is generated by mapping normal measurement data (training data). In the example of FIG. 10, the boundary threshold obtained by learning the feature data for one year of 100 manhole pump devices with two pumps installed as learning data is shown.

[0095] Each time the diagnosis processing unit 48 makes a primary diagnosis that the measurement data group is abnormal, based on the following mathematical formula, it multiplies the distance between the position of each measurement data group and the boundary threshold by a weight coefficient W based on a predetermined abnormal reference value Vnb and adds them. Each time each measurement data group is diagnosed as normal in the primary diagnosis, it subtracts a predetermined normal recovery evaluation value Vpb to calculate a cumulative evaluation value V, and finally diagnoses whether the manhole pump device is normal or abnormal based on the cumulative evaluation value. V = Vnb×W - Vpb In this embodiment, Vnb = 1, and it is set that Vnb < Vpb < Wmax×Vnb.

[0096] Even if it is diagnosed as abnormal in the primary diagnosis, there are mild abnormalities that may be diagnosed as normal later, and there are also abnormalities that will continue to be diagnosed as abnormal and eventually lead to serious failures. Even in such cases, each time it is diagnosed as abnormal in the primary diagnosis, a value obtained by multiplying a predetermined abnormal reference value by a weight coefficient based on the distance between the measurement data group and the boundary threshold is added, and each time it is diagnosed as normal, a predetermined normal recovery evaluation value is subtracted to calculate a cumulative evaluation value, and based on this cumulative evaluation value, a diagnosis that takes into account the degree of abnormality, such as an abnormality that requires maintenance in the near future, becomes possible for the normal or abnormal of the mechanical equipment.

[0097] As explained in Fig. 7, the cumulative evaluation value is subjected to a primary evaluation for each operation count, and if it is judged to be abnormal, the value of Vnb × W is added to the initial value of 0, and if it is judged to be normal, the value of Vpb is subtracted. When the cumulative evaluation value exceeds a predetermined threshold, a final abnormality is judged, and a message to that effect is sent to a mobile terminal owned by the manager. Note that the primary and final judgments are continuously made even after an abnormality is judged.

[0098] An example of a diagnostic map is shown in Fig. 11. The outside of the boundary threshold shown in the feature space shown in Fig. 10 is divided into eight regions, and each region is associated with one of the causes of anomaly. The diagnostic processing unit 48 diagnoses the cause of the anomaly based on the region in which each piece of feature data is plotted. For example, in the example of Fig. 11, when feature data is plotted in region 3, a foreign object entrapment anomaly is diagnosed, when feature data is plotted in region 5, a water pressure leakage anomaly is diagnosed, and when feature data is plotted in region 8, an air entrapment anomaly is diagnosed.

[0099] In the above example, the measurement data group is a combination of the pump drive current value and the water level inclination in the manhole, but the measurement data group is not limited to these data and can be set appropriately. For example, instead of the water level inclination, the operation time of the pump per start, which is correlated with the water level inclination, may be used as the measurement data.

[0100] As described above, the manhole pump diagnostic method according to the present invention is a method for diagnosing a manhole pump that is installed in a manhole and repeatedly starts and stops, and includes a sampling step of chronologically sampling measurement data indicating the characteristics of the manhole pump to obtain characteristic data for diagnosis, a normalization step of normalizing the characteristic data obtained in the sampling step based on predetermined statistical data, and a diagnosis step of diagnosing whether the manhole pump is normal or abnormal based on the characteristic data normalized in the normalization step and a preset boundary threshold value.

[0101] The method further includes a statistical data update step for updating, when a predetermined update condition is satisfied, predetermined statistical data based on the characteristic data before normalization that has been sampled during a previous predetermined period and diagnosed as normal in the diagnosis step.

[0102] In addition, the statistical data update step can also be configured to update and process specified statistical data based on a specified number or more of pre-normalized characteristic data sampled during the immediately preceding period and diagnosed as normal in the diagnosis step when a specified update condition is met.

[0103] It is preferable that the measurement data includes the operation time and number of times of the manhole pump.

[0104] The measurement data preferably includes the operation time of the manhole pump, the current value, and the water level in the manhole.

[0105] The predetermined update condition is preferably the passage of a predetermined time or the occurrence of an external trigger, and the boundary threshold is preferably maintained at an initially set value regardless of whether or not the statistical data update step is performed.

[0106] Each of the above-described embodiments is merely an example of the present invention, and the technical scope of the present invention is not limited by the description. It goes without saying that the specific configurations of each part, including the pump and water level gauge, and the threshold values ​​set for abnormality determination can be appropriately modified and designed within the scope of the effect of the present invention. [Explanation of symbols]

[0107] 10: Manhole pump equipment PA,PB: Pump 18,19: Water level gauge 21: Control unit 22: Storage part 24: Communications Department 30: Mobile devices 40: Monitoring device 41: Communications Department 42: Data processing section 44: Diagnostics Department 46: Normalization processing section 47: Statistical data update processing unit 48: Diagnostic processing unit

Claims

1. A method for diagnosing a manhole pump that is installed in a manhole and repeatedly starts and stops, comprising the steps of: A sampling step of obtaining diagnostic characteristic data by sampling measurement data indicating characteristics of the manhole pump in a time series manner; a normalization step of normalizing the characteristic data obtained in the sampling step based on predetermined statistical data; a diagnosis step of diagnosing whether the manhole pump is normal or abnormal based on the characteristic data normalized in the normalization step and a preset boundary threshold value; a statistical data update step of updating the predetermined statistical data based on characteristic data before normalization that has been diagnosed as normal in the diagnosis step during a predetermined period immediately preceding the normalization step when a predetermined update condition is satisfied; Equipped with A manhole pump diagnostic method in which the specified update condition is set to an external trigger that is generated in response to a predetermined time period corresponding to seasonal fluctuations in the amount of sewage inflow into the manhole, or in response to changes in the surrounding environment that cause an increase or decrease in the sewage source.

2. A method for diagnosing a manhole pump that is installed in a manhole and repeatedly starts and stops, comprising the steps of: A sampling step of obtaining diagnostic characteristic data by sampling measurement data indicating characteristics of the manhole pump in a time series manner; a normalization step of normalizing the characteristic data obtained in the sampling step based on predetermined statistical data; a diagnosis step of diagnosing whether the manhole pump is normal or abnormal based on the characteristic data normalized in the normalization step and a preset boundary threshold value; a statistical data update step of updating the predetermined statistical data based on a predetermined number or more of pre-normalized characteristic data diagnosed as normal in the diagnosis step during a immediately preceding period when a predetermined update condition is satisfied; Equipped with A manhole pump diagnostic method in which the specified update condition is set to an external trigger that is generated in response to a predetermined time period corresponding to seasonal fluctuations in the amount of sewage inflow into the manhole, or in response to changes in the surrounding environment that cause an increase or decrease in the sewage source.

3. 3. The method for diagnosing a manhole pump according to claim 1, wherein the measurement data includes an operating time and an operating frequency of the manhole pump.

4. 3. The method for diagnosing a manhole pump according to claim 1, wherein the measurement data includes an operating time of the manhole pump, a current value, and a water level in the manhole.

5. 5. The manhole pump diagnostic method according to claim 1, wherein the boundary threshold is maintained at an initially set value regardless of whether or not the statistical data updating step is performed.

6. A diagnostic device for a manhole pump that is installed in a manhole and repeatedly starts and stops, a normalization processing unit that normalizes diagnostic characteristic data obtained from measurement data indicating characteristics of the manhole pump sampled within a predetermined period based on predetermined statistical data; a diagnosis processing unit that diagnoses whether the manhole pump is normal or abnormal based on the characteristic data normalized by the normalization processing unit and a preset boundary threshold value; a statistical data update processing unit that updates the predetermined statistical data based on characteristic data before normalization that has been diagnosed as normal by the diagnosis processing unit during a predetermined period immediately preceding the normalization processing when a predetermined update condition is satisfied; Equipped with A manhole pump diagnostic device in which the specified update condition is set to an external trigger that is generated in response to a specified time period corresponding to seasonal fluctuations in the amount of sewage inflow into the manhole, or in response to changes in the surrounding environment that cause an increase or decrease in the sewage source.

7. A diagnostic device for a manhole pump that is installed in a manhole and repeatedly starts and stops, a normalization processing unit that normalizes diagnostic characteristic data obtained from measurement data indicating characteristics of the manhole pump sampled within a predetermined period based on predetermined statistical data; a diagnosis processing unit that diagnoses whether the manhole pump is normal or abnormal based on the characteristic data normalized by the normalization processing unit and a preset boundary threshold value; a statistical data update processing unit that updates the predetermined statistical data based on a predetermined number or more of pre-normalized characteristic data diagnosed as normal by the diagnosis processing unit during a immediately preceding period when a predetermined update condition is satisfied; Equipped with A manhole pump diagnostic device in which the specified update condition is set to an external trigger that is generated in response to a specified time period corresponding to seasonal fluctuations in the amount of sewage inflow into the manhole, or in response to changes in the surrounding environment that cause an increase or decrease in the sewage source.

Images