JP2016148963A - Failure sign diagnostic system and failure sign diagnostic method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent wrong diagnosis occurring by a use state of a movable body, environmental change or maintenance work.SOLUTION: A failure sign diagnostic system comprises: an equipment state prediction device configured to estimate an equipment deterioration state on the basis of history of an operation signal of equipment; an equipment state calculation device configured to calculate a deterioration state using the latest input operation signal of the equipment; and an inspection timing estimation device configured to estimate inspection timing on the basis of the deterioration state (prediction value) estimated by the equipment state prediction device and the deterioration state (current value) calculated by the equipment state calculation device. The inspection timing estimation device selects either inspection timing determination with the prediction value and inspection timing determination with the current value, on the basis of the difference (prediction error) between the predication value and the current value.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a failure sign diagnosis that estimates the state of a device and displays an inspection time.

  The introduction of state monitoring and maintenance that estimates the equipment state from the equipment operation signal during equipment operation and sets the inspection time according to the equipment state is progressing. For example, by analyzing the data stored in time series to detect a time series change in performance degradation of the heat source unit from the start of use, and estimating the degree of deterioration from the start of use based on the change over time, predetermined There has been proposed a heat source unit management device that calculates a maintenance period from the estimated degradation threshold and the estimated degree of degradation (Patent Document 1).

JP 2003-185232 A

However, in Patent Document 1, when the environmental status or usage status of a device changes, there is a possibility that the device status cannot be estimated correctly. In addition, even if the equipment characteristics have changed due to maintenance work (adjustment / replacement), there is a risk that the equipment status cannot be estimated correctly.

  In order to solve this problem, in the present invention, a device state prediction device that estimates a device deterioration state based on a history of device operation signals, and a device state that calculates a deterioration state from the most recently input operation signal of the device. An inspection time estimation device for estimating an inspection time based on a deterioration state (predicted value) estimated by the device state prediction device and a deterioration state (current value) calculated by the device state calculation device; The inspection time estimation device selects either the inspection time determination based on the prediction value or the inspection time determination based on the current value based on the difference (prediction error) between the prediction value and the current value.

  According to the present invention, it is possible to improve the estimation accuracy of the device state when the use state or the environmental state changes or when the device state changes due to maintenance work (adjustment / exchange). In addition, when the inspection work time is determined based on the estimation result of the equipment state, a less maintenance plan can be proposed.

An example of the block diagram of the failure sign diagnostic system to which this invention is applied. An example of the calculation method of the apparatus state prediction apparatus 101. Another example of a calculation method of the device state prediction apparatus 101. An example of the calculation method of the apparatus state calculation apparatus 102. An example of an inspection time estimation method based on predicted values. An example of a method for estimating the inspection time based on the current value. An example of the flowchart for implement | achieving this invention. An example of a time series graph display. Another example of the flowchart for implement | achieving this invention. An example of a record display. An example of the block diagram of the failure sign diagnostic system in Example 2. FIG. An example of an estimation accuracy calculation method. An example of the figure which shows the relationship between the estimated value of inspection time, and a predetermined value. An example of the block diagram of the failure sign diagnostic system in Example 3. FIG. 12 is another example of a flowchart for realizing the third embodiment.

  Embodiments of the present invention will be described with reference to the drawings. In the embodiment described below, a door is described as an example of a specific target device, but the application target of the present invention is not limited to this.

FIG. 1 is an example of a functional block diagram constituting the failure sign diagnosis system 100 of the present invention. A door opening / closing command signal is input from a door control device (not shown) to the door that is the target device, and the door opens and closes in response to the command signal. The door is provided with a sensor for monitoring the opening and closing operation. The failure sign diagnosis system 100 includes a device state prediction device 101, a device state determination device 102, and an inspection time determination device 103. A device operation signal including a door opening / closing command signal output from the control device and a door opening / closing detection signal measured by a sensor is input to the device state prediction device 101 and the device state determination device 102. The device state prediction apparatus 101, for example, based on a device operation signal input before the current time, for example, a door open / close command signal (“open” command or “close” command) or a door open / close operation (“open” operation or “ The door opening / closing operation time is calculated from the completion detection signal of the “closed” operation, and an environmental information signal such as ambient temperature that affects the opening / closing time is taken into account, and the predicted value of the device deterioration state at the current time is calculated from the history data. Here, the predicted value generated by the device state prediction apparatus 101 is obtained by calculating an estimated value of the current device state from the device operation signal input before the current time. On the other hand, the device state determination apparatus 102 calculates the current value of the device deterioration state based on the device operation signal input at the current time. Here, the current value generated by the device state determination device 102 is obtained by calculating the current device deterioration state from the device operation signal input at the current time. The inspection time determination device 103 calculates the inspection time of the equipment based on the predicted value and the current value of the equipment deterioration state.
As an actual estimation method in the device state prediction apparatus 101, a physical state model based on a state equation as shown in FIG. 2 or a statistical model as shown in FIG. 3 may be used. 2 and 3, u (n-1) is a device operation signal input before the current time, x (n) is an internal state at the current time, and f0, f1, and f2 are arbitrary functions ( For example, f0 is a noise removal function, f1 is a periodic function, and f2 is a trend function.) Ye (n) represents a predicted value at the current time.
Hereinafter, an example of a specific calculation method for calculating the prediction value by the device state prediction apparatus 101 using the statistical model shown in FIG. 3 will be described below. First, by using the noise removal function f0, the past door opening / closing time is calculated based on the history information of the past device operation signal, and the door opening / closing time x (n-1) is calculated by removing the operation variation by moving average. . A predicted value ye (n) is calculated from this x (n-1) using the periodic function f1 and the trend function f2. Here, the periodic function f1 is a function for calculating the influence of seasonal variation, and for example, a periodic function such as a sin function that varies with a one-year cycle is set. On the other hand, the trend function f2 is a term corresponding to the deterioration trend, and a linear equation such as f2 (n) = a * x (n) + b is used. The coefficients of the functions of f1 and f2 are set by a fitting means such as a least-squares method using past opening / closing operation data. With the above calculation method, it is possible to construct a model that predicts the current time or future device deterioration state from past door opening / closing history data.
On the other hand, a static model as shown in FIG. 4 is used for calculating the current value in the device state determination apparatus 102. In the static model, a function f3 for calculating the device deterioration state based on the device operation signal at the current time is used. In the function f3, the door opening / closing time is calculated from the door opening / closing operation command signal and the door opening / closing detection signal. Specifically, based on the device operation signal input at the current time, the door opening / closing time taken from the input of the “open” door opening / closing command signal to the completion of the door opening operation is calculated.
The inspection time determination method by the inspection time determination device 103 will be described with reference to FIGS. 5 and 6, the device deterioration state (door opening / closing time) is shown on the vertical axis, and the time is shown on the horizontal axis. Here, as the equipment deterioration state, it is desirable to select an index having a correlation with the equipment deterioration. In this embodiment, the door opening / closing time is used as an index. FIG. 5 shows a method for calculating the inspection time based on the predicted value of the equipment deterioration state. The inspection time determination device 103 estimates the transition of the future equipment deterioration state indicated by the solid line in FIG. 5 using the above-described statistical model or physical model. The time from the time at which the transition of the future device deterioration state intersects with a preset threshold value for determining the device deterioration state to the current time is calculated as the inspection time. Note that the “current value” in FIG. 5 is a value calculated by the device state determination apparatus 102 by the method shown in FIG. 4, and the “prediction error” used in the present invention is the “current value” at the current time as shown in FIG. It is the difference between the “current value” and the “predicted value”.
In the present invention, as will be described later, when this prediction error is smaller than a predetermined threshold value, the inspection time shown in FIG. 5 is displayed as an output of the failure sign diagnosis system 100. FIG. 6 shows a method for calculating the inspection time in consideration of the “current value”. Here, the inspection time is calculated from the point where the transition of the future equipment deterioration state (dotted line in FIG. 6) generated by linear prediction based on the “current value” for the past three times intersects with the determination threshold. Here, linear interpolation is used for the sake of simplicity, but a different complementing method such as a spline may be used as the method for calculating the inspection time based on the “current value”.
The inspection time based on the “predicted value” explained so far is based on the past history, so when there is an environment and usage that has not been experienced in the past, or when maintenance work (equipment adjustment / part replacement) is performed. May estimate the wrong inspection time. Therefore, in the present invention, the “current value” that is the device degradation state at the current time is generated and compared with the “predicted value” at the current time generated based on the past history. Next, if the prediction error, which is the difference between the predicted value and the current value, is large, it is determined that the inspection time estimated based on the predicted value is incorrect, and based on the current value calculated from the device operation signal at the current time. To estimate the inspection time. As a result, it is possible to prevent the inspection timing from being erroneously estimated due to the influence of past data even if an environmental situation or specification situation that has not occurred in the past occurs or maintenance work is performed. In other words, the prediction error that is the difference between the predicted value of the current time generated based on the past history of the device operation signal obtained in the past and the current value based on the device operation signal of the current time is less than the predetermined threshold value. If it is small, the inspection time is presented based on the past history of the device operation signal obtained in the past, and if the prediction error is larger than the predetermined threshold, the device for the past several times from the current time or the current time The inspection time is presented based on the operation signal. If the prediction error is small, it is considered that an unexpected situation change that has not been accumulated in the past history has occurred, so the inspection time based on the predicted value based on the past history is presented, and vice versa. When the prediction error is large, it is considered that an unexpected situation change that has not been accumulated in the past history has occurred, so the inspection time is presented based on the latest device operation signal. Therefore, even when unexpected changes occur, it is possible to reduce the influence of past history data on the judgment of the inspection time. For example, maintenance personnel attach grease to the door rail, Even when adjustments such as changing the set value of the open / close time are performed, an appropriate inspection time can be presented.

  FIG. 7 is an example of a control flowchart in the inspection time determination device 103 for realizing the present invention. In step S701, a prediction error is calculated from the aforementioned predicted value and current value. In step S702, it is determined whether or not the prediction error is smaller than the predetermined threshold value compared with a predetermined threshold value. If the prediction error is smaller, it is determined that the inspection time based on the predicted value is correct, and the process proceeds to step S703. If the prediction error is greater than or equal to the predetermined value, it is determined that the inspection time based on the prediction value is incorrect, and the process proceeds to step S704. Step S703 is a normal process, and the inspection time calculated based on the predicted value is displayed. One step S704 is processing at the time of abnormality, and displays the inspection time calculated based on the current value, and proceeds to step S705 to facilitate the investigation of the cause and displays a time series graph.

  FIG. 8 shows an example of a time-series graph display in step S705. FIG. 8a is a time series graph of the prediction error. This prediction error is calculated from the difference between the current value and the predicted value of the device deterioration state (door opening / closing time) shown in FIG. 8b. FIG. 8c shows a device operation signal for calculating the current value. The above is the door control command value, and the device deterioration state (door opening / closing time) is calculated from the time until the state value changes after the door control command rises, and the device deterioration state at the current time becomes the current value. The time series graphs shown in Fig. 8a, Fig. 8b, and Fig. 8c are effective in analyzing the cause of the increase in the prediction error, and present this information when the prediction error exceeds a predetermined threshold. Thus, it can be easily confirmed whether or not the inspection time estimation based on the current value has been properly performed, and the user can set the maintenance plan more reliably. Alternatively, as described below, not only the time series graph of operation signals but also maintenance work records and device self-diagnosis records may be displayed simultaneously.

FIG. 9 is another example of a control flowchart in the inspection time determination device 103 for realizing the present invention. The difference from FIG. 7 is step S905. Here, instead of displaying the time series graph in step S704 in FIG. 7, the maintenance work record of the most recently performed maintenance work and the self-diagnosis record most recently determined to be abnormal are displayed as shown in FIG. In FIG. 10a, date and time, person in charge, work code, and in this example, it is possible to easily confirm whether a security work affecting the door operation time is not performed as a maintenance work record. On the other hand, FIG. 10b shows date and time, location of occurrence, error code, error content, work code (link to maintenance work record), and the like as self-diagnosis records that are diagnosis results of the self-diagnosis function of the device. For example, if the opening / closing time of the door is constantly calculated by the self-diagnosis function and an abnormal code is output when a preset threshold value is exceeded, it is understood that a door abnormality has occurred. If the user can refer to the information shown in FIGS. 10a and 10b, it can be determined whether the inspection time presented by the system is appropriate. In this example, the system requests that the inspection time be advanced. However, it can be determined that this is a false alarm due to maintenance work. Or, despite the self-diagnosis result shown in FIG. 10b, when there is no work corresponding to FIG. 10a, it is possible to set a work plan so that inspection and adjustment can be performed quickly.

An embodiment in which maintenance work can be performed more reliably before a device failure occurs will be described with reference to FIGS. 11 to 13. As in the first embodiment, the failure sign diagnosis system 200 according to the present embodiment includes the device state prediction device 101, the device state determination device 102, and the inspection time determination device 203, and the inspection time determination device 203 outputs the inspection time of the device. However, if the output estimation accuracy of the inspection time can be presented to the user, the user can make a more reliable maintenance work plan. Therefore, the failure sign diagnosis system 200 of the present embodiment further includes an estimation accuracy determination device 204, which is based on the prediction value output from the device state prediction device 101 and the prediction error obtained from the inspection time determination device. Then, the estimated accuracy of the inspection time output from the inspection time determination device is calculated and output to the outside.
An example of a method for calculating the estimated accuracy of the inspection time in the estimated accuracy determination device 204 will be described with reference to FIG. Here, the estimation accuracy is calculated using the moving average from the prediction error and the prediction value, and this estimation accuracy is presented to the user together with the inspection time. With this estimated accuracy, if the estimated accuracy is poor, for example, if the estimated accuracy is poor, the user can set up an inspection work plan so that the inspection is performed earlier than the inspection time presented by the failure sign diagnosis system. Inspection work can be performed before troubles occur, and train operations can be more reliably prevented from being delayed or suspended.
FIG. 13 shows an example of a method for setting a predetermined threshold for comparison with the prediction error shown in step S702 of FIG. 7 or step S902 of FIG. In FIG. 12, the smaller the estimated value of the inspection time, the smaller the predetermined threshold for comparison with the prediction error. This reflects the fact that the closer the inspection time is to the current time, the more accurate the estimation is required. In this way, even when the inspection time approaches the current time, the predetermined threshold value for comparing with the prediction error in the inspection time determination device is set to be smaller as the estimated value of the inspection time becomes smaller. A more reliable inspection time can be presented.

With reference to FIGS. 14 and 15, an embodiment will be described in which the failure sign diagnosis system described in the first embodiment is applied to verify whether the maintenance work is properly performed. FIG. 14 shows an example of a functional block diagram constituting the failure sign diagnosis system 300 in the present embodiment. In this embodiment, in addition to the devices shown in the first embodiment, an input device 306 and a maintenance work monitoring device 305 are provided, and it is possible to determine whether or not the maintenance work has been properly performed. It is an embodiment. The maintenance work monitoring device 305 in this embodiment is a device that performs maintenance work based on maintenance work execution information input to the input device 306 by a maintenance worker or the like and information on the inspection time output from the inspection time determination device 103. It is determined whether or not the deterioration state has been improved. If the device deterioration state has not improved due to maintenance work, a warning is output to the outside.
FIG. 15 is an example of a control flowchart of the maintenance work monitoring apparatus 305 in the present embodiment. In step S1501, it is determined based on the maintenance work execution information whether or not maintenance work such as device adjustment and parts replacement has been performed. If the maintenance work has been performed, the process proceeds to step S1502. In step S1502, the inspection time before the maintenance calculated by the inspection time determination device 103 before the maintenance work is performed, the inspection time after the maintenance calculated by the inspection time determination device 103 after the maintenance work is performed, Based on the above, it is determined whether or not the inspection time after the maintenance is performed is longer than that before the maintenance is performed. If the inspection time after the maintenance is not longer than that before the maintenance is performed, the process proceeds to step S1503. Normally, if the maintenance work is properly performed, the device deterioration state is improved after the maintenance than before the maintenance. Therefore, the calculated inspection time is longer after the maintenance than before the maintenance. Therefore, if the calculated inspection time is not longer than that before the maintenance, it can be determined that there is a high possibility that the maintenance work is not properly performed. In step S1503, a warning is output from the maintenance work monitoring device 305 together with the inspection time of the corresponding equipment generated by the inspection time determination device 103. According to the present embodiment, it is possible to grasp the case where the user is not properly performing maintenance work on the device, and it is possible to quickly plan the re-inspection time.
In this embodiment, the calculated “inspection time” is compared before and after the maintenance work to determine whether or not the maintenance work is properly performed, but instead of the “inspection time”. The “deterioration state of the device” may be compared before and after the maintenance work to determine whether or not the maintenance work is being performed appropriately. In this case, in step S1502, based on the current value of the device deterioration state generated by the device state determination device, the current value after the maintenance work is lower than the current value before the maintenance work is performed. If it is determined that the maintenance work is properly performed, and the current value after the maintenance work is not lower than the current value before the maintenance work is performed, the maintenance work is being performed properly. If it is determined that there is not, the process proceeds to step S1503 to output a warning.

100, 200, 300: Predictive failure system
101: Equipment status prediction device
102: Device status calculation device
103, 203: Inspection time determination device
204: Estimated accuracy judgment device
305: Maintenance work monitoring device
306: Input device

Claims (8)

A device state prediction device that predicts a device deterioration state based on history information of device operation signals detected by a sensor;
A device state determination device for determining a device deterioration state from the operation signal most recently detected by the sensor,
An inspection time determination device that determines an inspection time based on a predicted value of the device deterioration state predicted by the device state prediction device and a current value of the device deterioration state determined by the device state determination device;
The inspection time determination device displays either an inspection time generated based on the predicted value or an inspection time generated based on the current value based on a prediction error that is a difference between the predicted value and the current value. A failure sign diagnosis system characterized by that.
In the failure sign diagnosis system according to claim 1,
When the prediction error is smaller than a predetermined threshold value, an inspection timing generated based on the predicted value is displayed.
In the failure sign diagnosis system according to claim 1 or 2,
When the prediction error is greater than or equal to a predetermined threshold value, the inspection timing generated based on the current value is displayed.
In the failure sign diagnostic system according to claim 3,
When the prediction error is equal to or greater than a predetermined threshold, the inspection time generated based on the current value is displayed, and at least one time series of the prediction error, the predicted value, the current value, and the operation signal A failure sign diagnosis system characterized by displaying data.
In the failure sign diagnostic system according to claim 3,
When the prediction error is equal to or greater than a predetermined threshold value, the inspection time generated based on the current value is displayed, and the maintenance work record of the most recently performed maintenance work or the self of the equipment that has recently been determined to be abnormal A predictive diagnosis system characterized by displaying a diagnostic record.
In the failure sign diagnosis system according to claim 1,
A failure sign diagnostic system, wherein the estimated accuracy of the inspection time is calculated based on the prediction error, and the estimated accuracy is displayed together with the inspection time.
In the failure sign diagnosis system according to claim 2,
The failure sign diagnosis system, wherein the predetermined threshold value is decreased as the estimated value of the inspection time approaches the time when the estimated value is calculated.
The failure sign diagnosis system according to claim 1 to 7,
Provided with an input device capable of inputting that the maintenance work of the device has been performed,
When information indicating that maintenance work has been performed is input to the input device, the inspection time generated before the maintenance is compared with the inspection time generated after the maintenance is performed. A failure sign diagnosis system that outputs a warning when the inspection time is not longer than the inspection time generated before maintenance.

Images