JP2003029818A - Failure diagnostic system and failure diagnostic program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a failure diagnostic system for allowing a manager to accurately grasp the defective state of each equipment in a package unit without bothering to go to check it. SOLUTION: A packet unit 10 is provided with a sound sensor 1 and a smell sensor 12. A failure diagnosing application server 16 collects the sound and smell data of each equipment from the sound sensor 11 and the smell sensor 12 through a network 14 in a batch, and obtains pattern data corresponding to each equipment by frequency-analyzing the data, and judges the defective state of each equipment by operating the matching processing with pattern data in a normal time preliminarily registered in a pattern data base 17, and notifies a computer 20 for monitoring of a manager of the judged result through the network 14. Thus, it is not necessary for the manager to go to check each equipment in the package unit 20, and it is possible to quickly facilitate countermeasures to the abnormal time by grasping the accurate state.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a failure diagnosis system and a failure diagnosis program for performing failure diagnosis on devices mixed in a package unit in a generator system or the like via a network.

[0002]

2. Description of the Related Art For example, in a generator system, there is a generator system in which devices such as a generator main body and auxiliary devices thereof are packaged and installed in one unit in order to save space. In a package unit in which such devices are mixed, a fault diagnosis for each device is usually performed by an administrator who regularly checks to see if there is any tendency for abnormality. When a device abnormality is found, it is artificially judged whether the device is replaced or not. Also,
In some cases, the measured value such as the temperature of the device is transmitted to a remote place, and the administrator monitors the trend of the measured value to artificially determine whether there is an abnormality.

[0003]

As described above, the current inspection method requires the administrator to periodically inspect each device in the package unit. Therefore, the cost and labor required for the inspection are very high, and if the abnormality of the device is detected in advance and the system stop due to the failure of the single device is to be avoided, the periodic inspection must be performed in a short cycle. Further, since it is artificially determined when to replace the device body or parts, there is a high possibility of making a mistake in the determination, and if a delay in response occurs, it may lead to failure.

The present invention has been made in view of the above points, and a failure diagnosis system capable of accurately ascertaining the failure state thereof even if an administrator does not inspect each device in a package unit. And to provide a failure diagnosis program.

[0005]

A failure diagnosis system of the present invention includes a package unit in which a plurality of devices are mixed, a server for performing a failure diagnosis for each of the devices in the package unit, and a diagnosis result from the server. A failure diagnosis system configured by a terminal on the administrator side and connected between the package unit and the server, and between the server and the terminal by a specific network. A sensor for detecting the state of each device is provided, and the server is
Data collecting means for collectively collecting data indicating the status of each device from the sensor via the network, and a failure for judging a failure status for each device based on the data collected by the data collecting means It is characterized by comprising a judging means and a notifying means for notifying the terminal of the judgment result by the failure judging means via the network.

According to the failure diagnosis system having such a configuration, the server collects data from the sensors installed in the package unit, and the failure state of each device in the package unit is based on the collected data. Can be individually judged and the judgment result can be provided to the administrator.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION A failure diagnosis system according to an embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram schematically showing a package unit which is a failure diagnosis target of this system.

The package unit 10 includes a device body 1
In addition to a and 1b, a large number of devices such as auxiliary devices 2a, 2b, and 2c are installed in a mixed manner. In the case of a generator system, the device main bodies 1a and 1b are the main bodies of the generator, and the auxiliary devices 2a, 2b and 2c correspond to pumps, valves, fans and the like. Such devices are packaged not only in generator systems but also in fuel cell systems and the like.

By the way, in the case where the devices are packaged, in the method of diagnosing a failure by installing a dedicated sensor for each device or part, there is a restriction on the installation space and a cost problem. Is not preferable. Therefore, in the present system, a sensor commonly used for each device in the unit is provided, and the sensor detects the sound and odor of each device.

In FIG. 1, 11 and 12 are a sound sensor for detecting a sound emitted by each device and an odor sensor for detecting an odor emitted by each device, which are commonly provided to each device. The sound sensor 11 and the odor sensor 12 are connected to a signal processing / transmission device 13 installed in the unit 10. The signal processing / transmission device 13 has a function of digitizing and transmitting the state signal (analog) of each device detected by the sound sensor 11 and the odor sensor 12.

FIG. 2 is a diagram showing the overall configuration of the failure diagnosis system of the present invention, and the dotted line in the figure shows the flow of data.

The signal processing / transmission device 13 installed in the package unit 10 is connected to the failure diagnosis application server 16 of the center 15 via the network 14. Although the Internet is assumed as the network 14, a LAN (Local Area Network)
Alternatively, a dedicated line such as a WAN (Wide Area Network) may be used.

The failure diagnosis application server 16 installed in the center 15 is a computer for realizing the failure diagnosis service of this system. The failure diagnosis application server 16 includes a pattern database 17 and an abnormality diagnosis database 18, and collects data from the sound sensor 11 and the odor sensor 12 in the package unit 10 so that the package unit 10 can be stored in the package unit 10 based on the data. The failure diagnosis of each device (1a, 1b, 2a to 2c) is performed. In the pattern database 17, pattern data when each device is in a normal state is registered in advance. Rules for abnormality diagnosis are registered in advance in the abnormality diagnosis database 18. The method of this abnormality diagnosis will be described later in detail with reference to FIG.

The result of the failure diagnosis by the failure diagnosis application server 16 is sent to the monitoring room 1 via the network 14.
9 monitoring computer (PC) 20. By receiving the failure diagnosis result at the monitoring computer 20, the administrator can grasp the failure state of each device in the package unit 10 while staying in the monitoring room 19.

FIG. 3 is a block diagram showing a circuit configuration of the signal processing / transmitting device 13 installed in the package unit 10.

The signal processing / transmission device 13 comprises a data collection unit 31, a low pass filter (LPF) 32, an A / D conversion unit 33, and a data transmission unit 34. Signals (analog) of sounds and odors of each device detected by the sound sensor 11 and the odor sensor 12 installed in the package unit 10 are collected by the data collecting unit 31, and low-frequency components are collected by the low-pass filter 32. It is removed and given to the A / D converter 33. Then, the data is converted into digital data by the A / D converter 33, and is collectively sent to the failure diagnosis application server 16 in the center 15 by the data transmitter 34 via the network 14.

The signal processing / transmission device 13 and the failure diagnosis application server 16 are always connected, and the status of each device detected by the sound sensor 11 and the odor sensor 12 is real time in the failure diagnosis application. It is provided to the server 16.

FIG. 4 is a block diagram showing the circuit configuration of the failure diagnosis application server 16 installed in the center 15.

The failure diagnosis application server 16 is
It is used as a failure diagnosis device of the present system, and is realized by a computer whose operation is controlled by reading a program recorded in a recording medium such as a magnetic disk.

As shown in FIG. 4, the failure diagnosis application server 16 includes a CPU 41 and a CPU 41.
43, R connected to the computer via the system bus 42
AM44, external storage device 45, input unit 47, display unit 4
8 and a communication unit 49. The CPU 41 controls the entire system, and reads a program stored in advance in the ROM 43 or a program provided by the recording medium 46 to execute a failure diagnosis process as shown in FIG. The ROM 43 stores various data including the control program of the CPU 41. RAM
Reference numeral 44 is used as a work memory and stores various information necessary for the present system.

The external storage device 45 comprises, for example, a magnetic disk device, a flexible disk device, an optical disk device, etc., and is used as an auxiliary storage device.
The pattern database 17 and the abnormality diagnosis database 18 described above are provided in the external storage device 45. It is also possible to record a program for realizing the present invention in the recording medium 46 and read it in the external storage device 45. The method of providing the program is not limited to the recording medium 46, and can be provided from the outside via a transmission medium such as the Internet.

The input unit 47 is composed of an input device such as a keyboard and a mouse, and is used when data is input or instructions are given. The display unit 48 is a CRT (Cathode-ray tu
be) and an LCD (Liquid Crystal Display), and is used when displaying data. The communication unit 49 performs communication processing according to a predetermined protocol, and here, realizes data communication between the center 15 and the package unit 10 or the monitoring room 20 via the network 14.

Although the structure of the monitoring computer 20 provided in the monitoring room 19 is not particularly shown, the basic structure is the same as that shown in FIG. 4 and includes the components as a general-purpose computer.

Next, the operation of this system will be described.

Each device (1a, 1b, 2a to 2c) in the package unit 1 of FIG. 1 emits its own sound or odor. These sounds and odors are detected by the sound sensor 11 and the odor sensor 12 which are common to the respective devices. Sound sensor 11
The data (analog) indicating the sound and odor of each device detected by the odor sensor 12 and the odor sensor 12 are collected by the signal processing / transmission device 13 and then converted into digital data to be converted into the center data.
It is transmitted to the failure diagnosis application server 16 in the network via the network 14. The data transmitted to the application server 6 includes the package unit 10
It will contain the sound and odor information of all devices inside.

The failure diagnosis application server 16 receives the data from the signal processing / transmitting device 13 to perform the failure diagnosis processing as described below, and the diagnosis result is sent to the monitoring computer in the monitoring room 19. 20 to provide.

FIG. 5 shows the failure diagnosis application server 1.
6 is a flowchart showing the operation of the failure diagnosis processing executed in 6.

When the application server 6 receives the data including the sound and odor information of all the devices (step S11), the received data is subjected to frequency analysis, and the sound or odor pattern data corresponding to each device is obtained. Is calculated (step S12). That is, since each device has a different sound and odor pattern, frequency and sound analysis is performed to obtain the sound and odor pattern data for each device, and the data is stored in a predetermined file in the RAM 44.

Then, pattern matching processing is performed between the sound and odor pattern data saved in the above file and the normal sound and odor pattern data registered in the pattern database 5 in advance to determine whether or not there is an abnormality in each device. To judge individually. For pattern matching, for example, a non-linear model such as a neural network is adopted to realize highly accurate matching processing (step S13). The state at this time is shown in FIG.

FIG. 6 is a diagram for explaining the pattern matching process.

The sound data 51 and the odor data 52 obtained in time series from the signal processing / transmitting device 13 in the package unit 10 include all the devices in the unit 10 (devices 1a, 1b, 2a in the example of FIG. 1). 2c) is included. Reference numerals 51 and 52 in the figure show the data waveforms. For such sound data 51 and odor data 52,
Fourier transform as shown in equations (1) and (2) is performed.

[0033]

[Equation 1]

In the above (1), Xn is a sound / smell data value, and n is the number of samplings (0 to N). In the above (2), Xk is a sound / odor spectrum value, k
Is the frequency. Also, exp is a Fourier function and J is a complex number. Although a more complicated conversion formula is actually used, since frequency analysis by Fourier transform is known, only a basic formula is shown here, and a detailed description will be omitted.

Reference numerals 53 and 54 in the figure show the sound spectrum and the odor spectrum after the Fourier transform, and include a plurality of patterns A, B and C. This pattern A, B, C
Corresponds to each device. That is, pattern A corresponds to the sound data of the first device, pattern B corresponds to the sound data of the second device, and pattern C corresponds to the third.
It corresponds to the sound data of the device. This also applies to the odor data.

By comparing the patterns A, B, C corresponding to the respective devices obtained in this way with the normal patterns A, B, C registered in the pattern database 17 in advance, the deviation (difference) The degree of.

In the example of FIG. 6, the signal of each data is shown as an analog waveform, but in reality, it is all processed digitally. Regarding the odor data, a method is adopted in which the pattern of each element that constitutes the odor is obtained for each device and the pattern is matched with the normal pattern.

Returning to FIG. 5, the failure diagnosis application server 16 shifts (different) between the pattern data of each device obtained from the received data and the pattern data at the normal time.
It is determined whether the degree of is within the normal range (step S14). In this case, a comprehensive judgment is made by combining both the sound and odor states. As a result, if it is within the normal range, that is, if the deviation is within the preset allowable range (Yes in step S14), it is determined that the device is in the normal state (step S15). Is notified to the monitoring computer 20 of the administrator via the network 14 (step S16).

On the other hand, if it exceeds the normal range, that is, if the deviation exceeds the preset allowable range (No in step S14), it is determined that the device is in an abnormal state, and the degree of the abnormality is determined. Judge (Step S1
7) Based on the abnormality determination rule registered in the abnormality diagnosis database 18, the administrator's monitoring computer 20 is notified via the network 14 of the degree of abnormality at that time (step S18). .

As a rule for the above abnormality determination, the degree of deviation between the collected pattern data and the normal pattern data is quantified, thresholds are set in several stages, and a stepwise diagnosis result is announced to the user. For example, if there is not much deviation in the pattern data, "just in case
Please check the equipment "or" Please replace the filter of the equipment within one week ", and if the deviation is large," Stop the system and check the ***.
Notify the administrator of the device abnormality including specific countermeasure information such as "Please replace it as soon as possible". The notification method may be by voice in addition to the message display.

Further, when making an abnormality determination, more accurate diagnosis result can be obtained by considering not only the size of the deviation from the normal pattern but also the direction of the deviation.

Further, not only in the event of an abnormality, but information about a future maintenance plan may be notified to the manager according to the current state of the equipment. The future maintenance plan defines the replacement timing and maintenance method of the main body and parts that should be performed after a predetermined period from the current state of the equipment in order to improve the safety of the equipment. It is registered in advance in the abnormality diagnosis database 18.

As described above, in the present system, the failure diagnosis application server 16 collects data from the sound sensor 11 and the odor sensor 12 sensor installed in the package unit 10, and the data in the package unit 10 is collected based on the data. Since the failure state of each device is individually determined, it is not necessary to provide a dedicated sensor for each device or part, and the facility investment can be reduced accordingly.

Further, the failure diagnosis application server 1
Since the result determined in 6 is provided to the administrator via the network 14 such as the Internet, the administrator can stay in the monitoring room 19 and have a failure of each device without directly inspecting each device. The state can be grasped, and moreover, an accurate judgment result can be obtained because it does not depend on artificial inspection. In this case, since the network 14 can always grasp the state of each device in real time, it becomes possible to promptly respond when an abnormality is found.

Further, since the maintenance plan regarding the replacement time of the device body and parts is provided based on the current state of the device, there is an advantage that the safety of each device can be enhanced.

In the above embodiment, both the sound and the odor of each device are detected, and the presence / absence of abnormality is determined based on the detected data. However, the determination is made based on either one of the data. It's okay.

Further, the material of the failure judgment is not limited to the sound and smell emitted by each device, but any information that can judge the failure state of each device can be applied.

Further, the methods described in the above-described embodiments are, as a program that can be executed by a computer, for example, a magnetic disk (flexible disk, hard disk, etc.), an optical disk (CD-RO).
It is also possible to write the data on a recording medium such as M, DVD, etc.) or a semiconductor memory and apply it to various devices, or to transmit the program itself via a transmission medium such as a network and apply it to various devices. A computer that realizes the present apparatus reads the program recorded in the recording medium or the program provided via the transmission medium, and the operation is controlled by the program to execute the above-described processing.

[0049]

As described above, according to the present invention, a failure diagnosis service using a network can be realized. As a result, the administrator does not have to go to inspect each device in the package unit, and because it does not rely on artificial inspection, it can grasp the accurate state, and the network can check the state of each device in real time. It is possible to respond quickly in the event of an abnormality.

[Brief description of drawings]

FIG. 1 is a diagram schematically showing a package unit of a failure diagnosis system according to an embodiment of the present invention.

FIG. 2 is a diagram showing an overall configuration of the failure diagnosis system.

FIG. 3 is a block diagram showing a circuit configuration of a signal processing / transmission device installed in the package unit.

FIG. 4 is a block diagram showing a circuit configuration of a failure diagnosis application server of the failure diagnosis system.

FIG. 5 is a flowchart showing a failure diagnosis process executed by the failure diagnosis application server.

FIG. 6 is a diagram for explaining pattern matching processing included in the failure diagnosis processing of FIG. 5;

[Explanation of symbols]

1a, 1b, 2a to 2c ... Equipment 10 ... Package unit 11 ... Sound sensor 12 ... Odor sensor 13 ... Signal processing / transmission device 15 ... Center 16 ... Failure diagnosis application server 17 ... Pattern database 18 ... Database for abnormality diagnosis 19 ... Monitoring room 20 ... Monitoring computer

Claims (8)

[Claims] 1. A package unit including a plurality of devices mixed therein, a server for performing a failure diagnosis for each device in the package unit, and a terminal on an administrator side for receiving a diagnosis result from the server. A failure diagnosis system in which a unit and the server and a server and the terminal are respectively connected by a specific network, and a sensor for detecting a state of each device is provided in the package unit. The server collects data indicating the status of each device from the sensor collectively via the network, and a failure for each device based on the data collected by the data collecting device. The failure judgment means for judging the state and the judgment result by the failure judgment means are transmitted through the network. Fault diagnosis system comprising: a notification means for notifying to said terminal. 2. A package unit including a plurality of devices mixed therein, a server for performing a failure diagnosis for each device in the package unit, and a terminal on an administrator side for receiving a diagnosis result from the server, the package comprising: A failure diagnosis system in which a unit and the server and a server and the terminal are respectively connected by a specific network, and a sensor for detecting a state of each device is provided in the package unit. The server collects data indicating the status of each of the devices from the sensor collectively via the network, and frequency-analyzes the data collected by the data collecting means to each device. The analysis means for obtaining the corresponding pattern data and the above-mentioned devices obtained by this analysis means The pattern data that has been registered is matched with the previously registered normal pattern data, and a failure determination unit that determines the failure state for each device based on the matching result; A failure diagnosis system comprising: a notification means for notifying the terminal via the terminal. 3. A package unit including a plurality of devices mixed therein, a server for performing a failure diagnosis for each device in the package unit, and a terminal on the side of an administrator that receives a diagnosis result from the server, the package comprising: A failure diagnosis system in which a unit and the server and a server and the terminal are respectively connected by a specific network, and a first system for detecting a sound emitted from each device in the package unit is provided. A sensor and a second sensor for detecting the odor emitted by each device are provided, and the server collectively collects the sound and odor data of each device from the first and second sensors via the network. Data collection means to collect and the sound and odor data collected by this data collection means by frequency analysis The analysis means for obtaining the corresponding pattern data, and the sound and odor pattern data corresponding to each of the above-mentioned devices obtained by this analysis means are respectively subjected to matching processing with the previously registered normal sound and odor pattern data, A failure determination means for comprehensively determining the failure status of each device based on the matching result of 1. and a notification means for notifying the terminal of the determination result by the failure determination means via the network. And a fault diagnosis system. 4. The failure according to claim 1, wherein the notifying means notifies the terminal of the failure status of the device as well as information on how to deal with the failure status via the network. Diagnostic system. 5. The notification means according to claim 1, wherein the notification means notifies the terminal of information on a future maintenance plan according to the current state of the device via the network. Fault diagnosis system. 6. A fault diagnosis program for performing a fault diagnosis on a plurality of devices mixed in a package unit using a specific network and notifying an administrator of the diagnosis result. A sensor that detects the status of each of the above devices is provided inside, and a computer installed on the network collectively collects data indicating the status of each of the above devices from the sensor via the network at once. Function, a failure judgment function for judging the failure state of each of the above devices based on the data collected by this data collection function, and the judgment result by this failure judgment function to the terminal held by the administrator via the network. A failure diagnosis program that realizes the notification function of notifying. 7. A fault diagnosis program for performing a fault diagnosis on a plurality of devices mixed in a package unit using a specific network and notifying an administrator of the diagnosis result. A sensor that detects the status of each of the above devices is provided inside, and a computer installed on the network collectively collects data indicating the status of each of the above devices from the sensor via the network at once. The function, the analysis function for frequency-analyzing the data collected by this data collection function to obtain the pattern data corresponding to each device, and the pattern data corresponding to each device obtained by this analysis function are registered in advance. Matching processing with the pattern data at the normal time, and based on the matching result Fault diagnostics and fault determination function for determining a fault condition, the determination result of the failure determining function to realize the notification function for notifying the terminal via the network. 8. A failure diagnosis program for performing a failure diagnosis on a plurality of devices mixed in a package unit by using a specific network and notifying an administrator of the diagnosis result. A first sensor that detects a sound emitted by each device and a second sensor that detects an odor emitted by each device are provided in advance, and a computer installed on the network includes the first and the second sensors. The data collection function that collects the sound and odor data of each of the above devices from the two sensors collectively via the above network, and the above each device by performing the frequency analysis of the sound and odor data collected by this data collection function. And the pattern data of sound and odor corresponding to each device obtained by this analysis function. For this reason, there is a failure judgment function that performs a matching process with the registered normal sound and odor pattern data and comprehensively judges the failure status of each of the above devices based on the matching results. A failure diagnosis program that realizes a notification function of notifying the terminal of the result via the network.