JP6914141B2 - Predictive detection device, Predictive detection method, Predictive detection program and Predictive detection system - Google Patents

Description

The present invention relates to a sign detection device, a sign detection method, a sign detection program, and a sign detection system.

In recent years, aging equipment has become a social problem. For businesses that operate and manage equipment, stable operation of aging equipment and reduction of maintenance costs are important management issues.

Moving equipment is vibrating. In addition, the vibration of moving equipment changes when a failure occurs. Therefore, it is conceivable to install an acceleration sensor in the equipment and measure the vibration of the equipment to detect a sign of equipment failure.

Japanese Patent Application Laid-Open No. 2008-528112

By the way, the signal detected by the acceleration sensor includes noise. Therefore, conventionally, an analog filter is provided in the signal line of the acceleration sensor to remove noise.

However, when an analog filter is provided in the signal line of the acceleration sensor, it may not be possible to detect a sign of failure. The vibration generated in the equipment due to the failure has various frequencies from high frequency to low frequency. Therefore, when an analog filter is provided in the signal line of the accelerometer, vibration due to equipment failure may be cut or attenuated by the analog filter, and a sign of failure may not be detected.

The present invention has been made in view of the above, and an object of the present invention is to provide a sign detection device, a sign detection method, a sign detection program, and a sign detection system capable of detecting a sign of failure.

In order to solve the above-mentioned problems and achieve the object, the predictive detection device of the present invention is the measurement data of the acceleration sampled by shifting the sampling phase by three or more acceleration sensors installed in the monitoring target. It is characterized by having a storage unit for storing the data, an analysis unit for performing analysis of variance of measurement data, and a detection unit for detecting a sign of a failure to be monitored based on the analysis result by the analysis unit.

The present invention has the effect of being able to detect signs of failure.

FIG. 1 is a diagram showing an example of the configuration of a sign detection system. FIG. 2 is a diagram showing an example of a functional configuration of the sensor device. FIG. 3 is a diagram showing an example of the relationship between the sampling rate and the noise level of electrical noise. FIG. 4 is a diagram illustrating aliasing. FIG. 5 is a diagram showing an example of the relationship between the sampling rate and the noise level of aliasing. FIG. 6 is a diagram showing an example of the relationship between noise. FIG. 7 is a diagram showing an example of a functional configuration of the sign detection device. FIG. 8 is a diagram showing an example of a data structure of measurement data. FIG. 9 is a diagram showing an example of the sampled waveform. FIG. 10 is a diagram showing an example in the case where unstable aliasing does not occur. FIG. 11 is a diagram showing an example of a case where unstable aliasing occurs. FIG. 12 is a diagram showing an example of a Fourier spectrum. FIG. 13 is a diagram showing an example of Fourier spectra having different measurement periods. FIG. 14 is a flowchart showing an example of the procedure of the tuning process. FIG. 15 is a flowchart showing an example of the procedure of the sign detection process. FIG. 16 is a diagram showing a computer that executes a sign detection program.

Hereinafter, examples of the sign detection device, the sign detection method, the sign detection program, and the sign detection system according to the present invention will be described in detail with reference to the drawings. The present invention is not limited to this embodiment. Then, each embodiment can be appropriately combined as long as the processing contents do not contradict each other.

[Typical configuration of predictive detection system]
First, a typical configuration of the sign detection system according to this embodiment will be described. FIG. 1 is a diagram showing an example of the configuration of a sign detection system. The sign detection system 10 includes a sensor device 11, a gateway 12, and a sign detection device 15.

The sensor device 11 is a device that measures vibration. In the sign detection system 10, a plurality of sensor devices 11 may be provided. In the example of FIG. 1, three sensor devices 11 are shown, but the number of sensor devices 11 can be arbitrary. Each of the sensor devices 11 is installed in a monitoring target for monitoring the occurrence of a failure. For example, when monitoring the occurrence of a failure of the air conditioning equipment, the sensor device 11 is installed on a fan belt or the like of the air conditioning equipment to be monitored.

The sensor device 11 can communicate with the gateway 12 by wireless communication. Examples of the wireless communication method include Bluetooth (registered trademark), ZigBee (registered trademark), wireless LAN, and specific low-power wireless such as 315 MHz band, 400 MHz band, and 920 MHz band. The sensor device 11 detects vibration at a predetermined measurement time and transmits the detected measurement data to the gateway 12.

The gateway 12 is a device that relays communication between the sensor device 11 and the sign detection device 15. The gateway 12 is capable of communicating with the sensor device 11 by wireless communication. Further, the gateway 12 supports mobile communication such as 3G and LTE (Long Term Evolution), and is capable of communicating with the base station 13 of mobile communication. The gateway 12 is capable of communicating with the network N via the base station 13. As the network N, any kind of communication network such as a mobile communication network, the Internet (Internet), a LAN (Local Area Network), and a VPN (Virtual Private Network) can be adopted regardless of whether it is wired or wireless. The sign detection device 15 is communicably connected to the network N. The gateway 12 transmits the measurement data received from each sensor device 11 to the sign detection device 15.

The sign detection device 15 is a device that detects a sign of a failure to be monitored. The sign detection device 15 is, for example, a computer such as a personal computer or a server computer. The sign detection device 15 may be implemented as one computer, or may be implemented as a computer system consisting of a plurality of computers. In this embodiment, the case where the sign detection device 15 is used as one computer will be described as an example.

The sign detection device 15 performs analysis of variance of the measurement data received from the sensor device 11 via the gateway 12, and detects a sign of failure to be monitored based on the analysis result.

[Sensor device configuration]
Next, the configuration of the sensor device 11 will be described. FIG. 2 is a diagram showing an example of a functional configuration of the sensor device. The sensor device 11 includes a plurality of acceleration sensors 20, a memory 21, a microcomputer 22, and a wireless communication unit 23.

Here, in recent years, the development of sensing and monitoring technology has been remarkable. In particular, Micro Electro Mechanical Systems (MEMS) make it possible to offer sensors at significantly lower prices. The sensor device 11 is configured by arranging an acceleration sensor 20, various ICs (Integrated Circuits) having various functions such as a memory 21, a microcomputer 22, and a wireless communication unit 23 on a substrate. The sensor device 11 may be configured by connecting a plurality of substrates on which various ICs are arranged. For example, the sensor device 11 may be configured by connecting a substrate on which a plurality of acceleration sensors 20 and a memory 21 are arranged and a substrate on which a microcomputer 22 and a wireless communication unit 23 are arranged. Further, the sensor device 11 may be configured as a one-chip device equipped with various IC functions.

The acceleration sensor 20 is a device that measures acceleration. The sensor device 11 is provided with a plurality of acceleration sensors 20. In the example of FIG. 2, four acceleration sensors 20 (20A, 20B, 20C, 20D) are provided. The number of acceleration sensors 20 provided in the sensor device 11 may be any number as long as it is 3 or more in order to perform analysis of variance described later.

In this embodiment, the acceleration sensor 20 is a three-axis acceleration sensor capable of measuring accelerations in three-axis directions orthogonal to each other. In the following, the three axial directions orthogonal to each other will be referred to as the XYZ directions. The acceleration sensor 20 measures acceleration in the XYZ directions orthogonal to each other.

The microcomputer 22 is a device that controls the sensor device 11. The microcomputer 22 uses four acceleration sensors 20 to sample the acceleration in the XYZ directions for a predetermined period by shifting the sampling phase at each predetermined measurement time. For example, the microcomputer 22 samples the acceleration in the XYZ direction by repeating the measurement of the acceleration one by one by the four acceleration sensors 20 for a predetermined period at each measurement time. The measurement time may be a timing for each fixed period such as every hour or every day. Further, the measurement time may be a predetermined time such as 9:00, 12:00, 17:00, or 21:00. The predetermined period may be any period as long as the vibration state of the monitored object can be measured, and is, for example, about 10 seconds.

In the sensor device 11, the sampling timing at which each acceleration sensor 20 samples the acceleration is shifted by changing the acceleration sensor 20 for measuring the acceleration one by one. As a result, each acceleration sensor 20 shifts the sampling phase.

The memory 21 is a device that stores various types of data. The memory 21 stores the measurement data of the acceleration in the XYZ directions measured by each acceleration sensor 20.

The wireless communication unit 23 is a device that wirelessly transmits and receives various types of information to and from other devices. Under the control of the microcomputer 22, the wireless communication unit 23 transmits the measurement data of the acceleration in the XYZ directions measured by each acceleration sensor 20 and stored in the memory 21 to the gateway 12. This measurement data is transmitted by the gateway 12 to the sign detection device 15.

Here, the noise generated in the measurement result of the acceleration sensor 20 will be described. The acceleration sensor 20 has an output data rate (ODR) as a setting parameter for measuring acceleration. The output data rate is a parameter indicating how often the output of the acceleration of the acceleration sensor 20 is updated. When the acceleration is measured periodically, it is necessary to set the output data rate of the acceleration sensor 20 to be higher than the sampling rate for measuring the acceleration. The acceleration sensor 20 can increase the sampling rate for measuring acceleration by increasing the output data rate setting, and can detect vibrations at high frequencies. However, when the output data rate of the acceleration sensor 20 is set high, the fluctuation of electrons becomes large, and the electrical noise included in the measured acceleration value becomes large.

FIG. 3 is a diagram showing an example of the relationship between the sampling rate and the noise level of electrical noise. FIG. 3 shows the range R1 of the vibration frequency that can be measured and the level of electrical noise included in the measured acceleration value when the output data rate (ODR) of the acceleration sensor 20 is set high. There is. Further, FIG. 3 shows the range R2 of the vibration frequency that can be measured and the level of electrical noise included in the measured acceleration value when the output data rate of the acceleration sensor 20 is set low. .. As shown in FIG. 3, when the output data rate is set high, it is possible to measure vibrations at high frequencies. On the other hand, the higher the output data rate is set, the higher the level of electrical noise included in the measured acceleration value.

The acceleration sensor 20 preferably has a low output data rate setting in order to reduce the level of electrical noise. However, if the output data rate is set low, the sampling rate cannot be increased, and unstable aliasing may occur due to vibration in a high frequency band.

FIG. 4 is a diagram illustrating aliasing. For example, when the AC signal of frequency fin is A / D converted and the data is D / A converted, a sampling rate fs of at least 2 × fin or more is required to restore the original AC signal from the data. FIGS. 4A to 4D show a state in which the AC signal is A / D converted at the sampling rate fs, respectively. Each vertical line indicates the sampling timing for sampling the AC signal. The point where the vertical line and the AC signal intersect is the data measured by sampling. For example, as shown in FIG. 4A, in the case of fs >> fin, the AC signal can be restored from the data. However, as shown in FIGS. 4B and 4C, when fs≈2 × fin and fs = 2 × fin, the frequency fin of the AC signal can be reproduced, but a swell (beat) occurs in the amplitude. .. For example, in FIG. 4B, the amplitude is wavy. Further, in the cases of (B) and (C) of FIG. 4, the measured amplitude becomes small depending on the sampling phase. The fin when fs = 2 × fin is the Nyquist frequency. Further, as shown in FIG. 4D, when fs <fin, the waveform W1 having a frequency lower than the frequency fin of the AC signal is observed. This waveform W1 is called an aliasing waveform. Waveforms with frequencies higher than the Nyquist frequency are detected as aliasing waveforms with lower frequencies than they really are. This aliasing waveform is an AC signal that does not actually exist and becomes noise.

The vibration generated in the equipment due to the failure has various frequencies from high frequency to low frequency. When the microcomputer 22 samples each acceleration sensor 20 at a sampling rate fs, aliasing is generated by vibration having a frequency higher than the Nyquist frequency of the sampling rate, and the measured acceleration value includes aliasing noise. ..

FIG. 5 is a diagram showing an example of the relationship between the sampling rate and the noise level of aliasing. In FIG. 5, the output data rate (ODR) of the acceleration sensor 20 is set to a constant value, and the vibration frequency range R3 that can be restored when the sampling rate is lowered, the noise level due to aliasing, and the sampling rate are increased. The frequency range R4 of the vibration that can be restored and the noise level due to aliasing are shown. When the sampling rate is lowered, aliasing occurs even at low frequencies. Therefore, as shown in FIG. 5, when the sampling rate is lowered, the aliasing noise level included in the measured acceleration value becomes high.

Therefore, conventionally, an analog filter that cuts a signal having a frequency higher than the Nyquist frequency of the sampling rate is provided in the signal line of the accelerometer to remove aliasing noise.

On the other hand, in the sensor device 11 according to the present embodiment, the signal line of the acceleration sensor 20 is not provided with an analog filter, and the signal indicating the acceleration output by the acceleration sensor 20 is A / D converted as it is at the timing of the sampling rate. And converted to data. Therefore, in the sensor device 11 according to the present embodiment, the acceleration value measured by each acceleration sensor 20 includes aliasing noise.

There is a trade-off between electrical noise due to the output data rate setting and aliasing noise due to the sampling rate setting.

FIG. 6 is a diagram showing an example of the relationship between noise. The noise of the acceleration sensor 20 increases when the output data rate is set high. However, unless the output data rate is set high, the vibration in the high frequency band cannot be evaluated.

On the other hand, in the acceleration sensor 20, the higher the sampling rate frequency, the more the aliasing waveform is reduced, and the aliasing noise is reduced.

The sensor device 11 sets the output data rate and sampling rate of each acceleration sensor 20 according to the frequency band of interest. In this embodiment, the sensor device 11 is capable of dynamically adjusting the output data rate and sampling rate of each acceleration sensor 20 from the sign detection device 15.

Due to the characteristics of the MEMS type acceleration sensor 20, the sensor device 11 can measure only in a band where the sampling rate is very limited. In addition to this, the sensor device 11 has a characteristic that the noise level increases as the output data rate becomes higher in frequency. The higher the noise level, the more difficult it is to determine whether the sign of equipment failure is noise or a signal.

In order to deal with such problems, we focus on a stable aliasing waveform that can be measured even at a limited sampling frequency and has reproducibility. In the sign detection system 10 according to the first embodiment, for example, the output data rate and the sampling rate of each acceleration sensor 20 are adjusted from the sign detection device 15 so that a signal having a stable aliasing waveform with reproducibility can be obtained.

[Configuration of sign detection device]
Next, the configuration of the sign detection device 15 will be described. FIG. 7 is a diagram showing an example of a functional configuration of the sign detection device. The sign detection device 15 includes a communication unit 30, an operation unit 31, a display unit 32, a storage unit 33, and a control unit 34.

The communication unit 30 is an interface for transmitting and receiving various information with other devices. The communication unit 30 is connected to the network N and transmits / receives various information. For example, the communication unit 30 receives the measurement data measured by each sensor device 11 via the network N.

The operation unit 31 is an input device that receives inputs for various operations. Examples of the operation unit 31 include an input device that accepts input for operations such as a mouse and a keyboard. The operation unit 31 accepts input of various information. The operation unit 31 receives an operation input from the user, and inputs the operation information indicating the received operation content to the control unit 34.

The display unit 32 is a display device that displays various types of information. Examples of the display unit 32 include display devices such as an LCD (Liquid Crystal Display) and a CRT (Cathode Ray Tube). The display unit 32 displays various information. For example, the display unit 32 displays various screens such as an operation screen. The sign detection device 15 may receive access from an external terminal device used by an administrator or the like, display various screens such as an operation screen on the terminal device, and receive operation input from various screens.

The storage unit 33 is a storage device that stores various types of data. For example, the storage unit 33 is a storage device such as a hard disk, an SSD (Solid State Drive), or an optical disk. The storage unit 33 may be a semiconductor memory in which data such as a RAM (Random Access Memory), a flash memory, and an NVSRAM (Non Volatile Static Random Access Memory) can be rewritten.

The storage unit 33 stores the OS (Operating System) and various programs executed by the control unit 34. For example, the storage unit 33 stores various programs including a sign detection program that executes a sign detection process described later. Further, the storage unit 33 stores various data used in the program executed by the control unit 34. For example, the storage unit 33 stores the measurement data 40.

The measurement data 40 is data including the value of the acceleration measured by the sensor device 11 installed in the monitoring target. For example, in the measurement data 40, the acceleration in the three axial directions measured by each acceleration sensor 20 of the sensor device 11 is stored in association with the measurement time at each measurement time.

FIG. 8 is a diagram showing an example of a data structure of measurement data. As shown in FIG. 8, the measurement data 40 has each item of date, hour, minute, second, sampling NO, system NO, sensor NO, X-axis acceleration, Y-axis acceleration, and Z-axis acceleration. The measurement data 40 may include various information other than the above.

The date item is an area for storing the date when the acceleration was measured. The item of hour, minute, and second is an area for storing the hour, minute, and second when the acceleration is measured. The sampling NO is an area for storing a number indicating the number of times the acceleration is measured. The sensor device 11 sequentially assigns a number indicating the number of times of measurement when the acceleration is measured to each of the acceleration sensors 20. The range of usable numerical values is defined for the numbers, and the numbers are assigned in order from the lower limit of the range, and when the upper limit of the range is assigned, the lower limit is assigned again. In the example of FIG. 8, the sampling number numbers 1 to 9999 can be used, are assigned in order from 1, and return to 1 when assigned up to 9999. The item of the system NO is an area for storing the identification information that identifies each sensor device 11. For example, a unique system number is given to the sensor device 11 as identification information for identifying each sensor device 11. In the item of system NO, the system NO of the sensor device 11 that has measured the acceleration is stored. The item of the sensor NO is an area for storing the identification information for identifying the acceleration sensor 20 provided in the sensor device 11. The acceleration sensor 20 is given, for example, a unique sensor number as identification information for identifying the acceleration sensor 20 for each sensor device 11. In the sensor NO item, the sensor NO of the acceleration sensor 20 that has measured the acceleration is stored. The items of X-axis acceleration, Y-axis acceleration, and Z-axis acceleration are areas for storing measured values of acceleration in the XYZ directions.

Return to FIG. The control unit 34 is a device that controls the sign detection device 15. As the control unit 34, electronic circuits such as a CPU (Central Processing Unit) and an MPU (Micro Processing Unit), and integrated circuits such as an ASIC (Application Specific Integrated Circuit) and an FPGA (Field Programmable Gate Array) can be adopted. The control unit 34 has an internal memory for storing programs and control data that define various processing procedures, and executes various processing by these. The control unit 34 functions as various processing units by operating various programs. For example, the control unit 34 has an adjustment unit 50, a storage unit 51, an analysis unit 52, a detection unit 53, and a notification unit 54.

The adjusting unit 50 adjusts the sampling rate and the output data rate of each acceleration sensor 20 of the sensor device 11 so that a reproducible and stable aliasing waveform can be obtained. For example, the adjustment unit 50 transmits instruction information instructing the start of tuning and setting information of the sampling rate and the output data rate to be changed to each sensor device 11 at a predetermined adjustment timing. For example, the adjusting unit 50 transmits setting information including the values of the sampling rate and the output data rate to be changed to the sensor device 11.

When the microcomputer 22 of each sensor device 11 receives the instruction information for instructing the start of tuning and the setting information via the wireless communication unit 23, the value of the output data rate included in the setting information is set in each acceleration sensor 20. do. Then, the microcomputer 22 shifts the sampling phase, and samples the acceleration by each acceleration sensor 20 at the sampling rate included in the setting information for a predetermined period. The microcomputer 22 transmits the sampled acceleration measurement data to the sign detection device 15 by the wireless communication unit 23.

Upon receiving the measurement data from each sensor device 11, the adjusting unit 50 changes the sampling rate and the output data rate, and repeatedly transmits the setting information to each sensor device 11 a predetermined time.

The adjusting unit 50 performs analysis of variance of the sampled measurement data for each combination of sampling rate and output data rate for each sensor device 11 and calculates an F value. For example, the adjusting unit 50 obtains the amplitude of the acceleration in a predetermined direction from the acceleration in the three-axis direction stored in the measurement data for each combination of the sampling rate and the output data rate for each sensor device 11, and performs analysis of variance of the acceleration. conduct. The predetermined direction may be, for example, the direction of gravity, any direction orthogonal to the direction of gravity, the direction in which the amplitude of acceleration is maximum, or the direction determined in advance. good. Further, the adjusting unit 50 may perform analysis of variance of acceleration for each of the three axial directions.

Here, the sampled waveform will be described. The signal detected by each acceleration sensor 20 of the sensor device 11 includes waveforms of various frequencies including aliasing waveforms. FIG. 9 is a diagram showing an example of the sampled waveform. The waveforms shown by the solid lines (A) to (D) in FIG. 9 simulate the analog input signal. Each part indicated by a circle (“◯”) in FIGS. 9 (A) to (D) simulates a sampled digital value. The waveforms shown by the broken lines in FIGS. 9A to 9D show values that linearly interpolate the sampled digital values.

The sampling theorem requires a sampling rate of at least twice the Nyquist frequency in order to almost reproduce the analog input signal with the sampling data. For example, FIG. 9A shows a sampling state in which a substantially analog input signal is reproduced. On the other hand, when the sampling rate is sampled at a frequency lower than twice the Nyquist frequency, the amplitude undulates due to the phase shift as shown in FIGS. 9 (B) and 9 (C), and a large fluctuation occurs in the time series. In some cases.

Further, as shown in FIG. 9D, a waveform having a frequency lower than the target frequency may be observed. The waveforms shown in FIGS. 9B and 9 have large fluctuations with respect to the phase shift of sampling, and the waveforms are not stable with respect to the phase shift, resulting in unstable aliasing. On the other hand, the waveform shown in FIG. 9D has a small fluctuation with respect to the phase shift of sampling, and the waveform is stable with respect to the phase shift. When the waveform is stable with respect to the phase shift, the F value becomes a small value.

Here, analysis of variance will be described. In the analysis of variance, a variance structure model such as the following equation (1) can be set.

X _ij = μ + τ _i + ε _ij (1)
X _ij is the jth measured value of level i. μ is the average of all measurements. τ _i is a biased effect due to the level i. ε _ij is the error due to the jth measured value of level i.

In addition, the variance has the following relationship. σ ² Total represents the variance of _{X ij.} σ ² A represents the variance of _{τ i.} σ ² Error represents the variance of _{ε ij.} σ ² Total is called total variance. σ ² A is called factor variance. σ ² Error is called error variance.

When the waveform is not stable with respect to the phase shift, the waveform sampled by each acceleration sensor 20 has a large fluctuation with respect to the sampling phase shift. Therefore, the change in the Fourier spectrum of the waveform sampled by each acceleration sensor 20 is also large. Therefore, by evaluating the degree of deviation of the Fourier spectrum of each acceleration sensor 20, it is possible to determine whether the waveform is stable with respect to the phase shift. The degree of bias of the Fourier spectrum of each accelerometer 20 can be evaluated by obtaining the F value using analysis of variance.

The adjusting unit 50 performs analysis of variance of the measurement data sampled by each acceleration sensor 20 for each combination of sampling rate and output data rate for each sensor device 11 and calculates an F value. For example, the adjusting unit 50 calculates a Fourier spectrum of a predetermined target band for a waveform indicating a change in acceleration for each acceleration sensor 20 from the sampled measurement data. Then, the adjusting unit 50 calculates the F value by performing the analysis of variance of the Fourier spectrum for each of the calculated acceleration sensors 20. The predetermined target band is, for example, a range from 0 Hz to the Nyquist frequency.

For example, in the above equation (1), the difference between the acceleration sensors 20 is identified by the level i, and the variance value of the Fourier spectrum of the frequency j of the acceleration sensor 20 of i is X _ij . The adjusting unit 50 obtains the overall average μ￣ of the Fourier spectra of the four acceleration sensors 20. Further, the adjusting unit 50 obtains _{the average μ i} ￣ of the Fourier spectrum of the acceleration sensor 20 of i for each acceleration sensor 20. Then, the adjusting unit 50 sets the total sum of squares SS _{Total of the Fourier spectra of the four accelerometers 20 and the sum of squares SS A} of the Fourier spectra of the individual accelerometers 20 from the following equations (2) to (4). , Sum of squares of error SS _Error . n is the number of samples of the measured values of each acceleration sensor 20.

In the analysis of variance, the relationship of Eq. (5) below holds for _{SS Total} , SS _A, and SS _Error.

SS _Total = SS _A + SS _Error (5)

Adjustment unit 50, by dividing the sum of squares SS _A degree of freedom of the sum of squares SS _A, calculates the mean square MS _A to cause the difference of the acceleration sensor 20. The adjustment unit 50, by dividing the sum of squares SS _Error degrees of freedom of sum of squares SS _Error, calculates the mean square MS _Error to cause the difference in error. Freedom of sum of squares SS _A is 1 minus the value of the number of the acceleration sensor 20. In this embodiment, the acceleration sensor 20 is 4 subarray, the degree of freedom of the sum of squares SS _A, a 3 (= 4-1). The _{degree of freedom of SS Error} is the total value obtained by subtracting 1 from the number of acceleration data measured by each acceleration sensor 20.

The adjusting unit 50 _{calculates the F value from the mean square MS A} and the mean square MS _Error by performing the calculation of the following equation (6).

F = MS _A / MS _Error (6)

The calculated F value indicates a significant difference due to the difference in the acceleration sensor 20.

The adjusting unit 50 compares the F values calculated for each combination of the sampling rate and the output data rate for each sensor device 11, and specifies the sampling rate and the output data rate having the smallest F value for each sensor device 11. do.

The adjusting unit 50 transmits instruction information for instructing the end of tuning and setting information including the specified sampling rate and output data rate for each sensor device 11. When there are a plurality of combinations of sampling rate and output data rate whose F value is sufficiently smaller than 1, the adjusting unit 50 transmits setting information including the sampling rate and output data rate of any F value. May be good. Further, when the adjusting unit 50 has a sampling rate and an output data rate at which the F value is less than the first threshold value, the adjusting unit 50 may transmit setting information including the sampling rate and the output data rate of the F value. Further, the adjusting unit 50 may repeat the change of the sampling rate and the output data rate up to a predetermined number of times until the F value becomes less than the first threshold value. The first threshold value is, for example, 1, but is not limited to this, and may be set according to the F value that is considered to have obtained a stable aliasing waveform. Further, the first threshold value may be set from the outside. For example, the first threshold value may be set from the operation screen or the like using the operation unit 31.

When the microcomputer 22 of the sensor device 11 receives the instruction information for instructing the end of tuning and the setting information via the wireless communication unit 23, the microcomputer 22 of the sensor device 11 sets the value of the output data rate included in the setting information to each acceleration sensor 20. ..

As a result, each sensor device 11 adjusts the sampling rate and the output data rate so that the F value indicating a significant difference due to the difference in the acceleration sensor 20 becomes small. The F value indicating a significant difference due to the difference in the acceleration sensor 20 becomes small when a reproducible and stable aliasing waveform can be obtained. Therefore, each sensor device 11 adjusts the sampling rate and the output data rate so as to obtain a reproducible and stable aliasing waveform.

The adjusting unit 50 may adjust either the sampling rate or the output data rate of the acceleration sensor 20.

Each sensor device 11 shifts the sampling phase for each measurement time, samples the acceleration by each acceleration sensor 20, and transmits the measurement data to the sign detection device 15.

Here, a specific example will be described. FIG. 10 is a diagram showing an example in the case where unstable aliasing does not occur. In FIGS. 10A to 10D, waveforms 60A to 60D based on data obtained by sampling a sine wave having a natural frequency of 10 Hz with a phase change of 100 Hz for 10 seconds are shown. The sine wave was generated by a small shaking table, and the acceleration was measured by the sensor device 11 fixed to the shaking table.

Since the waveforms 60A to 60D have different phases when sampled, the phases are out of phase. In FIGS. 10A to 10D, Fourier spectra 61A to 61D showing the amplitudes of the waveforms 60A to 60D for each frequency are shown. The Fourier spectra 61A to 61D are obtained by, for example, filtering the waveforms 60A to 60D by a band path filter of 1 Hz to 25 Hz by a fast Fourier transform (FFT) to calculate the Fourier spectrum. In the absence of unstable aliasing, the Fourier spectra 61A-61D are similar even if the sampling is out of phase, with a significant peak in amplitude at 10 Hz, as shown in FIG. In the Fourier spectra 61A to 61D shown in FIG. 10, the F value is calculated to be 0.2.

FIG. 11 is a diagram showing an example of a case where unstable aliasing occurs. In FIGS. 11A to 11D, waveforms 62A to 62D based on data obtained by sampling a sine wave having a natural frequency of 75 Hz with a phase change of 100 Hz for 10 seconds are shown. The sine wave was generated by a small shaking table, and the acceleration was measured by the sensor device 11 fixed to the shaking table.

The waveforms 62A to 62D also have different phases when sampled, so that the phases are out of phase. In FIGS. 11A to 11D, Fourier spectra 63A to 63D showing the amplitudes of the waveforms 62A to 62D for each frequency are shown. The Fourier spectra 63A to 63D are obtained by, for example, filtering the waveforms 62A to 62D by a bandpass filter of 1 Hz to 25 Hz by a fast Fourier transform to calculate the Fourier spectrum. When unstable aliasing occurs, as shown in FIG. 11, in the Fourier spectra 63A to 63D, the waveform changes due to the out-of-phase sampling, and noise is generated at various frequencies. In the Fourier spectra 63A to 63D shown in FIG. 11, the F value is calculated to be 5.02.

By adjusting the sampling rate and output data rate of the acceleration sensor 20 of each sensor device 11 so that the F value becomes small, the sign detection device 15 can measure the measurement data of the reproducible and stable aliasing waveform from each sensor device 11. Will be obtained.

Returning to the description of FIG. The storage unit 51 stores various types of data. For example, the storage unit 51 stores the measurement data of each sensor device 11 received from the gateway 12 in the storage unit 33 as the measurement data 40.

The analysis unit 52 performs an analysis of variance of the measurement data 40. For example, the analysis unit 52 obtains the amplitude of the acceleration in a predetermined direction from the acceleration in the three-axis direction stored in the measurement data 40, and performs an analysis of variance of the acceleration. The predetermined direction may be, for example, the direction of gravity, any direction orthogonal to the direction of gravity, the direction in which the amplitude of acceleration is maximum, or the direction determined in advance. good. Further, the analysis unit 52 may perform analysis of variance of acceleration for each of the three axial directions.

Here, the vibration of the equipment changes when a failure is occurring. Therefore, it is possible to determine whether or not the vibration is changing by performing an analysis of variance of a plurality of measurement timing data stored in the measurement data 40.

Based on the measurement data 40, the analysis unit 52 calculates a Fourier spectrum of a predetermined target band for a waveform indicating a change in acceleration measured by each acceleration sensor 20 for each measurement time. Then, the analysis unit 52 calculates the variance value for each frequency of the Fourier spectrum from each of the calculated Fourier spectra for each measurement time.

FIG. 12 is a diagram showing an example of a Fourier spectrum. In FIGS. 12A to 12D, waveforms 64A to 64D based on data obtained by sampling a sine wave having a natural frequency of 10 Hz with a phase change of 100 Hz for 10 seconds are shown. The sine wave was generated by a small shaking table, and the acceleration was measured by the sensor device 11 fixed to the shaking table.

Since the waveforms 64A to 64D have different phases when sampled, the phases are out of phase. In FIGS. 12A to 12D, Fourier spectra 65A to 65D showing the amplitudes of the waveforms 64A to 64D for each frequency are shown. The analysis unit 52 calculates the variance of the values of the Fourier spectra 65A to 65D for each frequency. For example, the analysis unit 52 calculates the variance value of the values of the Fourier spectra 65A to 65D for each same frequency. FIG. 12 (E) shows a waveform 67 showing the dispersion values of the Fourier spectra 65A to 65D for each frequency. Hereinafter, a waveform showing a dispersion value for each frequency of the Fourier spectrum, such as waveform 67, is also referred to as a “Fourier spectrum dispersion waveform”.

The analysis unit 52 calculates the F value by performing the analysis of variance of the variance value for each frequency of the Fourier spectrum at each measurement time. For example, the analysis unit 52 calculates the variance value for each frequency of the Fourier spectrum for each measurement time for three or more measurement times from the measurement data 40. As a result, the Fourier spectrum dispersion waveform 67 shown in FIG. 12 (E) is calculated for each measurement time. Then, the analysis unit 52 calculates the F value by performing the analysis of variance of the variance value for each frequency of the calculated Fourier spectra of the three or more measurement periods.

For example, in the above equation (1), the difference in measurement time is identified by the level i, and the variance value of the Fourier spectrum of the frequency j at the measurement time of i is X _ij . The analysis unit 52 obtains the overall average μ￣ of the Fourier spectrum dispersion waveforms of all the measurement periods targeted for the analysis of variance. For example, the analysis unit 52 obtains the overall average μ￣ of the variance values for each frequency of the Fourier spectrum at all measurement periods. _{Further, the analysis unit 52 obtains the average μ i} ￣ of the Fourier spectrum dispersion waveform of the measurement time of i for each measurement time to be analyzed by the variance analysis. _{For example, the analysis unit 52 obtains the average μ i} ￣ of the variance values for each frequency of the Fourier spectrum at the measurement time of i. Then, the analysis unit 52 uses the above equations (2) to (4) to determine the total sum of squares SS _Total of the Fourier spectrum variance waveforms of all the measurement periods targeted for analysis of variance and the individual measurement periods. _{Find the sum of squares SS A} of the Fourier spectrum ANOVA waveform and the sum of squares SS _{Error of} the error. n is the number of samples of the measured value at each measurement period.

Analysis unit 52, by dividing the sum of squares SS _A degree of freedom of the sum of squares SS _A, calculates the mean square MS _A to cause a difference in the measurement period. The analysis unit 52, by dividing the sum of squares SS _Error degrees of freedom of sum of squares SS _Error, calculates the mean square MS _Error to cause the difference in error. Freedom of sum of squares SS _A is 1 minus the number of measurement time to the subject of analysis of variance. The _{degree of freedom of SS Error} is the total value obtained by subtracting 1 from the number of acceleration data measured at each measurement period.

The analysis unit 52 _{calculates the F value from the mean square MS A} and the mean square MS _Error using the above equation (6).

The calculated F value indicates a significant difference due to the difference in measurement time.

Here, the vibration of moving equipment is not constant even when it is operating normally, and may change depending on the operating state of the equipment. For example, when the monitoring target is an air conditioner, the operating state of the air conditioner changes depending on the time of day due to the influence of changes in temperature and the like. It is preferable that the three or more measurement periods used for calculating the F value are measurement periods in which the monitored object is in the same operating state. For example, if the monitoring target is a fan belt of an air conditioner, and the air conditioner changes its operating state over time, but the operating state changes at the same time on each day, the analysis unit 52 performs the same operation on each day. It is preferable to calculate the F value from the data whose measurement time is the time.

Further, the measurement time used for calculating the F value preferably includes the measurement time of the operating state in which the monitored object is operating normally. For example, when the monitoring target is a fan belt of the air conditioning equipment and the air conditioning equipment is operating normally at the measurement time A, the analysis unit 52 obtains an F value from the measurement data 40 of a plurality of measurement times including the measurement time A. It is preferable to calculate.

Here, a specific example will be described. FIG. 13 is a diagram showing an example of Fourier spectra having different measurement periods. In FIGS. 13A to 13D, Fourier spectrum dispersion waveforms 68A to 68D showing the dispersion values for each frequency of the measurement periods 1 to 4 are shown. In FIGS. 13 (A) to 13 (D), a Fourier spectrum is calculated from measurement data obtained by sampling a sine wave having a natural frequency frequency of 10 Hz at 200 Hz with a phase change of 200 Hz for 10 seconds, and each calculated Fourier spectrum is used for each measurement time. The dispersion value for each frequency of the Fourier spectrum is calculated. The sine wave was generated by a small shaking table, and the acceleration was measured by the sensor device 11 fixed to the shaking table. However, at the measurement time 4, in order to simulate a sign of failure of the equipment, the bolts fixing the shaking table and the sensor device 11 were loosened, and the measurement was performed with some backlash.

The analysis unit 52 performs analysis of variance of the Fourier spectrum for each measurement period to calculate the F value. For example, when the ANOVA of the Fourier spectrum dispersion waveforms 68A to 68C shown in FIG. 13 is performed, the F value is calculated to be 0.15. On the other hand, when the ANOVA of the Fourier spectrum dispersion waveforms 68A to 68D shown in FIG. 13 is performed, the F value is calculated to be 1.29. In this way, when the ANOVA is performed by adding the Fourier spectrum dispersion waveform 68D simulating the failure of the equipment, the F value is calculated to be large.

As described above, when no abnormality has occurred in the equipment and the change in vibration at each measurement period is small, the F value calculated by the analysis unit 52 is a small value. On the other hand, when an abnormality is occurring in the equipment and the vibration at each measurement time is changing, the F value calculated by the analysis unit 52 becomes a large value. Therefore, by evaluating the F value, it can be determined whether the vibration is changing, and it can be determined that a failure is occurring.

Returning to the description of FIG. The detection unit 53 detects a sign of a failure to be monitored based on the analysis result by the analysis unit 52. For example, in the detection unit 53, the F value indicating a significant difference due to the difference in the measurement time, which is calculated by the analysis unit 52 by performing the analysis of variance of the Fourier spectrum for each measurement time, is equal to or larger than a predetermined second threshold value. Is determined. Then, when the F value is equal to or higher than the second threshold value, the detection unit 53 detects that there is a sign of failure. The second threshold value is, for example, 1, but is not limited to this, and may be set according to the F value regarded as a sign of failure. Further, the second threshold value may also be set from the outside.

When the detection unit 53 detects a sign of a failure of the monitoring target, the notification unit 54 notifies the monitoring target of the detection of the failure. For example, the notification unit 54 displays a message or the like notifying the display unit 32 of the detection of a failure with respect to the monitoring target. In addition, the notification unit 54 transmits information for notifying the detection of a failure to the monitoring target to an external terminal device used by the administrator or the like.

As a result, the sign detection device 15 can detect the sign of failure.

[Processing flow]
Various processes performed by the sign detection device 15 will be described. First, the flow of the tuning process in which the sign detection device 15 tunes the sampling rate and the output data rate of the acceleration sensor 20 of the sensor device 11 will be described.

FIG. 14 is a flowchart showing an example of the procedure of the tuning process. This tuning process is executed at a predetermined adjustment timing, for example, the timing at which the sensor device 11 is installed, the periodic timing such as the date and time or monthly, or the timing instructed by the administrator.

The adjusting unit 50 transmits the instruction information instructing the start of tuning and the setting information of the sampling rate and the output data rate to be changed to the sensor device 11 at a predetermined adjustment timing (step S10).

When the sensor device 11 receives the instruction information for instructing the start of tuning and the setting information, the sensor device 11 sets the value of the output data rate included in the setting information in each acceleration sensor 20 and performs acceleration sampling for a predetermined period. Then, the sensor device 11 transmits the sampled acceleration measurement data to the sign detection device 15.

The adjusting unit 50 determines whether or not the measurement data has been received (step S11). If the measurement data has not been received (step S11: No), the process proceeds to step S11 again.

When the measurement data is received (step S11: Yes), the adjusting unit 50 calculates the Fourier spectrum of a predetermined target band for the waveform indicating the change in acceleration for each acceleration sensor 20 from the received measurement data (step S11: Yes). S12). The adjusting unit 50 performs analysis of variance of the Fourier spectrum for each acceleration sensor 20 and calculates an F value indicating a significant difference due to the difference in the acceleration sensor 20 (step S13).

The adjusting unit 50 determines whether the calculated F value is less than the first threshold value (step S14). When the F value is not less than the first threshold value (step S14: No), the adjusting unit 50 determines whether or not the sampling rate and the output data rate have been changed a predetermined number of times or more (step S15). When the number of changes is not more than a predetermined number (step S15: No), the adjusting unit 50 changes the sampling rate and the output data rate (step S16). For example, the adjusting unit 50 changes to a sampling rate higher than the currently set sampling rate. Further, the adjusting unit 50 changes to the lowest output data rate among the output data rates updated at a frequency equal to or higher than the sampling rate to be changed.

The adjusting unit 50 transmits the changed sampling rate and output data rate setting information to the sensor device 11 (step S17), and proceeds to step S11 described above.

On the other hand, when the number of changes becomes a predetermined number or more (step S15: Yes), the adjusting unit 50 specifies the sampling rate and the output data rate at which the calculated F value is the smallest (step S18). The adjustment unit 50 transmits instruction information for instructing the end of tuning and setting information including the specified sampling rate and output data rate (step S19), and ends the process.

On the other hand, when the F value becomes less than the first threshold value (step S14: Yes), the adjusting unit 50 specifies the sampling rate and the output data rate from which the F value is calculated (step S20). The adjustment unit 50 transmits instruction information for instructing the end of tuning and setting information including the specified sampling rate and output data rate (step S21), and ends the process.

Next, the flow of the sign detection process in which the sign detection device 15 detects the sign of failure will be described. FIG. 15 is a flowchart showing an example of the procedure of the sign detection process. This sign detection process is executed at a predetermined timing, for example, when measurement data is received from each sensor device 11.

The storage unit 51 stores the received measurement data of the sensor device 11 as measurement data 40 in the storage unit 33 (step S50).

The analysis unit 52 determines whether or not three or more measurement data of the measurement time are obtained from the sensor device 11 for each sensor device 11 (step S51). When the measurement data of three or more measurement times is not obtained (step S51: No), the process ends.

On the other hand, when measurement data of three or more measurement periods are obtained (step S51: Yes), the analysis unit 52 determines a waveform indicating the change in acceleration measured by each acceleration sensor 20 for each measurement period. The Fourier spectrum of the target band of is calculated (step S52). Then, the analysis unit 52 calculates the variance value for each frequency of the Fourier spectrum from each of the calculated Fourier spectra for each measurement time (step S53). The analysis unit 52 performs analysis of variance of the variance value for each frequency of the Fourier spectrum of each measurement time, and calculates an F value indicating a significant difference due to the difference in the measurement time (step S54).

The detection unit 53 determines whether the F value indicating the significant difference due to the calculated difference in measurement time is equal to or greater than the second threshold value (step S55). When the F value is less than the second threshold value (step S55: No), the process ends.

On the other hand, when the F value is equal to or higher than the second threshold value (step S55: Yes), the detection unit 53 detects that there is a sign of failure to be monitored (step S56). The notification unit 54 notifies the monitoring target of the detection of a failure (step S57), and ends the process.

[effect]
As described above, the sign detection device 15 according to the present embodiment stores the measurement data 40 of the acceleration sampled by shifting the sampling phase by three or more acceleration sensors 20 installed in the monitoring target. Store in 33. The sign detection device 15 performs an analysis of variance of the measurement data 40. The sign detection device 15 detects a sign of failure to be monitored based on the analysis result. As a result, the sign detection device 15 can detect the sign of failure.

Further, the sign detection device 15 according to the present embodiment calculates a Fourier spectrum of a predetermined target band for a waveform indicating a change in acceleration sampled by three or more acceleration sensors 20 respectively. The sign detection device 15 performs analysis of variance of the calculated Fourier spectrum for each acceleration sensor 20 and calculates an F value indicating a significant difference due to the difference in the acceleration sensor 20. The sign detection device 15 adjusts one or both of the sampling rate and the output data rate of the acceleration sensor 20 so that the F value becomes the smallest or the F value becomes less than the first threshold value. do. As a result, the sign detection device 15 can adjust the output data rate and the sampling rate of the acceleration sensor 20 of the sensor device 11 so that a signal having a stable aliasing waveform with reproducibility can be obtained. As a result, the sign detection device 15 can obtain reproducible and stable measurement data of the aliasing waveform from the sensor device 11, so that the accuracy of detecting the sign of failure can be improved.

Further, the sign detection device 15 according to the present embodiment stores the measurement data 40 of the acceleration sampled for a predetermined period at a predetermined measurement time by shifting the sampling phase by three or more acceleration sensors 20. From the measurement data 40, the sign detection device 15 calculates a Fourier spectrum of a predetermined target band for a waveform indicating a change in acceleration measured by each acceleration sensor 20 for each measurement time. The sign detection device 15 calculates the variance value for each frequency of the Fourier spectrum for each measurement time from each calculated Fourier spectrum. The sign detection device 15 performs analysis of variance of the variance value for each frequency of the calculated Fourier spectrum of each measurement time, and calculates an F value indicating a significant difference due to the difference in the measurement time. The sign detection device 15 detects that there is a sign of failure when the calculated F value is equal to or higher than a predetermined second threshold value. As a result, the sign detection device 15 can detect the sign of failure.

Further, the sign detection device 15 according to the present embodiment calculates the variance value for each frequency of each Fourier spectrum for each measurement time for three or more measurement times from the measurement data 40. The sign detection device 15 performs analysis of variance of the variance value for each frequency of the Fourier spectrum of the calculated three or more measurement timings, and calculates the F value indicating a significant difference due to the difference in the measurement timings. As a result, the sign detection device 15 can calculate an F value indicating a significant difference due to a difference in measurement time by analysis of variance.

Further, the sign detection device 15 according to the present embodiment calculates the variance value for each frequency of each Fourier spectrum for each measurement time from the measurement data 40 of a plurality of measurement times in which the monitoring target is in the same operating state. The sign detection device 15 performs analysis of variance of the variance values of the calculated Fourier spectra of the Fourier spectra at a plurality of measurement times for each frequency, and calculates an F value indicating a significant difference due to the difference in the measurement times. As a result, the sign detection device 15 can accurately detect the sign of failure.

Further, the sign detection device 15 according to the present embodiment shows a significant difference due to the difference in the measurement time from the measurement data of a plurality of measurement times including the measurement time of the operating state in which the monitored object is operating normally. Calculate the value. As a result, the sign detection device 15 can accurately detect the change in vibration from the operating state in which it is operating normally, and can accurately detect the sign of failure.

Further, in the sensor device 11 according to the present embodiment, three or more acceleration sensors 20 are provided on the same substrate, and the acceleration measurement is repeated one by one by the three or more acceleration sensors 20 during the sampling period. As a result, the sensor device 11 can measure the acceleration by each acceleration sensor 20 by shifting the sampling phase.

Although the embodiments relating to the disclosed device have been described so far, the disclosed technique may be implemented in various different forms other than the above-described embodiments. Therefore, other examples included in the present invention will be described below.

For example, in the above embodiment, the case where the sensor device 11 communicates with the sign detection device 15 via the gateway 12 and the network N has been described, but the disclosed device is not limited thereto. For example, the sensor device 11 and the sign detection device 15 may directly perform wireless communication. Further, the sensor device 11 constitutes a multi-hop wireless network corresponding to multi-hop wireless communication such as, for example, a 920 MHz band multi-hop radio, and each sensor device 11 directly or via another sensor device 11 is a sign detection device. You may communicate with 15.

Further, in the above embodiment, the case where the sensor device 11 transmits the measurement data to the sign detection device 15 and the sign detection device 15 performs the analysis of variance of the measurement data to detect the sign of failure has been described. The device is not limited to this. For example, the sensor device 11 may be provided with the functions of the analysis unit 52, the detection unit 53, and the notification unit 54, and the sensor device 11 may perform analysis of variance of the measurement data and detect a sign of failure. In this case, the sensor device 11 corresponds to the sign detection device of the present invention. Further, the sensor device 11 may be provided with the function of the adjusting unit 50, and the sensor device 11 may independently tune the output data rate and the sampling rate of each acceleration sensor 20.

Further, each component of each of the illustrated devices is a functional concept, and does not necessarily have to be physically configured as shown in the figure. That is, the specific state of distribution / integration of each device is not limited to the one shown in the figure, and all or part of the device is functionally or physically distributed / physically in arbitrary units according to various loads and usage conditions. Can be integrated and configured. For example, each processing unit of the adjustment unit 50, the storage unit 51, the analysis unit 52, the detection unit 53, and the notification unit 54 may be integrated as appropriate. Further, the processing of each processing unit may be appropriately separated into the processing of a plurality of processing units. Further, each processing function performed in each processing unit may be realized by a CPU and a program analyzed and executed by the CPU, or may be realized as hardware by wired logic. ..

[Sign detection program]
Further, the various processes described in the above-described embodiment can also be realized by executing a program prepared in advance on a computer system such as a personal computer or a workstation. Therefore, an example of a computer system that executes a program having the same function as that of the above embodiment will be described below. FIG. 16 is a diagram showing a computer that executes a sign detection program.

As shown in FIG. 16, the computer 300 includes a CPU (Central Processing Unit) 310, an HDD (Hard Disk Drive) 320, and a RAM (Random Access Memory) 340. Each of these 300 to 340 parts is connected via the bus 400.

The HDD 320 stores in advance a sign detection program 320a that has the same functions as the storage unit 51, the analysis unit 52, the detection unit 53, the notification unit 54, and the adjustment unit 50. The sign detection program 320a may be separated as appropriate.

Further, the HDD 320 stores various information. For example, the HDD 320 stores various data used for detecting a sign of failure, such as the above-mentioned measurement data 40.

Then, the CPU 310 reads the sign detection program 320a from the HDD 320 and executes it to execute the same operation as each processing unit of the embodiment. That is, the sign detection program 320a executes the same operations as the storage unit 51, the analysis unit 52, the detection unit 53, the notification unit 54, and the adjustment unit 50.

The above-mentioned sign detection program 320a does not necessarily have to be stored in the HDD 320 from the beginning.

For example, the program is stored in a "portable physical medium" such as a flexible disk (FD), a CD-ROM, a DVD disk, a magneto-optical disk, or an IC card inserted into the computer 300. Then, the computer 300 may read the program from these and execute the program.

Further, the program is stored in an "other computer (or server)" connected to the computer 300 via a public line, the Internet, a LAN, a WAN, or the like. Then, the computer 300 may read the program from these and execute the program.

10 Predictive detection system 11 Sensor device 15 Predictive detection device 20 Accelerometer 21 Memory 22 Microcomputer 23 Wireless communication unit 30 Communication unit 31 Operation unit 32 Display unit 33 Storage unit 34 Control unit 40 Measurement data 50 Adjustment unit 51 Storage unit 52 Analysis unit 53 Detection unit 54 Notification unit

Claims (11)

A storage unit that stores measurement data of acceleration sampled by shifting the sampling phase by three or more acceleration sensors installed in the monitoring target.
An analysis unit that performs analysis of variance of the measurement data,
Based on the analysis result by the analysis unit, the detection unit that detects a sign of failure of the monitoring target and the detection unit
A sign detection device characterized by having. A Fourier spectrum of a predetermined target band for a waveform showing a change in acceleration sampled by each of the three or more acceleration sensors is calculated, and a dispersion analysis of the calculated Fourier spectrum for each acceleration sensor is performed to obtain the acceleration sensor. The sampling rate and output data of the accelerometer are calculated so that the F value indicating a significant difference due to the difference is the smallest, or the F value is less than the first threshold value. The predictive detection device according to claim 1, further comprising an adjusting unit for adjusting the setting of either one or both of the rates. The storage unit stores the measurement data of the acceleration sampled for a predetermined period at a predetermined measurement time by shifting the sampling phase by the three or more acceleration sensors.
From the measurement data, the analysis unit calculates a Fourier spectrum of a predetermined target band for a waveform indicating a change in acceleration measured by each acceleration sensor for each measurement time, and from each calculated Fourier spectrum for each measurement time. The dispersion value for each frequency of the Fourier spectrum is calculated, and the variance analysis of the dispersion value for each frequency of the calculated Fourier spectrum at each measurement time is performed to calculate the F value indicating a significant difference due to the difference in the measurement time. death,
The sign detection device according to claim 1 or 2, wherein the detection unit detects that there is a sign of failure when the F value calculated by the analysis unit is equal to or higher than a predetermined second threshold value. The analysis unit calculates the variance value for each frequency of the Fourier spectrum for each of the measurement timings for three or more measurement timings from the measurement data, and calculates the variance value for each frequency of the Fourier spectrum of the three or more measurement timings. The predictive detection device according to claim 3, wherein an analysis of variance of the variance value is performed to calculate an F value indicating a significant difference due to the difference in measurement time. The analysis unit calculates the variance value for each frequency of the Fourier spectrum for each measurement period from the measurement data of a plurality of measurement periods in which the monitoring target is in the same operating state, and calculates the Fourier spectrum of the plurality of measurement periods. The predictive detection device according to claim 3 or 4, wherein the F value indicating a significant difference due to the difference in measurement time is calculated by performing analysis of variance of the dispersion value for each frequency. The analysis unit calculates an F value indicating a significant difference due to the difference in the measurement timing from the measurement data of a plurality of measurement timings including the measurement timing of the operating state in which the monitoring target is operating normally. The predictive detection device according to any one of claims 3 to 5, which is characterized. Three or more accelerometers installed in the monitoring target perform analysis of variance of the measurement data described in the storage unit that stores the measurement data of the acceleration sampled by shifting the sampling phase.
Detecting signs of failure of the monitored object based on the analysis results,
A predictive detection method characterized by a computer performing processing. Three or more accelerometers installed in the monitoring target perform analysis of variance of the measurement data described in the storage unit that stores the measurement data of the acceleration sampled by shifting the sampling phase.
Detecting signs of failure of the monitored object based on the analysis results,
A predictive detection program characterized by having a computer perform processing. A sensor device installed in the monitoring target and transmitting measurement data of the sampled acceleration by shifting the sampling phase by three or more acceleration sensors.
A storage unit that stores the measurement data received from the sensor device, an analysis unit that performs analysis of variance of the measurement data, and a detection unit that detects a sign of failure of the monitoring target based on the analysis result by the analysis unit. With a sign detector with,
A sign detection system characterized by being equipped with. The ninth aspect of the sensor device is characterized in that three or more acceleration sensors are provided on the same substrate, and acceleration measurement is repeated one by one by the three or more acceleration sensors during a sampling period. Described predictive detection system. The sign detection device is
A Fourier spectrum of a predetermined target band is calculated for a waveform indicating a change in acceleration sampled by each of the three or more accelerometers, and a dispersion analysis of the calculated Fourier spectrum for each accelerometer is performed by each accelerometer. The sampling rate and output data of the accelerometer are calculated so that the F value indicating a significant difference between the sampled waveforms is calculated so that the F value is the smallest or the F value is less than the first threshold value. It also has an adjustment unit that adjusts one or both of the rate settings.
The sensor device is characterized in that the three or more acceleration sensors sample acceleration by shifting the sampling phase at one or both of the sampling rate and the output data rate adjusted by the adjusting unit. The sign detection system according to claim 9 or 10.

Images