Signal processing method and signal processing device
The signal processing method uses correlation filter processing to enhance periodic components and reduce noise in asynchronous waveforms, achieving a high S/N ratio correlation signal for accurate analysis without altering the object's state.
Patent Information
- Application Number
- JP2020124345
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2020-07-21
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2040-07-21
AI Technical Summary
Existing methods for reducing asynchronous waveform components using synchronous addition also reduce correlated waveform components that are not synchronized, requiring complex calculations and envelope processing.
A signal processing method involving correlation filter processing between a template signal and a processing target signal, generated based on a source signal from a steady-state object, to enhance periodic components and reduce ergodic noise without envelope processing.
This method effectively reinforces periodic signal components and reduces noise, resulting in a high Signal-to-Noise Ratio (S/N) correlation signal, allowing accurate analysis without altering the object's state and reducing memory requirements.
Smart Images

Figure 0007585640000001 
Figure 0007585640000002 
Figure 0007585640000003
Abstract
Description
[Technical field]
[0001] The present invention relates to a signal processing method and a signal processing device. [Background technology]
[0002] Synchronous addition is known as a method for reducing components that are asynchronous with the target stationary repetitive waveform. However, this method has the problem that it also reduces waveform components that are correlated with the repetitive waveform but are not synchronized with the synchronous addition timing. As a method for addressing this problem, Non-Patent Document 1 proposes a method in which envelope processing is performed on the target stationary repetitive time series waveform, and the obtained waveform is subjected to spectrum analysis. [Prior art documents] [Non-patent literature]
[0003] [Non-Patent Document 1] Pete Sopcik and Dara O'Sullivan, "How Sensor Performance Enables Condition-Based Monitoring Solutions", Analog Dialogue 53-06, June 2019. Summary of the Invention [Problem to be solved by the invention]
[0004] However, in envelope processing, it is necessary to perform smoothing processing after rectifying the time series waveform, and it is necessary to appropriately select the cutoff frequency of the smoothing filter so that the desired signal component is appropriately extracted, so that the method described in Non-Patent Document 1 involves complex calculations. [Means for solving the problem]
[0005] One aspect of the signal processing method according to the present invention includes: a process-target signal generating step of generating a process target signal which is a time-series signal based on a source signal which is a time-series signal corresponding to a physical quantity generated in a steady state of the object; and a correlation signal generating step of performing correlation filtering between the processing target signal and a template signal, which is a time-series signal generated based on the source signal, to generate a correlation signal, which is a time-series signal.
[0006] One aspect of the signal processing device according to the present invention is a signal-to-be-processed circuit that generates a signal-to-be-processed, which is a time-series signal, based on a source signal that is a time-series signal corresponding to a physical quantity generated in a steady state of an object; a memory circuit for storing a template signal, which is a time-series signal generated based on the source signal; The correlation signal generating circuit performs correlation filtering between the template signal and the processing target signal to generate a correlation signal that is a time-series signal. [Brief description of the drawings]
[0007] [Figure 1] FIG. 4 is a flowchart showing the procedure of a signal processing method according to the first embodiment. [Diagram 2] FIG. 4 is a flowchart showing an example of a procedure of a template signal generating step in the first embodiment. [Diagram 3] FIG. 4 is a flowchart showing an example of a procedure of a processing target signal generating step in the first embodiment. [Figure 4] FIG. 4 is a flowchart showing an example of a procedure of a correlation signal generating step in the first embodiment. [Diagram 5] FIG. 2 is a diagram showing a specific example of an object and a source signal. [Figure 6] FIG. 4 is a diagram showing an example of a signal in a first period extracted from an original signal. [Figure 7] FIG. 4 is a diagram showing an example of a template signal. [Figure 8] FIG. 4 is a diagram showing an example of a waveform of a signal to be processed. [Figure 9] FIG. 4 is a diagram showing an example of a waveform of a correlation signal. [Figure 10]9 is an enlarged view of the waveform of the signal to be processed shown in FIG. 8. [Figure 11] FIG. 10 is an enlarged view of the waveform of the correlation signal shown in FIG. [Figure 12] FIG. 1 is a diagram showing an example of the configuration of a signal processing device according to a first embodiment. [Figure 13] FIG. 11 is a flowchart showing an example of a procedure of a template signal generating step in the second embodiment. [Figure 14] FIG. 11 is a flowchart showing an example of a procedure of a process target signal generating step in the second embodiment. [Figure 15] FIG. 13 is a flowchart showing an example of a procedure of a template signal generating step in the third embodiment. [Figure 16] FIG. 13 is a flowchart showing an example of a procedure of a correlation signal generating step in the third embodiment. [Figure 17] FIG. 13 is a flowchart showing the procedure of a signal processing method according to a fourth embodiment. [Figure 18] FIG. 13 is a flowchart showing the procedure of a signal processing method according to a fifth embodiment. [Figure 19] FIG. 13 is a flowchart showing another procedure of the signal processing method according to the fifth embodiment. [Figure 20] FIG. 13 is a flowchart showing an example of a procedure of a waveform analysis step in the fifth embodiment. [Figure 21] FIG. 4 is a diagram showing an example of a frequency spectrum obtained by performing an FFT on a correlation signal. [Figure 22] FIG. 13 is a diagram showing an example of a frequency spectrum of a comparative example. [Diagram 23] FIG. 13 is a diagram showing another example of a frequency spectrum obtained by performing FFT on a correlation signal. [Figure 24] FIG. 13 is a diagram showing another example of a frequency spectrum of the comparative example. [Diagram 25] FIG. 13 is a diagram showing an example of the configuration of a signal processing device according to a fifth embodiment. [Figure 26] FIG. 23 is a flowchart showing an example of a procedure of a waveform analysis step in the sixth embodiment. [Figure 27] FIG. 12 is an enlarged view of the waveform of the correlation signal shown in FIG. [Figure 28]FIG. 23 is a flowchart showing the procedure of a signal processing method according to a seventh embodiment. [Figure 29] FIG. 23 is a flowchart showing another procedure of the signal processing method according to the seventh embodiment. [Diagram 30] FIG. 13 is a diagram showing an example of the configuration of a signal processing device according to a seventh embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0008] Preferred embodiments of the present invention will be described in detail below with reference to the drawings. Note that the embodiments described below do not unduly limit the contents of the present invention described in the claims. Furthermore, not all of the configurations described below are necessarily essential components of the present invention.
[0009] 1. First embodiment 1-1.Signal processing method FIG. 1 is a flow chart showing the procedure of the signal processing method according to the first embodiment.
[0010] 1, the signal processing method of the first embodiment includes a template signal generating step S1, a processing target signal generating step S2, and a correlation signal generating step S3. The signal processing method of the present embodiment is performed by, for example, a signal processing device 1.
[0011] First, in a template signal generating step S1, the signal processing device 1 generates a template signal, which is a time-series signal, based on a source signal, which is a time-series signal corresponding to a physical quantity generated in a steady state of an object. The object is an object to be subjected to signal processing, and the type of the object may be, in particular, The physical quantity is not limited, and may be, for example, various devices such as a motor having a rotating mechanism or a vibrating mechanism, a structure such as a bridge or a building that vibrates due to an external force, or an electric circuit that generates a periodic signal. The steady state of an object is a state in which the object repeats a predetermined operation, and may be, for example, a state in which the object, that is, an object, a device or a structure, is rotating or vibrating, or a state in which the object, that is, an electric circuit, is generating a periodic signal. The type of physical quantity is not particularly limited, and may be, for example, acceleration, angular velocity, pressure, current, voltage, etc.
[0012] Next, in a processing target signal generating step S2, the signal processing device 1 generates a processing target signal, which is a time-series signal, based on the source signal.
[0013] Next, in a correlation signal generating step S3, the signal processing device 1 performs correlation filtering between the template signal generated in step S1 and the processing target signal generated in step S2 to generate a correlation signal that is a time-series signal.
[0014] Then, the signal processing device 1 repeats steps S1, S2, and S3 until the signal processing is completed (N in step S4).
[0015] In this manner, in the present embodiment, the template signal and the processing target signal are updated until the signal processing is completed. By updating the template signal, even if the period of a signal component having periodicity included in the processing target signal changes due to a change over time of the object, a correlation signal including the signal component can be generated.
[0016] FIG. 2 is a flow chart showing an example of the procedure of the template signal generating step S1 in FIG.
[0017] As shown in FIG. 2, in the template signal generating step S1, the signal processing device 1 first extracts a signal of a first period having a length t1 from the source signal (step S11).
[0018] Next, the signal processing device 1 generates a template signal by multiplying the signal in the first period by a window function (step S12), and ends the template signal generating step S1.
[0019] Thus, in this embodiment, the template signal is a signal based on the signal of the first period cut out from the source signal, and its time length is equal to the length t1 of the first period. Also, the template signal is a signal obtained by multiplying the signal cut out from the source signal by a window function. This makes it possible to reduce the influence of noise caused by discontinuity between the first sample value and the last sample value of the template signal on the correlation signal. The type of window function is not particularly limited, and examples of the window function include a Hanning window function, a rectangular window function, a Gaussian window function, a Hamming window function, a Blackman window function, and a Kaiser window function.
[0020] FIG. 3 is a flow chart showing an example of the procedure of the processing target signal generating step S2 in FIG.
[0021] As shown in FIG. 3, in the process of generating a signal to be processed S2, the signal processing device 1 first extracts a signal of a second period having a length t2 that does not overlap with the first period from the source signal (step S21).
[0022] Next, the signal processing device 1 multiplies the signal in the second period by a window function to generate a signal to be processed (step S22), and ends the signal to be processed generation step S2.
[0023] In this manner, in this embodiment, the processing target signal overlaps with the first period cut out from the source signal. The time length of the second period is equal to the length t2 of the second period. Therefore, since there is no period in which the template signal and the processing target signal contain exactly the same signal, the correlation filter process effectively reduces signal components having ergodic properties such as noise.
[0024] Moreover, the processing target signal is a signal obtained by multiplying a signal extracted from the source signal by a window function. This makes it possible to reduce the influence of noise caused by discontinuity between the first sample value and the last sample value of the processing target signal on the correlation signal. The type of the window function is not particularly limited, and examples of the window function include a Hanning window function, a rectangular window function, a Gaussian window function, a Hamming window function, a Blackman window function, and a Kaiser window function.
[0025] In this embodiment, the source signal is a digital signal with a constant sampling rate, and the template signal and the processing target signal, both of which are extracted from the source signal, have the same sampling period Δt. This ensures that the orthogonality of signal components with different frequencies is ensured in the correlation filter process, and signal components with the same frequency are correctly extracted. However, the template signal and the processing target signal may have different sampling periods. In this case, the signal processing device 1 may linearly convert at least one of the template signal and the processing target signal with respect to the time axis before the correlation signal generation step S3 so that the sampling periods of the template signal and the processing target signal are the same.
[0026] In this embodiment, the template signal is shorter than the target signal. <t2である。
[0027] FIG. 4 is a flow chart showing an example of the procedure of the correlation signal generating step S3 in FIG.
[0028] As shown in FIG. 4, in the correlation signal generation step S3, the signal processing device 1 first sets t=0 (step S31), shifts the template signal and the processing target signal by time t, and performs a product-sum operation to obtain a sample value of the correlation signal at time t (step S32).
[0029] Next, the signal processing device 1 sets t=t+Δt (step S33) and repeats steps S32 and S33 until t>t2-t1 (N in step S34). When t>t2-t1 (Y in step S34), the correlation signal generating step S3 is terminated.
[0030] As described above, in this embodiment, the correlation filter process is a product-sum operation process between the template signal and the signal to be processed, and the signal processing device 1 performs a product-sum operation process between the template signal and the signal to be processed while shifting the time t until it matches the difference between the time length t2 of the signal to be processed and the time length t1 of the template signal, to generate a correlation signal. The envelope of the correlation signal generated by the correlation filter process and the envelope of the signal to be processed are nonlinear. That is, in this embodiment, since no constraint is placed on maintaining the linearity of the envelope before and after the correlation filter process, ergodic noise contained in the signal to be processed is efficiently reduced.
[0031] The sum-of-products operation may be a finite impulse response (FIR) filter process for the signal to be processed. In this case, the coefficients of the FIR filter process are defined based on a template signal. For example, n2 sample values included in the signal to be processed are input in order to an n1-tap FIR filter in which n1 sample values included in the template signal are used as n1 coefficient values, thereby obtaining a correlation signal as an output signal of the FIR filter. Note that n1=t1 / Δt and n2=t2 / Δt. In this way, by using the sum-of-products operation as an FIR filter process, the correlation filter process can be easily realized.
[0032] In this embodiment, the source signal includes a plurality of signal components having periodicity, and the plurality of signal components have different periods. This allows the target object to have a plurality of mechanisms with different operating frequencies. The time length of the template signal is longer than the period of a first signal component among the plurality of signal components, and the difference t2-t1 between the time length of the processing target signal and the time length of the template signal is longer than the period of the first signal component. This allows the correlation signal to include one or more periods of the first signal component, and therefore allows the first signal component to be the analysis target.
[0033] Fig. 5 is a diagram showing a specific example of an object and a source signal. In the example of Fig. 5, the object is a dry pump 200, which includes a booster pump 210 and a main pump 220. The booster pump 210 includes a first pump chamber 211, and rotates a rotor (not shown) housed in the first pump chamber 211 at high speed by the action of a built-in motor, gears, bearings, etc. (not shown). As a result, the first pump chamber 211 sucks in air from outside the dry pump 200 and discharges it.
[0034] The main pump 220 includes a second pump chamber 221, and rotates a rotor (not shown) housed in the second pump chamber 221 at high speed by the action of a built-in motor, gears, bearings, etc. (not shown). As a result, the second pump chamber 221 sucks in the air discharged from the first pump chamber 211 and discharges it to the outside of the dry pump 200.
[0035] In the steady state of the dry pump 200, the motor, gears, bearings, rotors, etc. built into the booster pump 210 and the motor, gears, bearings, rotors, etc. built into the main pump 220 are operating, and these operations generate vibrations of various frequencies. The sensor 20 is installed at a position where these vibrations are transmitted. The sensor 20 may be any sensor that can detect vibrations, and may be, for example, an acceleration sensor or an angular velocity sensor. The sensor 20 is only required to be able to detect these vibrations even slightly, and there are few restrictions on where it can be installed.
[0036] The output signal of the sensor 20 is input to the analog front-end 30. The analog front-end 30 performs amplification processing and A / D (Analog to Digital) conversion processing on the output signal of the sensor 20, and outputs a source signal that is a digital time-series signal.
[0037] Figures 6 to 9 show examples of various signal waveforms obtained from the source signal shown in Figure 5. In Figures 6 to 9, the horizontal axis represents time and the vertical axis represents sample values. The sampling rate of the sensor 20 is 1 kHz.
[0038] Fig. 6 is a diagram showing an example of a signal of a first period cut out from an original signal. Fig. 7 is a diagram showing an example of a template signal generated by multiplying the signal of the first period shown in Fig. 6 by a Hanning window function. In the example of Fig. 7, the time length t1 of the template signal is 8.192 seconds, and the template signal includes 8192 samples. The signal of Fig. 6 may be used as the template signal.
[0039] Fig. 8 is a diagram showing an example of the waveform of the signal to be processed. The time length t2 of the signal to be processed is 24.576 seconds (=16.384 seconds + 8.192 seconds), and the signal to be processed contains 24576 samples. Fig. 8 shows the first 16.384 seconds of the signal.
[0040] Fig. 9 is a diagram showing an example of the waveform of a correlation signal after correlation filter processing of the template signal shown in Fig. 7 and the processing target signal shown in Fig. 8. In the example of Fig. 9, the time length of the correlation signal is The time length t2 of the signal to be processed is equal to the difference between the time length t2 of the template signal and t1, ie, 16.384 seconds, and the correlation signal contains 16384 samples.
[0041] FIG. 10 is an enlarged view of the waveform of the processing target signal shown in FIG. 8 from 0 to 0.5 seconds. FIG. 11 is an enlarged view of the waveform of the correlation signal shown in FIG. 9 from 0 to 0.5 seconds. Since the dry pump 200 is in a steady state, the source signal contains signal components of various frequencies caused by the operation of the motor, gear, bearing, rotor, etc. built in the booster pump 210 and the motor, gear, bearing, rotor, etc. built in the main pump 220. The template signal shown in FIG. 7 and the processing target signal shown in FIG. 8 are both generated based on a signal cut out from the source signal, and therefore contain signal components of various frequencies in common. Signal components of the same frequency contained in the template signal and the processing target signal are amplified by the correlation filter process. On the other hand, among the signal components not correlated with the operation of the dry pump 200, noise having ergodicity is attenuated by the correlation filter process. Therefore, the signal components correlated with the operation of the dry pump 200 are unclear in the processing target signal shown in FIG. 10, whereas they are clear in the correlation signal shown in FIG. 11.
[0042] 1-2.Signal processing device Fig. 12 is a diagram showing an example of the configuration of a signal processing device 1 that realizes the above-mentioned signal processing method. As shown in Fig. 12, the signal processing device 1 includes a processing circuit 10, a sensor 20, an analog front end 30, a memory circuit 40, an operation unit 50, a display unit 60, a sound output unit 70, and a communication unit 80. Note that the signal processing device 1 may be configured such that some of the components in Fig. 12 other than the processing circuit 10 and the sensor 20 are omitted or changed, or other components are added.
[0043] The sensor 20 detects a physical quantity occurring in a steady state of an object, and outputs a signal having a magnitude corresponding to the detected physical quantity. The output signal of the sensor 20 is input to the analog front end 30.
[0044] The analog front end 30 performs amplification, A / D conversion, and other processes on the output signal of the sensor 20, and outputs a digital time-series signal.
[0045] The processing circuit 10 performs signal processing on the source signal, which is a digital time-series signal output from the analog front end 30. Specifically, the processing circuit 10 executes a signal processing program 41 stored in the storage circuit 40, and performs various calculation processing on the source signal. In addition, the processing circuit 10 performs various processes according to operation signals from the operation unit 50, a process of transmitting a display signal for displaying various information on the display unit 60, a process of transmitting a sound signal for generating various sounds to the sound output unit 70, a process of controlling the communication unit 80 for performing data communication with an external device, and the like. The processing circuit 10 is realized by, for example, a CPU (Central Processing Unit) or a DSP (Digital Signal Processor).
[0046] The processing circuit 10 executes the signal processing program 41 to function as a template signal generating circuit 11, a processing target signal generating circuit 12, and a correlation signal generating circuit 13. That is, the signal processing device 1 includes the template signal generating circuit 11, the processing target signal generating circuit 12, and the correlation signal generating circuit 13.
[0047] The template signal generation circuit 11 generates a template signal 42, which is a time-series signal, based on a source signal, which is a time-series signal corresponding to a physical quantity generated in a steady state of an object. That is, the template signal generation circuit 11 executes the template signal generation step S1 in Fig. 1, specifically, steps S11 and S12 in Fig. 2. The template signal 42 generated by the template signal generation circuit 11 is stored in the storage circuit 40.
[0048] The processed signal generation circuit 12 generates a processed signal 43, which is a time-series signal, based on the source signal. That is, the processed signal generation circuit 12 executes the processed signal generation step S2 in Fig. 1, specifically steps S21 and S22 in Fig. 3. The processed signal 43 generated by the processed signal generation circuit 12 is stored in the storage circuit 40.
[0049] The correlation signal generation circuit 13 performs correlation filtering on the template signal 42 generated by the template signal generation circuit 11 and the processing target signal 43 generated by the processing target signal generation circuit 12 to generate a correlation signal 44 which is a time-series signal. That is, the correlation signal generation circuit 13 executes the correlation signal generation step S3 in Fig. 1, specifically steps S31 to S36 in Fig. 4. The correlation signal 44 generated by the correlation signal generation circuit 13 is stored in the storage circuit 40.
[0050] The storage circuitry 40 has a ROM (Read Only Memory) and a RAM (Random Access Memory), not shown. The ROM stores various programs such as a signal processing program 41 and predetermined data, and the RAM stores signals and data generated by the processing circuitry 10, such as a template signal 42, a processing target signal 43, and a correlation signal 44. The RAM is also used as a working area for the processing circuitry 10, and stores programs and data read from the ROM, data input from an operation unit 50, and signals and data temporarily generated by the processing circuitry 10.
[0051] The operation unit 50 is an input device including operation keys, button switches, etc., and outputs to the processing circuit 10 an operation signal in response to an operation by a user.
[0052] The display unit 60 is a display device configured with an LCD (Liquid Crystal Display) or the like, and displays various information based on a display signal output from the processing circuit 10. The display unit 60 may be provided with a touch panel that functions as the operation unit 50. For example, the display unit 60 may display a waveform image of the correlation signal 44 based on the display signal output from the processing circuit 10.
[0053] The sound output unit 70 is configured with a speaker or the like, and generates various sounds based on the sound signal output from the processing circuit 10. For example, the sound output unit 70 may generate sounds indicating the start or end of signal processing based on the sound signal output from the processing circuit 10.
[0054] The communication unit 80 performs various controls for establishing data communication between the processing circuit 10 and an external device. For example, the communication unit 80 may transmit the correlation signal 44 to the external device.
[0055] At least a part of the template signal generating circuit 11, the processing target signal generating circuit 12, and the correlation signal generating circuit 13 may be realized by dedicated hardware. The signal processing device 1 may be a single device or may be configured by multiple devices. For example, the sensor 20 and the analog front end 30 may be included in a first device, and the processing circuit 10, the memory circuit 40, the operation unit 50, the display unit 60, the sound output unit 70, and the communication unit 80 may be included in a second device separate from the first device. For example, the processing circuit 10 and the memory circuit 40 may be realized by a device such as a cloud server, which generates the correlation signal 44 and transmits the generated correlation signal 44 to a terminal including the operation unit 50, the display unit 60, the sound output unit 70, and the communication unit 80 via a communication line.
[0056] 1-3.Effects According to the first embodiment described above, the template signal generated based on the source signal and the processing target signal both contain signal components having periodicity correlated with physical quantities occurring in the steady state of the object to be signal processed. Therefore, by performing correlation filter processing on the template signal and the processing target signal, the signal components having this periodicity are reinforced, and it is possible to detect the signal components having periodicity contained in the processing target signal without performing envelope processing. can.
[0057] Furthermore, among the signal components that are not correlated with the physical quantity, noises having ergodicity weaken each other. Therefore, an effect equivalent to synchronously adding the signal components having periodicity contained in the template signal for the number of periods can be obtained, and a correlation signal having a high S / N ratio can be obtained even if the amplitude of the signal components having periodicity contained in the processing target signal is small. Therefore, according to the first embodiment, restrictions on the installation location of the acquisition unit for the target object and the physical quantity are relaxed.
[0058] Furthermore, according to the first embodiment, signal processing can be performed when the object is in a steady state, so that a desired correlation signal can be obtained without changing the motion state of the object.
[0059] 2. Second embodiment In the following, in the second embodiment, the same components as those in the first embodiment are given the same reference numerals, explanations that overlap with the first embodiment are omitted or simplified, and differences from the first embodiment are mainly described.
[0060] The steps of the signal processing method of the second embodiment are the same as those in Fig. 1, and therefore are not shown in the figure. In the signal processing method of the second embodiment, the steps of the template signal generating step S1 and the processing target signal generating step S2 are different from those in the first embodiment, but the step of the correlation signal generating step S3 is the same as that in the first embodiment.
[0061] FIG. 13 is a flowchart showing an example of the procedure of the template signal generating step S1 in the second embodiment.
[0062] As shown in FIG. 13, in the template signal generating step S1, the signal processing device 1 first cuts out a signal of a first period having a length t1 from the source signal (step S101).
[0063] Next, the signal processing device 1 multiplies the signal of the first period by a window function, removes the DC component, and generates a template signal (step S102), and ends the template signal generating step S1. For example, the signal processing device 1 may perform high-pass filtering on the signal obtained by multiplying the signal of the first period by the window function, to remove the DC component.
[0064] FIG. 14 is a flowchart showing an example of the procedure of the processing target signal generating step S2 in the second embodiment.
[0065] As shown in FIG. 14, in the process of generating a signal to be processed S2, the signal processing device 1 first cuts out a signal of a second period having a length t2 that does not overlap with the first period from the source signal (step S201).
[0066] Next, the signal processing device 1 multiplies the signal in the second period by a window function, removes the DC component, and generates a signal to be processed (step S202), and ends the signal to be processed generation step S2.
[0067] An example of the configuration of the signal processing device 1 in the second embodiment is the same as that in Fig. 12, and therefore will not be illustrated. However, the template signal generation circuit 11 executes steps S101 and S102 in Fig. 13. Furthermore, the processing target signal generation circuit 12 executes steps S201 and S202 in Fig. 14. Other configurations and functions of the signal processing device 1 in the second embodiment are the same as those in the first embodiment, and therefore will not be described.
[0068] In this way, according to the procedure of Fig. 13 and the procedure of Fig. 14, the template signal and the processing target signal are signals from which the DC component has been removed. Only one of the signal processing circuits 40 and 41 may be a signal from which the DC component has been removed. That is, in this embodiment, at least one of the template signal and the processing target signal is a signal from which the DC component has been removed. Therefore, the correlation signal obtained by the correlation filter process is biased by the correlation amount of the DC component contained in the template signal and the processing target signal, but since the bias is reduced, the maximum value of the correlation signal is reduced and the data amount of the correlation signal is reduced. Therefore, according to the second embodiment, the size of the storage circuit 40 for storing the correlation signal in the signal processing device 1 can be reduced.
[0069] 3. Third embodiment Hereinafter, for the third embodiment, components similar to those in the first or second embodiment are given the same symbols, explanations that overlap with the first or second embodiment are omitted or simplified, and mainly the differences from the first and second embodiments are described.
[0070] The steps of the signal processing method of the third embodiment are the same as those in Fig. 1, and therefore are not shown in the figure. In the signal processing method of the third embodiment, the steps of the template signal generating step S1 and the correlation signal generating step S3 are different from those of the first and second embodiments, and the step of the processing target signal generating step S2 is the same as that of the first or second embodiment.
[0071] FIG. 15 is a flowchart showing an example of the procedure of the template signal generating step S1 in the third embodiment.
[0072] As shown in FIG. 15, in the template signal generating step S1, the signal processing device 1 first sets an integer i=0 (step S111) and cuts out a signal of a first period having a length t1 from the source signal (step S112).
[0073] Next, the signal processing device 1 multiplies the signal in the first period by a window function, removes the DC component, and generates the i-th template signal (step S113).
[0074] Next, the signal processing device 1 repeats steps S112 and S113 by setting i to i+1 (step S115) until i=N (N in step S114). Then, when i=N (Y in step S114), the signal processing device 1 ends the template signal generating step S1.
[0075] Thus, in this embodiment, the signal processing device 1 generates the first to N-th template signals based on the source signal in the template signal generating step S1.
[0076] FIG. 16 is a flowchart showing an example of the procedure of the correlation signal generating step S3 in the third embodiment.
[0077] As shown in FIG. 16, in the correlation signal generation step S3, the signal processing device 1 first sets integer i=1 (step S311), then sets t=0 (step S312), shifts the i-th template signal and the signal to be processed by time t, and performs a product-sum operation to obtain the sample value of the i-th correlation signal at time t (step S313).
[0078] Next, the signal processing device 1 sets t=t+Δt (step S314) and repeats steps S313 and S314 until t>t2-t1 (N in step S315). When t>t2-t1 (Y in step S315), the signal processing device 1 sets i=i+1 (step S317) unless i=N (N in step S316) and repeats the processes from step S312 onwards. Here, N is a predetermined integer equal to or greater than 2.
[0079] The signal processing device 1 repeats the process from i=i+1 (step S316) until i=N (step S316). S317), and repeats steps S312 to S315. Then, when i=N (Y in step S316), the signal processing device 1 adds, averages, or weighted averages the first to Nth correlation signals to generate a correlation signal (step S318), and ends the correlation signal generation step S3.
[0080] As described above, in the present embodiment, in the correlation signal generating step S3, the signal processing device 1 performs correlation filter processing between the signal to be processed and each of the first to Nth template signals, which are time-series signals generated based on the source signal, to generate first to Nth correlation signals, and then adds, averages or takes a weighted average of the first to Nth correlation signals to generate a correlation signal.
[0081] An example of the configuration of the signal processing device 1 in the third embodiment is the same as that in Fig. 12, and therefore illustration thereof is omitted. However, the template signal generation circuit 11 executes steps S111 to S115 in Fig. 15. Furthermore, the correlation signal generation circuit 13 executes steps S311 to S319 in Fig. 16. Other configurations and functions of the signal processing device 1 in the third embodiment are the same as those in the first or second embodiment, and therefore description thereof is omitted.
[0082] According to the third embodiment described above, the ergodic noise contained in the first to Nth correlation signals is reduced by the correlation filter process, and the ergodic noise is further reduced by adding, averaging or weighted averaging the first to Nth correlation signals, thereby obtaining a correlation signal having a higher S / N ratio.
[0083] 4. Fourth embodiment Hereinafter, for the fourth embodiment, components similar to those of any of the first to third embodiments will be given the same symbols, explanations that overlap with those of any of the first to third embodiments will be omitted or simplified, and the following will mainly describe the contents that are different from any of the first to third embodiments.
[0084] FIG. 17 is a flow chart showing the procedure of the signal processing method according to the fourth embodiment.
[0085] As shown in FIG. 17, the signal processing device 1 first performs a template signal generating step S1, a processing target signal generating step S2, and a correlation signal generating step S3, similarly to any of the first to third embodiments.
[0086] Then, the signal processing device 1 repeats steps S2 and S3 until the signal processing is completed (N in step S4).
[0087] In this manner, in this embodiment, the processing target signal is updated but the template signal is not updated until the signal processing is completed.
[0088] In the signal processing method of the fourth embodiment, the procedure of the template signal generating step S1, the procedure of the correlation signal generating step S3, and the procedure of the processing target signal generating step S2 are the same as those of any of the first to third embodiments.
[0089] The signal processing device 1 may perform the processing target signal generating step S2 using a template signal generated in advance. That is, in this embodiment, the template signal generating step S1 may be omitted.
[0090] An example of the configuration of the signal processing device 1 in the fourth embodiment is the same as that in Fig. 12, and therefore is not illustrated. However, the processing circuit 10 executes steps S1 to S4 in Fig. 17. Other configurations and functions of the signal processing device 1 in the fourth embodiment are the same as those of any of the first to third embodiments, and therefore are not described.
[0091] According to the fourth embodiment described above, since the template signal is not updated, when the period or intensity of a signal component having periodicity contained in the signal to be processed changes due to a change over time of the object, the intensity of the signal component contained in the correlation signal changes. Therefore, according to the fourth embodiment, it is possible to easily grasp the change over time of the object and its cause.
[0092] 5. Fifth embodiment Hereinafter, for the fifth embodiment, components similar to those of any of the first to fourth embodiments will be given the same symbols, explanations that overlap with those of any of the first to fourth embodiments will be omitted or simplified, and the following will mainly describe the differences from any of the first to fourth embodiments.
[0093] FIG. 18 is a flowchart showing the procedure of the signal processing method according to the fifth embodiment.
[0094] As shown in FIG. 18, the signal processing device 1 first performs a template signal generating step S1, a processing target signal generating step S2, and a correlation signal generating step S3, similarly to any of the first to third embodiments.
[0095] Next, the signal processing device 1 performs a waveform analysis step S5 of analyzing the waveform of the correlation signal.
[0096] Then, the signal processing device 1 repeats steps S1, S2, S3, and S5 until the signal processing is completed (N in step S6).
[0097] In the procedure of FIG. 18, the template signal and the processing target signal are updated until the signal processing is completed.
[0098] FIG. 19 is a flowchart showing another procedure of the signal processing method according to the fifth embodiment.
[0099] As shown in FIG. 19, the signal processing device 1 first performs a template signal generating step S1, a processing target signal generating step S2, and a correlation signal generating step S3, similarly to the fourth embodiment.
[0100] Next, the signal processing device 1 performs a waveform analysis step S5 of analyzing the waveform of the correlation signal.
[0101] Then, the signal processing device 1 repeats steps S2, S3, and S5 until the signal processing is completed (N in step S6).
[0102] In the procedure of FIG. 19, the processing target signal is updated but the template signal is not updated until the signal processing is completed.
[0103] 18 and 19, in this embodiment, the signal processing device 1 performs a waveform analysis step S5 of analyzing the waveform of a correlation signal. In the waveform analysis step S5, the signal processing device 1 may perform analysis by focusing on the frequency or period of a signal having periodicity included in the correlation signal.
[0104] FIG. 20 is a flow chart showing an example of the procedure of the waveform analysis step S5 in FIG. 18 or FIG.
[0105] 20, first, the signal processing device 1 performs frequency analysis on the correlation signal (step S51). The frequency analysis may be, for example, a fast Fourier transform (FFT), a wavelet transform, an autocorrelation analysis, or the like.
[0106] Next, the signal processing device 1 calculates the frequency and intensity of the periodic signal contained in the correlation signal based on the frequency analysis result in step S51 (step S52), and ends the processing of the waveform analysis step S5.
[0107] As described above, in this embodiment, in the waveform analysis step S5, the signal processing device 1 performs frequency analysis on the correlation signal to calculate the frequency of the signal having periodicity that is included in the correlation signal.
[0108] Fig. 21 is a diagram showing an example of a frequency spectrum obtained by performing an FFT on a correlation signal in the waveform analysis step S5 for a relatively new dry pump 200 that has been in operation for about several months. Fig. 22 is a diagram showing, as a comparative example, an example of a frequency spectrum obtained when an FFT is directly performed on a signal to be processed. In Figs. 21 and 22, the horizontal axis represents frequency and the vertical axis represents intensity.
[0109] Comparing Fig. 21 and Fig. 22, it can be seen that the peaks of the signal components of various frequencies are clearer in Fig. 21 than in Fig. 22, and a correlation signal with a high S / N ratio (Signal to Noise Ratio) in which noise is significantly reduced can be obtained by performing correlation filter processing on the signal to be processed. Therefore, the signal processing device 1 can calculate the frequency and intensity of various signal components based on the frequency spectrum in Fig. 21.
[0110] For example, the frequency at which the inner ring of the bearing rotates in the main pump 220 is the inner ring rotation frequency f rm The design value of is known, and in FIG. 21, the signal component whose intensity peaks near this design value is the inner ring rotation frequency f rm Similarly, the frequency at which the inner ring of the bearing rotates in the booster pump 210 corresponds to the fundamental wave of the inner ring rotation frequency f rb The design value of is known, and in FIG. 21, the signal component whose intensity peaks near this design value is the inner ring rotation frequency f rm Therefore, the signal processing device 1 detects the inner ring rotation frequency f of the main pump 220 in the frequency spectrum of FIG. rm The fundamental wave and its second to fifth harmonics, and the inner ring rotation frequency f rb It is possible to calculate the intensity corresponding to each frequency of the fundamental wave and its second to fifth harmonics.
[0111] The balance of the bearing rotation and vibration can be determined by whether the odd or even harmonics are strong. When the odd harmonics are strong, the rotation and vibration are symmetrical, whereas when the even harmonics are strong, the rotation and vibration are asymmetrical. Therefore, it is preferable to be able to detect many harmonics, since it is possible to determine changes in the symmetry of the rotation and vibration, i.e., deterioration of the bearings, from the intensity and changes over time of each harmonic. For example, the inner ring rotation frequency f of the main pump 220 is rm and the inner ring rotation frequency f of the booster pump 210 rb The fifth harmonics of are unclear in Fig. 22 but are very clear in Fig. 21. Therefore, the signal processing device 1 can reliably detect these fifth harmonics as well by performing correlation filtering.
[0112] As another example, Fig. 23 shows a frequency spectrum obtained by performing an FFT on the correlation signal in the waveform analysis step S5 for a relatively old dry pump 200 with an operating time of nearly 10 years. Also, Fig. 24 shows, as a comparative example, an example of a frequency spectrum obtained when an FFT is directly performed on the signal to be processed. In Figs. 23 and 24, the horizontal axis represents frequency and the vertical axis represents intensity.
[0113] 23 and 24, for example, the inner ring rotation frequency f rm and the inner ring rotation frequency f of the booster pump 210 rb The fifth harmonics of are very unclear in Fig. 24 but are very clear in Fig. 23. Therefore, the signal processing device 1 can reliably detect these fifth harmonics as well by performing correlation filtering.
[0114] 21 and 23, the inner ring rotation frequency f rm The fundamental wave and its second to fifth harmonics are intensified, and the inner ring rotation frequency f rbThe intensities of the fundamental wave and its second to fifth harmonics are significantly different between FIG. 21 and FIG. 23. Also, the inner ring rotation frequency f rm 21 and 23, the frequencies of the fundamental waves are also slightly different. This indicates that the main pump 220 and the booster pump 210 of the relatively old dry pump 200 have deteriorated due to changes over time. Although the target dry pumps 200 are different in Fig. 21 and Fig. 23, the signal processing device 1 can determine the deterioration of one dry pump 200 over time by periodically performing a waveform analysis on that dry pump 200.
[0115] Fig. 25 is a configuration example of a signal processing device 1 that realizes the signal processing method of the fifth embodiment. As shown in Fig. 25, the signal processing device 1 includes a processing circuit 10, a sensor 20, an analog front end 30, a memory circuit 40, an operation unit 50, a display unit 60, a sound output unit 70, and a communication unit 80. Note that the signal processing device 1 may be configured such that some of the components in Fig. 25 are omitted or changed, or other components are added.
[0116] The configurations and functions of the sensor 20, analog front end 30, memory circuit 40, operation unit 50, display unit 60, sound output unit 70 and communication unit 80 are the same as those of any of the first to fourth embodiments, and therefore will not be described.
[0117] The processing circuit 10 executes the signal processing program 41 to function as a template signal generating circuit 11, a processing target signal generating circuit 12, a correlation signal generating circuit 13, and a waveform analyzing circuit 14. That is, the signal processing device 1 includes the template signal generating circuit 11, the processing target signal generating circuit 12, the correlation signal generating circuit 13, and the waveform analyzing circuit 14.
[0118] The functions of the template signal generating circuit 11, the processing target signal generating circuit 12 and the correlation signal generating circuit 13 are similar to those of any of the first to fourth embodiments, and therefore a description thereof will be omitted.
[0119] The waveform analysis circuit 14 analyzes the waveform of the correlation signal 44 generated by the correlation signal generation circuit 13. The waveform analysis circuit 14 may focus on the frequency or period of the signal having periodicity included in the correlation signal 44 for analysis. In this embodiment, the waveform analysis circuit 14 performs frequency analysis on the correlation signal 44 to calculate the frequency and intensity of the signal having periodicity included in the correlation signal 44. That is, the waveform analysis circuit 14 executes the waveform analysis step S5 in FIG. 18 or FIG. 19, specifically steps S41 and S42 in FIG. 20.
[0120] The display unit 60 may display information on the analysis result by the waveform analysis circuit 14 based on a display signal output from the processing circuit 10. The sound output unit 70 may generate a sound indicating the start or end of analysis by the waveform analysis circuit 14 based on a sound signal output from the processing circuit 10. The communication unit 80 may transmit information on the analysis result by the waveform analysis circuit 14 to an external device. The information on the analysis result by the waveform analysis circuit 14 is information such as the frequency of a periodic signal included in the correlation signal 44.
[0121] According to the fifth embodiment described above, a correlation signal having a high S / N ratio can be obtained by the correlation filter process, so that the waveform analysis of the correlation signal can be performed with high accuracy. In particular, according to the fifth embodiment, a signal having periodicity is emphasized by the correlation filter process, so that the waveform analysis in the frequency domain can be performed with high accuracy by the frequency spectrum of the correlation signal.
[0122] 6. Sixth embodiment Hereinafter, for the sixth embodiment, components similar to those of any of the first to fifth embodiments will be given the same symbols, explanations that overlap with those of any of the first to fifth embodiments will be omitted or simplified, and mainly the contents that differ from any of the first to fifth embodiments will be described.
[0123] The steps of the signal processing method of the sixth embodiment are the same as those in Fig. 18 or 19, and therefore are not shown in the figures. In the signal processing method of the sixth embodiment, the steps of the template signal generating step S1, the processing target signal generating step S2, and the correlation signal generating step S3 are the same as those in any of the first to fifth embodiments.
[0124] As in the fifth embodiment, in the sixth embodiment, the signal processing device 1 performs a waveform analysis step S5 of analyzing the waveform of the correlation signal. In the waveform analysis step S5, the signal processing device 1 may perform analysis by focusing on the frequency or period of a signal having periodicity included in the correlation signal. However, in the sixth embodiment, the procedure of the waveform analysis step S5 is different from that in the fifth embodiment.
[0125] FIG. 26 is a flowchart showing an example of the procedure of the waveform analysis step S5 in the sixth embodiment.
[0126] 26, first, the signal processing device 1 extracts a plurality of feature points of a signal having periodicity included in a correlation signal (step S501). The feature points may be local maximum points, local minimum points, inflection points, zero crossing points, etc.
[0127] Next, the signal processing device 1 calculates the period of the signal having periodicity contained in the correlation signal based on the time intervals of the plurality of feature points extracted in step S501 (step S502), and ends the processing of the waveform analysis step S5.
[0128] As described above, in this embodiment, in the waveform analysis step S5, the signal processing device 1 calculates the period of the periodic signal included in the correlation signal based on the time intervals between multiple feature points of the periodic signal.
[0129] 27 is an enlarged view of the waveform of the correlation signal from 0 to 0.1 seconds shown in FIG. 11. For example, in the waveform analysis step S5, the signal processing device 1 may calculate the time interval T1 between two maximum points P1 and P2 that are separated by 10 periods of the main signal component included in the correlation signal, and calculate the period T=T1×1 / 10 of the main signal component. The signal processing device 1 may also calculate the time interval T2 between two zero crossing points P3 and P4 that are separated by 10 periods of the main signal component included in the correlation signal, and calculate the period T=T2×1 / 10 of the main signal component. Furthermore, the signal processing device 1 may calculate the frequency f=1 / T of the main signal component. For example, the signal processing device 1 may compare the frequency f of the main signal component with a design value to determine the deterioration of the dry pump 200.
[0130] An example of the configuration of the signal processing device 1 in the sixth embodiment is the same as that in Fig. 25, and therefore is not illustrated. However, the waveform analysis circuit 14 executes steps S501 and S502 in Fig. 26. Other configurations and functions of the signal processing device 1 in the sixth embodiment are the same as those in any of the first to fifth embodiments, and therefore are not described.
[0131] According to the sixth embodiment described above, a correlation signal having a high S / N ratio can be obtained by the correlation filter process, so that the waveform analysis of the correlation signal can be performed with high accuracy. In particular, according to the sixth embodiment, a signal having periodicity is emphasized by the correlation filter process, so that the waveform analysis in the time domain can be performed with high accuracy.
[0132] 7. Seventh embodiment Hereinafter, for the seventh embodiment, components similar to those of any of the first to sixth embodiments will be given the same symbols, explanations that overlap with those of any of the first to sixth embodiments will be omitted or simplified, and mainly the contents that differ from any of the first to sixth embodiments will be described.
[0133] FIG. 28 is a flowchart showing the procedure of the signal processing method according to the seventh embodiment.
[0134] As shown in FIG. 28, the signal processing device 1 first performs a template signal generating step S1, a processing target signal generating step S2, and a correlation signal generating step S3, similarly to any of the first to sixth embodiments.
[0135] Next, the signal processing device 1 performs a waveform analysis step S5 of analyzing the waveform of the correlation signal, similarly to the fifth or sixth embodiment.
[0136] Next, the signal processing device 1 performs an abnormality determination step S7 for determining the presence or absence of an abnormality in the target object based on the analysis result of the waveform analysis step S5.
[0137] Then, the signal processing device 1 repeats steps S1, S2, S3, S5, and S7 until the signal processing is completed (N in step S8).
[0138] In the procedure of FIG. 28, the template signal and the processing target signal are updated until the signal processing is completed.
[0139] FIG. 29 is a flowchart showing another procedure of the signal processing method according to the seventh embodiment.
[0140] As shown in FIG. 29, the signal processing device 1 first performs a template signal generating step S1, a processing target signal generating step S2, and a correlation signal generating step S3, similarly to any of the first to sixth embodiments.
[0141] Next, the signal processing device 1 performs a waveform analysis step S5 of analyzing the waveform of the correlation signal, similarly to the fifth or sixth embodiment.
[0142] Next, the signal processing device 1 performs an abnormality determination step S7 for determining the presence or absence of an abnormality in the target object based on the analysis result of the waveform analysis step S5.
[0143] Then, the signal processing device 1 repeats steps S2, S3, S5, and S7 until the signal processing is completed (N in step S8).
[0144] In the procedure of FIG. 29, the processing target signal is updated but the template signal is not updated until the signal processing is completed.
[0145] As shown in Fig. 28 and Fig. 29, in this embodiment, the signal processing device 1 performs an abnormality determination step S7 for determining whether or not the object is abnormal based on the analysis result of the waveform analysis step S5. In the abnormality determination step S7, for example, the signal processing device 1 may determine that the object is normal if the period or frequency of the signal having periodicity included in the correlation signal calculated in the waveform analysis step S5 is within a predetermined range, and may determine that the object is abnormal if the period or frequency is not within the predetermined range. In addition, the signal processing device 1 may determine that the object is normal if the amount of change in the period or frequency from the initial operation of the object is within a predetermined range, and may determine that the object is abnormal if the amount of change is not within the predetermined range.
[0146] Fig. 30 shows an example of the configuration of a signal processing device 1 that realizes the signal processing method of the seventh embodiment. As shown in Fig. 30, the signal processing device 1 includes a processing circuit 10, a sensor 20, an analog front end 30, a memory circuit 40, an operation unit 50, a display unit 60, a sound output unit 70, and a communication unit 80. Note that the signal processing device 1 may be configured such that some of the components in Fig. 30 are omitted or changed, or other components are added.
[0147] The configurations and functions of the sensor 20, analog front end 30, memory circuit 40, operation unit 50, display unit 60, sound output unit 70 and communication unit 80 are the same as those of any of the first to sixth embodiments, and therefore will not be described.
[0148] The processing circuit 10 executes the signal processing program 41 to function as a template signal generating circuit 11, a processing target signal generating circuit 12, a correlation signal generating circuit 13, a waveform analysis circuit 14, and an abnormality determination circuit 15. In other words, the signal processing device 1 includes the template signal generating circuit 11, the processing target signal generating circuit 12, the correlation signal generating circuit 13, the waveform analysis circuit 14, and the abnormality determination circuit 15.
[0149] The functions of the template signal generating circuit 11, the processing target signal generating circuit 12, and the correlation signal generating circuit 13 are the same as those of any of the first to fourth embodiments, and therefore their description will be omitted. Also, the function of the waveform analyzing circuit 14 is the same as that of the fifth or sixth embodiment, and therefore its description will be omitted.
[0150] The abnormality determination circuit 15 determines whether or not the object is abnormal based on the analysis result by the waveform analysis circuit 14. For example, the abnormality determination circuit 15 may determine that the object is normal when the period or frequency of the signal having periodicity included in the correlation signal 44 calculated by the waveform analysis circuit 14 is within a predetermined range, and may determine that the object is abnormal when the period or frequency is not within the predetermined range. The abnormality determination circuit 15 may also determine that the object is normal when the amount of change in the period or frequency from the initial operation of the object is within a predetermined range, and may determine that the object is abnormal when the amount of change is not within the predetermined range. That is, the abnormality determination circuit 15 executes the abnormality determination step S7 in FIG. 28 or FIG. 29.
[0151] The display unit 60 may display information on the result of the determination by the abnormality determination circuit 15 based on a display signal output from the processing circuit 10. Furthermore, the sound output unit 70 may generate a sound indicating the result of the determination by the abnormality determination circuit 15 based on a sound signal output from the processing circuit 10. Furthermore, the communication unit 80 may transmit information on the result of the determination by the abnormality determination circuit 15 to an external device.
[0152] According to the seventh embodiment described above, a correlation signal having a high S / N ratio can be obtained by correlation filtering, so that waveform analysis of the correlation signal can be performed with high accuracy. Based on the result of the accurately performed waveform analysis of the correlation signal, it is possible to accurately determine whether or not an abnormality exists in the target object.
[0153] 8. Variations In each of the above embodiments, the template signal is a signal obtained by multiplying a signal extracted from the source signal by a window function, but the template signal may be the signal itself extracted from the source signal.
[0154] In the first and third to seventh embodiments, the update frequency of the template signal and the update frequency of the processing target signal are the same, but they may be different from each other. The update frequency of the template signal may be less than the update frequency of the processing target signal, for example, the update frequency of the template signal may be 1 / M of the update frequency of the processing target signal, where M is an integer equal to or greater than 2. stomach.
[0155] In addition, in the above first and third to seventh embodiments, the signal processing device 1 generates the template signal, but the template signal may be generated by a device other than the signal processing device 1 and written into the memory circuit 40 of the signal processing device 1.
[0156] The above-described embodiment and modifications are merely examples, and the present invention is not limited to these. For example, the embodiments and modifications can be appropriately combined.
[0157] The present invention includes configurations that are substantially the same as the configurations described in the embodiments (for example, configurations with the same functions, methods and results, or configurations with the same purpose and effect). The present invention also includes configurations in which non-essential parts of the configurations described in the embodiments are replaced. The present invention also includes configurations that achieve the same effects as the configurations described in the embodiments or that can achieve the same purpose. The present invention also includes configurations in which publicly known technology is added to the configurations described in the embodiments.
[0158] The following can be derived from the above-described embodiment and modifications.
[0159] One aspect of the signal processing method includes: a process-target signal generating step of generating a process target signal which is a time-series signal based on a source signal which is a time-series signal corresponding to a physical quantity generated in a steady state of the object; and a correlation signal generating step of performing correlation filtering between the processing target signal and a template signal, which is a time-series signal generated based on the source signal, to generate a correlation signal, which is a time-series signal.
[0160] In this signal processing method, the template signal generated based on the source signal and the processing target signal both contain signal components having periodicity correlated with physical quantities occurring in the steady state of the object being processed, and therefore, by correlation filter processing of the template signal and the processing target signal, the signal components having this periodicity are reinforced, making it possible to detect the signal components having periodicity contained in the processing target signal without performing envelope processing.
[0161] Furthermore, among the signal components that are not correlated with the physical quantity, noises that have ergodicity are destructive. Therefore, an effect equivalent to synchronously adding the periodic signal components contained in the template signal for the number of periods can be obtained, and a correlation signal with a high S / N ratio can be obtained even if the amplitude of the periodic signal components contained in the processing target signal is small. Therefore, this signal processing method eases restrictions on the installation location of the target object and the physical quantity acquisition unit.
[0162] Furthermore, according to this signal processing method, signal processing can be performed when the object is in a steady state, so that a desired correlation signal can be obtained without changing the motion state of the object.
[0163] In one aspect of the signal processing method, the source signal includes a plurality of signal components having periodicity; The plurality of signal components may have different periods.
[0164] According to this signal processing method, the target object can have a plurality of mechanisms with different operating frequencies.
[0165] In one aspect of the signal processing method, The template signal has a period longer than a period of a first signal component of the plurality of signal components. and shorter than the signal to be processed; A difference in length between the signal to be processed and the template signal may be longer than a period of the first signal component.
[0166] According to this signal processing method, the correlation signal contains one or more periods of the first signal component, so that the first signal component can be the subject of analysis.
[0167] In one aspect of the signal processing method, The template signal may be a signal obtained by multiplying a signal extracted from the source signal by a window function.
[0168] According to this signal processing method, it is possible to reduce the influence of noise caused by discontinuity between the first sample value and the last sample value of the template signal on the correlation signal.
[0169] One aspect of the signal processing method includes: The method may further include a template signal generating step of generating the template signal based on the source signal.
[0170] In one aspect of the signal processing method, The template signal may be updated.
[0171] According to this signal processing method, even if the period of a periodic signal component included in the signal to be processed changes due to a change over time of the object, it is possible to generate a correlation signal including the signal component. Furthermore, the shorter the update interval of the template signal, the greater the intensity of the signal component included in the correlation signal can be.
[0172] In one aspect of the signal processing method, The template signal may not be updated.
[0173] According to this signal processing method, when the period or intensity of a periodic signal component contained in the signal to be processed changes due to changes in the object over time, the intensity of the signal component contained in the correlation signal changes, making it easier to grasp the changes in the object over time and their causes.
[0174] In one aspect of the signal processing method, the template signal is a signal based on a signal of a first period extracted from the source signal, The target signal may be a signal based on a signal of a second period that is extracted from the source signal and does not overlap with the first period.
[0175] According to this signal processing method, there is no period in which the template signal and the processing target signal contain exactly the same signal, so that signal components having ergodic properties such as noise are effectively reduced by correlation filtering.
[0176] In one aspect of the signal processing method, The template signal and the processing target signal may have the same sampling period.
[0177] According to this signal processing method, the orthogonality of signal components with different frequencies is guaranteed in the correlation filter processing, and signal components with the same frequency can be correctly extracted.
[0178] In one aspect of the signal processing method, At least one of the template signal and the processing target signal may be a signal from which a direct current component has been removed.
[0179] In this signal processing method, the correlation signal obtained by the correlation filter process is biased by the correlation of the DC components contained in the template signal and the processing target signal, but since the bias is reduced, the maximum value of the correlation signal becomes smaller and the data amount of the correlation signal becomes smaller. Therefore, according to this signal processing method, the size of the memory for holding the correlation signal can be reduced.
[0180] In one aspect of the signal processing method, the correlation filtering process is a multiply-and-accumulate process of the template signal and the processing target signal, The envelope of the correlation signal and the envelope of the signal to be processed may be non-linear.
[0181] According to this signal processing method, since no constraint is placed on maintaining the linearity of the envelope before and after correlation filtering, it is possible to efficiently reduce ergodic noise contained in the signal to be processed.
[0182] In one aspect of the signal processing method, the multiply-and-accumulate operation is an FIR filter process for the processing target signal, The coefficients of the FIR filtering may be defined based on the template signal.
[0183] According to this signal processing method, correlation filtering can be easily realized.
[0184] In one aspect of the signal processing method, For an integer N greater than or equal to 2, In the correlation signal generating step, The template signal may be used as a first template signal, and correlation filtering may be performed between the processing target signal and each of first to Nth template signals, which are time-series signals generated based on the source signal, to generate first to Nth correlation signals, and the correlation signal may be generated by adding, averaging or weighted averaging the first to Nth correlation signals.
[0185] According to this signal processing method, the correlation filter process reduces the ergodic noise contained in the first to Nth correlation signals, and the ergodic noise can be further reduced by adding, averaging or weighted averaging the first to Nth correlation signals, thereby obtaining a correlation signal with a higher S / N ratio.
[0186] One aspect of the signal processing method includes: The step of analyzing the waveform of the correlation signal may be included.
[0187] According to this signal processing method, a correlation signal having a high S / N ratio can be obtained by correlation filtering, so that waveform analysis of the correlation signal can be performed with high accuracy.
[0188] In one aspect of the signal processing method, The waveform analysis step may focus on the frequency or period of a signal having periodicity contained in the correlation signal to perform the analysis.
[0189] According to this signal processing method, a signal having periodicity is emphasized by correlation filtering. Therefore, it is possible to perform accurate waveform analysis focusing on the period or frequency of a periodic signal.
[0190] In one aspect of the signal processing method, The waveform analysis step may include calculating the frequency by performing a frequency analysis on the correlation signal.
[0191] According to this signal processing method, waveform analysis in the frequency domain can be performed with high accuracy.
[0192] In one aspect of the signal processing method, In the waveform analysis step, the period may be calculated based on time intervals between a plurality of characteristic points of the periodic signal.
[0193] According to this signal processing method, waveform analysis in the time domain can be performed with high accuracy.
[0194] One aspect of the signal processing method includes: The method may further include an abnormality determination step of determining whether or not there is an abnormality in the object based on the analysis result of the waveform analysis step.
[0195] According to this signal processing method, the presence or absence of an abnormality in the target object can be determined with high accuracy based on the results of an accurate waveform analysis of the correlation signal.
[0196] One aspect of the signal processing device includes: a signal-to-be-processed circuit that generates a signal-to-be-processed, which is a time-series signal, based on a source signal that is a time-series signal corresponding to a physical quantity generated in a steady state of an object; a memory circuit for storing a template signal, which is a time-series signal generated based on the source signal; The correlation signal generating circuit performs correlation filtering between the template signal and the processing target signal to generate a correlation signal that is a time-series signal.
[0197] In this signal processing device, the template signal generated based on the source signal and the processing target signal both contain signal components having periodicity correlated with physical quantities occurring in the steady state of the object being processed, and therefore, by correlation filter processing of the template signal and the processing target signal, the signal components having this periodicity are reinforced, making it possible to detect the signal components having periodicity contained in the processing target signal without performing envelope processing.
[0198] Furthermore, among the signal components that are not correlated with the physical quantity, noises having ergodicity are destructive. Therefore, an effect equivalent to synchronously adding the periodic signal components contained in the template signal for the number of periods can be obtained, and a correlation signal with a high S / N ratio can be obtained even if the amplitude of the periodic signal components contained in the processing target signal is small. Therefore, with this signal processing device, restrictions on the installation location of the acquisition unit for the target object and physical quantity can be relaxed.
[0199] Furthermore, according to this signal processing device, since signal processing can be performed when the object is in a steady state, a desired correlation signal can be obtained without changing the motion state of the object. [Explanation of symbols]
[0200] 1 ... signal processing device, 10 ... processing circuit, 11 ... template signal generating circuit, 12 ... processing target signal generating circuit, 13 ... correlation signal generating circuit, 14 ... waveform analysis circuit, 15 ... abnormality determination circuit, 20 ... sensor, 30 ... analog front end, 40 ... memory circuit, 41 ... signal processing program, 42 ... template signal, 43 ... processing target signal, 44 ... correlation signal, 50 ... operation unit, 60: display unit, 70: sound output unit, 80: communication unit, 200: dry pump, 210: booster pump, 211: first pump chamber, 220: main pump, 221: second pump chamber
Claims
1. a process-target signal generating step of generating a process target signal which is a time-series signal based on a source signal which is a time-series signal corresponding to a physical quantity generated in a steady state of the object; a correlation signal generating step of performing correlation filtering between a template signal, which is a time-series signal generated based on the source signal, and the processing target signal to generate a correlation signal, which is a time-series signal; the source signal includes a plurality of signal components having periodicity; The plurality of signal components have different periods, the template signal has a period longer than a period of a first signal component of the plurality of signal components and shorter than a period of the processing target signal; A signal processing method, wherein a difference in length between the signal to be processed and the template signal is longer than a period of the first signal component.
2. a process-target signal generating step of generating a process target signal which is a time-series signal based on a source signal which is a time-series signal corresponding to a physical quantity generated in a steady state of the object; a correlation signal generating step of performing correlation filtering between a template signal, which is a time-series signal generated based on the source signal, and the processing target signal to generate a correlation signal, which is a time-series signal; A signal processing method, wherein at least one of the template signal and the processing target signal is a signal from which a direct current component has been removed.
3. a process-target signal generating step of generating a process target signal which is a time-series signal based on a source signal which is a time-series signal corresponding to a physical quantity generated in a steady state of the object; a correlation signal generating step of performing correlation filtering between a template signal, which is a time-series signal generated based on the source signal, and the processing target signal to generate a correlation signal, which is a time-series signal; The method of signal processing, wherein the template signal is not updated.
4. In claim 1 or 2, The template signal is updated.
5. In any one of claims 1 to 4, A signal processing method, wherein the template signal is a signal obtained by multiplying a signal extracted from the source signal by a window function.
6. In any one of claims 1 to 5, A signal processing method comprising: a template signal generating step of generating the template signal based on the source signal.
7. In any one of claims 1 to 6, A signal processing method, wherein the template signal and the processing target signal have the same sampling period.
8. In any one of claims 1 to 7, the correlation filtering process is a multiply-and-accumulate process of the template signal and the processing target signal, A signal processing method, wherein an envelope of the correlation signal and an envelope of the processing target signal are nonlinear.
9. In claim 8, the product-sum operation is an FIR filter process for the processing target signal, A signal processing method, wherein coefficients of the FIR filter processing are defined based on the template signal.
10. In any one of claims 1 to 9, A signal processing method comprising a waveform analysis step of analyzing a waveform of the correlation signal.
11. In claim 10, A signal processing method, wherein the waveform analysis step performs analysis by focusing on a frequency or a period of a signal having periodicity contained in the correlation signal.
12. In claim 11, A signal processing method, wherein the waveform analysis step calculates the frequency by performing frequency analysis on the correlation signal.
13. In claim 11, A signal processing method, wherein the waveform analysis step calculates the period based on time intervals between a plurality of characteristic points of the periodic signal.
14. In any one of claims 10 to 13, The signal processing method further comprises an abnormality determination step of determining whether or not the object is abnormal based on the analysis result of the waveform analysis step.
15. a signal-to-be-processed circuit that generates a signal-to-be-processed, which is a time-series signal, based on a source signal that is a time-series signal corresponding to a physical quantity generated in a steady state of an object; a memory circuit for storing a template signal, which is a time-series signal generated based on the source signal; A correlation filter process is performed between the template signal and the processing target signal to obtain a time series signal. a correlation signal generating circuit for generating a correlation signal, the source signal includes a plurality of signal components having periodicity; The plurality of signal components have different periods, the template signal has a period longer than a period of a first signal component of the plurality of signal components and shorter than a period of the processing target signal; A signal processing device, wherein a difference in length between the processing target signal and the template signal is longer than a period of the first signal component.
16. a signal-to-be-processed circuit that generates a signal-to-be-processed, which is a time-series signal, based on a source signal that is a time-series signal corresponding to a physical quantity generated in a steady state of an object; a memory circuit for storing a template signal, which is a time-series signal generated based on the source signal; a correlation signal generating circuit that performs correlation filtering between the template signal and the processing target signal to generate a correlation signal that is a time-series signal, The template signal is not updated.
17. a signal-to-be-processed circuit that generates a signal-to-be-processed, which is a time-series signal, based on a source signal that is a time-series signal corresponding to a physical quantity generated in a steady state of an object; a memory circuit for storing a template signal, which is a time-series signal generated based on the source signal; a correlation signal generating circuit that performs correlation filtering between the template signal and the processing target signal to generate a correlation signal that is a time-series signal, At least one of the template signal and the processing target signal is a signal from which a direct current component has been removed.
Citation Information
Patent Citations
Crack detecting apparatus
JP1991191859A
Method of setting sensitivity of ultrasonic flaw detecting device
JP1998062394A
Stationary section determining method and segmentation device for voiced sound
JP1999305788A
Method for precise measurement of waveforms
JP2013506145A
