JP2008298525A - Waveform abnormality detector - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a simple waveform abnormality detector for preventing an erroneous abnormality determination even if a signal waveform to be monitored rapidly rises or falls when there are a time error and a peak value error between a reference signal waveform and the signal waveform to be monitored. <P>SOLUTION: In the waveform abnormality detector, a comparison upper limit and a comparison lower limit are created. An overlap portion of the waveforms forms an upper envelope waveform at the comparison upper limit when larger one of a forward correction value and a backward correction value which are obtained by adding a bias value to a reference value of the previously-created reference signal waveform and forwardly and backwardly moving on a temporal axis is used. An overlap portion of the waveforms forms a lower envelope waveform at the comparison lower limit when smaller one of the forward correction value and the backward correction value which are obtained by subtracting the bias value from the reference value and forwardly and backwardly moving on the temporal axis is used. The abnormality determination is implemented when a monitored value of the signal waveform to be monitored is out of a band between the comparison upper limit and the comparison lower limit. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention compares a monitored value that is a peak value corresponding to a time of a monitored signal waveform generated by a signal source to be inspected with a reference value that is a peak value corresponding to a time of a preset reference signal waveform, The present invention relates to a waveform abnormality detection device that compares and determines at every predetermined time whether there is a difference between the peak values of both waveforms that exceeds an allowable error range.

  Determine whether the output waveform of the sensor that detects the operating state of the target object, such as an electrical component or actuator, matches the expected reference waveform within the allowable error range. Techniques for monitoring are well known. For example, a test apparatus for testing an electrical component having an A / D conversion unit that converts an analog signal into a digital signal, the waveform generation unit generating a test waveform used for testing the electrical component, and applying the test waveform An expected value output unit that outputs an expected value to be output from the electrical component based on the test waveform, an output value of the electrical component to which the test signal is applied, and the expected value are compared. A comparison unit for determining whether the electrical component is good or bad. As a result, there has been disclosed a technique for preventing an electric device under test from being judged as defective based on a test waveform including an error (see, for example, Patent Document 1).

  Further, a simple waveform monitoring device that is incorporated in an injection molding machine and monitors a waveform set based on the pressure of a hydraulic cylinder included in the injection molding machine, the sensor for generating pressure data of the hydraulic cylinder, and the above sensor A reference pressure waveform is set based on the acquired pressure data and an actual measurement waveform is displayed. When the pressure data exceeds a predetermined range with respect to the reference pressure waveform, the actual measurement waveform is marked and displayed. And a discriminator. As a result, before the molded product is molded, the defective product is identified by the mark attached to the measured value waveform, thereby eliminating the appearance inspection performed after the molded product is molded and preventing the outflow of defective products. And the technique of the simple waveform monitoring apparatus which can aim at the improvement of productivity by detecting the malfunction of a molding machine, incidental equipment, and a molding condition during shaping | molding is disclosed (for example, refer patent document 2).

JP 2001-033527 A JP 2004-216663 A

According to Patent Document 1, an expected value output unit that is generated based on a test waveform includes an error included in the test waveform, and as a result, an electrical component under test within an allowable error range is determined to be defective. However, a method of generating an expected value waveform by the expected value output unit is not discussed.
According to Patent Document 2, upper and lower limit waveforms with respect to the reference pressure waveform are set, and whether or not the actually measured waveform is within the upper and lower limit waveform band is displayed as a mark on the actually measured waveform. However, if the monitored signal waveform has a characteristic that increases or decreases sharply, even if there is a slight phase difference (waveform time difference) that is originally allowed, the error of the peak value becomes large, and abnormality determination is made. There are drawbacks to be done.

  An object of the present invention is a case where a time error and a peak value error are included between a reference signal waveform and a monitored signal waveform, and even if the monitored signal waveform suddenly rises or falls, an error occurs. The present invention provides a simple waveform abnormality detection device that does not perform abnormality determination.

  The waveform abnormality detection device according to the present invention compares the time-corresponding peak value of the monitored signal waveform generated by the signal source to be inspected with the preset time-corresponding peak value of the reference signal waveform. A waveform abnormality detection device that compares and determines at every predetermined time whether or not a peak value of a waveform exceeds an allowable error range, wherein the reference value of the reference signal waveform is an estimation theory of a peak value of a reference signal waveform corresponding to time Value or an actual correction value of the standard sample of the signal source to be inspected, and an upward correction value obtained by adding a predetermined first bias value to the reference value is previously determined on the time axis. Using the larger value of the backward correction value obtained by moving the forward correction value obtained by moving only the moving time or the above-mentioned upward correction value on the time axis for a predetermined moving time later, the overlapping part of the waveform is Comparing to form an envelope waveform Band upper limit data generating means for generating a value, and forward correction obtained by moving a downward correction value obtained by subtracting a predetermined second bias value from the reference value previously by a predetermined moving time on the time axis Value or the lower correction value obtained by moving the lower correction value later on the time axis for a predetermined movement time, and using the smaller value of the backward correction value, the comparison lower limit value at which the overlapping portion of the waveform forms the lower envelope waveform The band lower limit data generating means for generating, and the band comparing means for performing abnormality determination when the monitored value of the monitored signal waveform is outside the band between the comparison upper limit value and the comparison lower limit value. .

  In the waveform abnormality detection method according to the present invention, the comparison upper limit value and the comparison lower limit value obtained by correcting the reference waveform back and forth and up and down are used as comparison reference values. Therefore, there is an effect that it is possible to determine whether there is an error in the peak value of the monitored signal waveform and the presence or absence of a time error by simple band comparison means. In particular, the rise and fall of the signal waveform of the monitored signal are steep, and there is an effect that an erroneous comparison and determination is not performed with respect to a difference in peak values generated in a waveform that changes within an allowable time error.

Embodiment 1 FIG.
1 is a circuit block diagram of a waveform abnormality detection apparatus according to Embodiment 1 of the present invention.
As shown in FIG. 1, a waveform abnormality detection apparatus 100A according to the first embodiment of the present invention is mainly composed of a microprocessor 110. The microprocessor 110 includes a program memory 111A in which various control programs are stored, and arithmetic processing. A RAM memory 112 is connected by a bus.
The microprocessor 110 reads out various control programs stored in the program memory 111A, and proceeds with processing according to instructions described in the program.
The operation start command signal DR generated by the microprocessor 110 is supplied to the signal source 120A as a trigger signal via the interface circuit 113.
The inspected signal source 120A generates an analog measurement signal based on this trigger signal, and a digital value is input to the microprocessor 110 as measurement data DATA via the AD converter 114.
The microprocessor 110 supplies an AD conversion command ST to the AD converter 114 at regular intervals, and the AD converter 114 returns an AD conversion completion signal OK to the microprocessor 110. At this point, the microprocessor 110 transfers and writes the measurement data DATA to the RAM memory 112.
Note that the transfer address of the measurement data DATA to the RAM memory 112 is sequentially updated according to the measurement timing, and the measurement value corresponding to each measurement time is stored.
The setting indicator 130A is connected to the microprocessor 110 via a serial interface circuit (not shown), and supplies an operation start command or performs a man-machine interface for displaying an abnormality detection state.

Next, FIG. 2, which is a characteristic diagram showing the reference signal waveform, monitored signal waveform, and reference band waveform in the waveform abnormality detection device according to the first embodiment of the present invention, will be described.
FIG. 2A shows a waveform of a trigger signal generated by the microprocessor 110. FIG. 2B shows a reference signal waveform X that responds to a trigger signal, and the reference corresponding to each time T _i (i is a time number, 0, 1, 2,..., N). A curve connecting the coordinate points (T _i , X _i ) of the value X _i is the reference signal waveform X.
The reference signal waveform X is, for example, an output waveform of a temperature sensor of a ceramic heater driven in response to a trigger signal or an output waveform of a pressure sensor of a hydraulic pump, and a standard sample product was applied. The waveform is the case.
FIG. 2C shows a monitored signal waveform Z when a similar trigger signal is given to the signal source 120A to be inspected. The coordinate point (T _i , Z) of the monitored value Z _i corresponding to each time T _i is shown. _The curve connecting _i ) is the monitored signal waveform Z.
Figure 2 (D) are band lower waveform L and upper limit bandwidth waveform shows a U, the coordinate points (T _{i, L i)} a curve connecting the band lower limit waveform comparison limit value L _i corresponding to each time T _i A curve that connects the coordinate points (T _i , U _i ) of the comparison upper limit value U _i corresponding to each time T _i is the band upper limit waveform U.

Next, FIG. 3 which is a list showing how to generate the band lower limit waveform L and the band upper limit waveform U in the waveform abnormality detection device according to the first embodiment of the present invention and FIG. 4 which is a characteristic diagram will be described. .
In FIG. 3, start time T ₀ (time number i = 0) of time T _i corresponds to, for example, the rise time of the trigger signal in FIG. 2A, and end time T _n (time number i = n). Is the end time of waveform monitoring, and the waveform monitoring period T _n −T ₀ and the measurement interval (T _n −T ₀ ) / n between the monitoring periods have predetermined values.
The peak value of the reference signal waveform X corresponding to each time T _i is shown as a reference value X _i .
The upward correction value A0 _i and the downward correction value B0 _{i that} are the crest values at each time of the upward movement waveform A0 and the downward movement waveform B0 in FIG. 4A are expressed by the equations (1) and (1), as shown in the table of FIG. Calculated by (2).

A0 _i = X _i + α _i where α _i = α + Δα × X _i (1)
B0 _i = X _i -β _i where β _i = β + Δβ × X _i (2)

That is, the addition bias value α _i is composed of a fixed bias value α and a proportional bias value Δα × X _i , and the proportionality coefficient is indicated by Δα.

The subtraction bias value β _i is composed of a fixed bias value β and a proportional bias value Δβ × X _i , and the proportional coefficient is indicated by Δβ.
As this bias value, only a fixed bias value is used, only a proportional bias value is used, or a combination of both is used. Usually, an addition bias value α _i and a subtraction bias value β _i are The same value is used.
A forward movement waveform A1 and a backward movement waveform A2 in FIG. 4B are waveforms obtained by moving the upward movement waveform A0 in FIG. 4A before and after the time axis, respectively, and the forward movement waveform A1 at time T _i and forward correction value A1 _i and rear correction value A2 _i is a peak value of the backward movement waveform A2 is being used above correction value A0 _i that as shown in Table such as, for example, two of the measurement interval has moved back and forth in FIG. 3 Yes.
Figure 4 (C) forward motion waveform B1 and rearward movement waveform B2 in is a waveform that has moved respectively before and after the time-axis downward movement waveform B0 in FIG. 4 (A), the forward movement waveform B1 at time T _i As the front correction value B1 _i and the rear correction value B2 _i , which are the peak values of the backward movement waveform B2, as shown in the table of FIG. 3, for example, the upward correction value B0 _{i that} is moved back and forth between two measurement intervals is used. Yes.

The comparison upper limit value U _i of the band upper limit waveform U shown in FIG. 2D and the comparison lower limit value L _i of the band lower limit waveform L are shown by the equations (3) and (4).

U _i = max (A1 _i , A0 _i , A2 _i ) (3)
L _i = min (B1 _i , B0 _i , B2 _i ) (4)

That is, the comparison upper limit value U _i corresponds to the largest value among the upper correction value A0 _{i at} time T _i , the front correction value A1 _i of the upper correction value A0 _i , and the rear correction value A2 _i .
Further, corresponding to the smallest value of among the comparison limit value _{L i} forward correction value of the time the lower correction value in _{T i} B0 _i and its lower correction value B0 _i B1 _i and rear correction value B2 _i.
Ideally, a large number of movement waveforms with a slight time difference between the forward correction value A1 _i and the rear correction value A2 _i are assumed, and the maximum values of these many movement waveforms are used. However, in the case of a simple waveform shape, the larger one of the two waveforms of the front correction value A1 _i and the rear correction value A2 _i can be used.
Similarly, for the comparison lower limit value L _i , a large number of movement waveforms with a slight time difference between the forward correction value B1 _i and the rear correction value B2 _i are assumed, and the minimum values of these many movement waveforms are set. Although it is ideal to use, it is also possible to use the smaller one of the two waveforms of the forward correction value B1 _i and the backward correction value B2 _i if the waveform has a simple shape.

Next, FIG. 5 which is a characteristic diagram showing a reference band waveform and a monitored signal waveform related to the waveform abnormality detection device according to the first embodiment of the present invention, and waveform abnormality detection according to the first embodiment of the present invention. FIG. 6, which is a list for explaining comparative determination related to the apparatus, will be described.
In FIG. 5, it is determined at each time whether or not the monitored signal waveform Z has a peak value between the band upper limit waveform U and the band lower limit waveform L that are the reference band waveform.
6, the monitoring value Z _i is made excessively small determination is less than comparison limit value L _i, the monitored value Z _i is compared lower limit L _i above, proper determination is made equal to or less than the comparison the upper limit U _i, monitoring If the value Z _i exceeds the comparison upper limit value U _i , an overdetermination is made.
For example, if the underdetermination is performed once at any time, the underdetermination flag is set and the undernotification is performed.
Similarly, if the over-determination is performed once at any time, an over-determination flag is set and over-notification is performed.
Alternatively, the number of times of underdetermination and the number of times of overdetermination can be counted by a counter, respectively, and when this count value exceeds a predetermined threshold value, an underestimation or overestimation or abnormality notification can be performed.

Next, FIG. 7 which is an operation flowchart of the reference signal waveform measurement in the waveform abnormality detection device according to the first embodiment of the present invention will be described.
In FIG. 7, a process 700 is a step in which the microprocessor 110 starts a reference signal waveform measurement operation based on an operation command from the setting indicator 130 </ b> A, and a time counter (not shown) as the process 700 is executed. Will start timing.
Subsequent step 701 is a standby time determination step for waiting for a predetermined time until the trigger signal generation time shown in FIG. 2A. If the predetermined time has not elapsed, NO is determined and step 701 is performed. If a predetermined time has elapsed, a determination of YES is made and the process proceeds to step 702.
Step 702 the time by the time counter is equal to or a generation time period of the trigger signal, the process proceeds to step 703a performs determination of YES if Itare the time T ₀ to generating a trigger signal, the trigger This is a time zone determination step in which NO is determined and the process proceeds to step 703b in the period before the generation of the signal and at the end of the generation.

In step 703a, the microprocessor 110 generates a trigger command signal DR and supplies the trigger signal to the signal source 120A to be inspected which is a standard sample product via the interface circuit 113. The signal source 120A to be inspected is shown in FIG. ) Is generated.
In step 703b, when the trigger command signal DR is not generated or the trigger command signal DR that has already been generated is stopped, the signal source 120A to be inspected, which is a standard sample product, is shown in FIG. As described above, the generation of the reference signal waveform X is stopped.
Step 704, which is executed subsequent to step 703a or step 703b, is a determination step of a measurement time interval for waiting for the AD converter 114 to arrive at the generation timing of the AD conversion command ST. If the measurement time arrives, YES is determined. A determination is made and the process proceeds to step 705. If it has not arrived, a NO determination is made and the process returns to step 704.
Step 705 is a step in which the microprocessor 110 generates an AD conversion command ST to the AD converter 114, and a subsequent step 706 determines whether or not the AD conversion completion signal OK generated by the AD converter 114 has been received. If not received, NO is determined and the process returns to Step 706. If received, YES is determined and the process proceeds to Step 707a.

Step 707a is a step of accessing the address ADD _i of the RAM memory 112 corresponding to the measurement time T _i , and the first time is T ₀ .
Subsequent step 707b, the step of writing to the AD converter 114 digital conversion input data DATA to the address accessed by the process 707a as the reference value _{X i} RAM memory 112 by, subsequent step 707c adds 1 to the time number i In step 708, it is determined whether or not it is the measurement completion time. If it is not completed, NO is determined and the process returns to step 701. If completed, YES is determined and operation end process 709 is performed. This is a determination step for shifting to.
Note that the operation required time of a series of processes from the process 701 to the process 708 is sufficiently shorter than the measurement time interval, and the measurement timing at each time is adjusted by the standby time in the process 704 and measured at regular intervals. Is to be done.
The determination of the measurement completion time in step 708 is the operation completion after a predetermined time from the generation stop of the trigger signal in step 702, and the generation period of the trigger signal is a predetermined value set in advance.

Step 707a, 707b, step block 710 constituted by 707c is a reference data writing means, a data table of measured reference value _{X i} trigger signal according to step 703a is corresponds to a time _{T i} from occurring Is generated and stored in the RAM memory 112.
The data table of time T _i vs. reference value X _i is measured by another offline device (not shown), and the measurement result is transferred and written from the setting display 130A to the RAM memory 112.
In addition, a theoretical value that does not depend on actual measurement values can be applied to the data table of time T _i versus reference value X _i .

Next, FIG. 8 which is an operation flowchart of the reference band waveform generation in the waveform abnormality detection device according to the first embodiment of the present invention will be described.
In FIG. 8, step 800 is an operation start step of the microprocessor 110 for calculating and generating the comparison upper limit value U _i and the comparison lower limit value L _i that form the band lower limit waveform L and the band upper limit waveform U of FIG. The subsequent step 801 is a step of setting i = 0 as the data number, the subsequent step 802 is a step of reading the reference value X _i from the address ADD _i of the RAM memory 112, and the subsequent step 803a is an upward correction by the equation (1). calculating a value A0 _i, followed by step 803b is a step of extracting a forward correction value A1 _i values of upper correction value _{A0 i + k} at time _{T i + k} at time _{T i,} the front correction time difference δ _{= T i + k} It has become a -T _i.
Subsequent step 803c is a step of extracting the value of the upper correction value _{A0 i-k} at time _{T i-k} as the rear correction value A2 _i at time _{T i,} the backward correction time difference δ ₌ T _{i -T i- k} .
The subsequent step 803d is a step of calculating the maximum value among the upper correction value A0 _i , the front correction value A1 _i and the rear correction value A2 _{i as} the comparison upper limit value U _i at the time T _i , and from step 803a to step 803d. The configured process block 810 serves as band upper limit data generation means.

Step subsequent step 804a, the extracting step of calculating a lower correction value B0 _i by equation (2), followed by step 804b, the value of the lower compensation value _{B0 i + k} at time _{T i + k} as the forward correction value B1 _i at time _{T i} The forward correction time difference is δ = T _{i + k} −T _i .
Subsequent step 804c is a step of extracting the value of the lower compensation value _{B0 i-k} at time _{T i-k} as the rear correction value B2 _i at time _{T i,} the backward correction time difference δ ₌ T _{i -T i- k} .
Subsequent step 804d is lower correction value B0 _i, a step of calculating a comparison limit value _{L i} at time _{T i} the minimum value of the forward correction value B1 _i and rear correction value B2 _i, by the step 804a Step 804d The configured process block 820 serves as band lower limit data generation means.
A subsequent step 805 adds 1 to the time number i, and a subsequent step 806 determines whether or not the time number i has exceeded the final number n. If not, NO is determined and step 802 is performed. If YES, the determination of YES is made, and the operation end process 807 is entered.

As described above, FIG. 8 shows the band upper limit data generating means 810 and the band lower limit data stored in the program memory 111A based on the data table of the time T _i versus the reference value X _i stored in the RAM memory 112 in FIG. A data table for the comparison upper limit value U _i and the comparison lower limit value L _i is generated by a control program comprising the generation means 820, and is written and stored in a predetermined address area of the RAM memory 112.
However, if the data table of the comparison upper limit value U _i and the comparison lower limit value L _i is generated by another offline device and it is not necessary to change the data table, the comparison display upper limit value U _i is set from the setting display 130A. and if transfer write data table compares the lower limit value L _i in the RAM memory 112, control program as the program memory 111A made of band limit data generating means 810 and bandwidth limit data generating means 820 Metropolitan is unnecessary, the time T _i A data table for the reference value X _i is also unnecessary.
However, if the user wants to generate a data table of the comparison upper limit value U _i and the comparison lower limit value L _{i on} the basis of the data table of the time T _i versus the reference value X _i , the program memory 111A or the setting indicator 130A It is necessary to provide a control program comprising band upper limit data generating means 810 and band lower limit data generating means 820 in the inside.
In this case, in order to prevent the comparison determination condition from being inadvertently changed, a measure for enabling a setting change using a specific password is performed.

Next, FIG. 9 which is an operation flowchart of the monitored signal waveform measurement related to the waveform abnormality detection device according to the first embodiment of the present invention will be described.
FIG. 9 shows a case where a signal source to be monitored is connected as the signal source 120A to be inspected in FIG. 1, and performs the same operation as FIG. Has been changed to 900.
However, the data table generated by the measurement data writing means 910 in FIG. 9 is the monitored value Z _i of the monitored signal waveform Z at time T _i , and this data table is the data table of the reference value X _i. It is stored in the RAM memory 112 in a different address area.
Next, FIG. 10, which is an operation flowchart of the comparison determination related to the waveform abnormality detection device according to the first embodiment of the present invention, will be described.
In FIG. 10, a process 1000 is a step in which the microprocessor 110 starts a comparison determination operation, a subsequent process 1001 is a step of setting the time number i to zero, and a subsequent process 1002 a is a monitoring value stored in the RAM memory 112. The step of reading Z _i , the subsequent step 1002b is the step of reading the comparison upper limit value U _i and the comparison lower limit value L _i stored in the RAM memory 112, and the subsequent step 1003a is the monitoring value Z read in step 1002a. _i is equal to or comparative less than the lower limit L _i read in step 1002b, than if the process proceeds to step 1004a performs determination is YES, the performed determination of NO be less than This is a determination step for shifting to step 1003b.

In step 1003b, it is determined whether or not the monitoring value Z _i read in step 1002a exceeds the comparison upper limit value U _i read in step 1002b. The process shifts to 1004b, and if it does not exceed, it is a determination step of determining NO and proceeding to step 1004c.
Step 1004a is a step of setting an under-determination flag FL _i , step l004b is a step of setting an over-determination flag FU _i , and step 1004c is a step of setting an appropriate flag FJ _i. Following the steps 1004a, 1004b, and 1004c, steps In step 1005, 1 is added to the time number i.
In the subsequent step 1006, it is determined whether or not the time number i has exceeded the final number n. If not, the determination is NO and the process returns to the step 1002a. This is a determination step for shifting to 1007.
In step 1007, after the comprehensive determination output is performed, the operation is shifted to the operation end step 1008.

A process block 1010 composed of the processes 1003a and 1003b serves as a band comparison unit.
Further, as the overall determination output in step 1007, for example, in step 1004a or step 1004b, if any of the under determination flag FL _i or the over determination flag FU _i is set, an abnormality notification output is generated, Whether or not the underdetermination flag is set can be displayed on the screen by operating the setting indicator 130A.
In addition, the section of the monitored signal waveform Z is divided into, for example, a first period of rising, a second period in the middle, and a third period of falling, and an underdetermination flag FL _i and an overdetermination flag FU _i are set in each period. It is also possible to display whether or not it has been set.
It is also possible to provide a counter that counts the number of underdeterminations and the number of overdeterminations in each section, and displays the number of occurrences of abnormality for each section.

In the above description, the comparison determination according to FIG. 10 is performed after the writing of the measurement data according to FIG. 9 is completed. However, the measurement time interval is a time interval having a margin compared to the arithmetic processing capability of the microprocessor 110. in some cases, after performing the measurement at time T _i, the until the next measurement can be performed to complete the online determination of the comparison determination of the monitoring value Z _i at the time T _i.

As is apparent from the above description, the waveform abnormality detection device according to the first embodiment of the present invention has a monitoring value Z _i that is a peak value corresponding to the time of the monitored signal waveform Z generated by the inspected signal source 120A, and By comparing the reference value X _i that is the peak value corresponding to the time of the set reference signal waveform X, it is compared and determined at every predetermined time whether there is a difference between the peak values of both waveforms exceeding the allowable error range. In the waveform abnormality detection device, the upper and lower limit values of the allowable error range are generated by the band upper limit data generation unit 810 and the band lower limit data generation unit 820, and the comparison determination is executed by the band comparison unit 1010. Yes.
The reference value X _i of the reference signal waveform X is created based on the estimated theoretical value of the peak value of the reference waveform corresponding to the passage of time or the measured value of the standard sample.
Upper limit bandwidth data generating means 810, an upper correction value A0 _i obtained by adding the reference value X first bias value predetermined to _i alpha _i, respectively moved back and forth above the correction value A0 _i time axis Of the obtained forward correction value A1 _i and backward correction value A2 _i , at least the larger one of the forward correction value A1 _i and the backward correction value A2 _i is used, and the overlapping portion of the waveform represents the upper envelope waveform. A comparison upper limit value U _i to be formed is generated.
Band lower data generation unit 820, and a lower correction value B0 _i obtained by subtracting the reference value X _i from the predetermined second bias value beta _i, respectively moved back and forth beneath the correction value B0 _i time axis Of the obtained forward correction value B1 _i and backward correction value B2 _i , at least the smaller one of the forward correction value B1 _i and the backward correction value B2 _i is used, and the overlapping portion of the waveform forms the lower envelope waveform. A comparison lower limit L _i is generated.
The band comparison means 1010 makes an abnormality determination when the monitored value Z _i of the monitored signal waveform Z is outside the band between the comparison upper limit value U _i and the comparison lower limit value L _i .

The first bias value alpha _i or a second bias value beta _i is subtracted from a lower correction value B0 _i, to be added to the above correction value A0 _i is the common constant value at each time of the reference signal waveform X Or a value proportional to the reference value X _i at each time of the reference signal waveform X.
Thus, the bias value for the upper or lower correction value is fixed or a value proportional to the reference value is used. Accordingly, the allowable error range can be accurately set according to whether the error component of the monitored signal waveform is a drift component due to a temperature error or the like, or due to fluctuations in variation of the amplification gain.

As the comparison upper limit value U _i in the band upper limit data generation means 810, the maximum value of three numerical values of the upper correction value A0 _i , the forward correction value A1 _i related to the upper correction value, and the rear correction value A2 _i is applied. As the comparison lower limit value L _i in the generation means 820, the minimum value of the three numerical values of the lower correction value B0 _i , the forward correction value B1 _i and the rear correction value B2 _i relating to the lower correction value is applied. Therefore, the comparison upper limit value and the comparison lower limit value can be obtained by a simple algebra calculation.

The monitored value Z _i of the monitored signal waveform Z generated by the signal source 120A to be inspected is stored in the RAM memory 112 via the AD converter 114 and the microprocessor 110, and cooperates with the microprocessor 110. The program memory 111 </ b> A includes a control program that becomes the band comparison unit 1010.
The program memory 111A or the RAM memory 112 stores the time-related comparison upper limit value U _i and the comparison lower limit value L _i generated based on at least the band upper limit data generation unit 810 and the band lower limit data generation unit 820. ing.
As described above, a microprocessor for detecting an abnormality of the monitored signal waveform is provided in cooperation with the program memory or the RAM memory in which the comparison upper limit value and the comparison lower limit value are stored. Therefore, since the program memory only needs to perform the abnormality determination by the band comparison means, a small and inexpensive waveform abnormality detection device can be obtained.

The program memory 111 </ b> A further includes a control program serving as a band upper limit data generation unit 810 and a band lower limit data generation unit 820.
As described above, the waveform abnormality detection apparatus includes the band upper limit data generation unit and the band lower limit data generation unit. Therefore, the comparison upper limit value and the comparison lower limit value corresponding to the signal source to be inspected having various monitored signal waveforms can be generated by itself, and abnormality determination for various signal waveforms can be performed.

The microprocessor 110 is connected to a setting indicator 130A serving as a man-machine interface.
The setting indicator 130A notifies the abnormal state of the monitored signal waveform, and setting change means for the values of the comparison upper limit value U _i and the comparison lower limit value L _i stored in the program memory 111A or the RAM memory 112, Alternatively, any setting changing means for changing the reference value X _i with respect to the reference signal waveform X, or setting changing means for the upper correction bias value α _i , the lower correction bias value β _i, and the time width setting for forward and backward movement correction I have.
In this way, the setting constant can be changed by the setting indicator 130A. Therefore, the user using the waveform abnormality detection device can change the determination level of abnormality detection, and transfer and write the reference value of the reference signal waveform measured off-line as information data.

Embodiment 2. FIG.
Hereinafter, FIG. 11 which is a circuit block diagram of the waveform abnormality detection device according to the second embodiment of the present invention will be described focusing on the differences from FIG.
In each figure, the same numerals indicate the same or corresponding parts.
As shown in FIG. 11, a waveform abnormality detection apparatus 100B according to the second embodiment of the present invention is mainly composed of a microprocessor 110. The microprocessor 110 includes a program memory 111B in which various control programs are stored and an arithmetic process. A RAM memory 112 is connected by a bus.
A measurement command signal generated by the signal source 120B to be inspected is supplied to the microprocessor 110 as a trigger signal input TR via the interface circuit 115.
In response to the generation of the measurement command signal, the signal source 120B to be inspected generates an analog measurement signal, and a digital value is input to the microprocessor 110 as measurement data DATA via the AD converter 114. .
The microprocessor 110 supplies an AD conversion command ST to the AD converter 114 at regular intervals, and the AD converter 114 returns an AD conversion completion signal OK to the microprocessor 110. At this time, the microprocessor 110 performs measurement. Data DATA is transferred and written to the RAM memory 112.
However, the transfer address of the measurement data DATA to the RAM memory 112 is sequentially updated according to the measurement timing, and the measurement value corresponding to each measurement time is stored.
The setting indicator 130B is connected to the microprocessor 110 via a serial interface circuit (not shown), and performs a man-machine interface for displaying an abnormality detection state.

Next, FIG. 12 which is a characteristic diagram showing a reference signal waveform, a monitored signal waveform and a reference band waveform in the waveform abnormality detection apparatus according to the second embodiment of the present invention will be described.
FIG. 12A shows a waveform of a measurement command signal generated by the signal source to be inspected 120B.
FIG. 12B shows a reference signal waveform X that responds to a measurement command signal, and a curve connecting the coordinate points (T _i , X _i ) of the reference value X _i corresponding to each time T _i is a reference signal. A waveform X is obtained.
Note that FIG. 12B shows a plurality of reference signal waveforms X when a plurality of measurements are performed for a plurality of standard sample products, and the waveform change rate is less than a minute predetermined value close to zero. The vertical error width, which is the difference between the maximum value and the minimum value in the flat period, is indicated by ΔX.
Further, the front-to-back error width, which is the maximum time difference between the waveform portions showing steep waveform changes, is indicated by ΔT, and the front-to-back error width ΔT is the time when the upper limit signal waveform Xu has a steep waveform change greater than or equal to a predetermined value. There is a time difference between the first time and the second time, which is the time when the lower limit signal waveform Xl has undergone a steep waveform change of a predetermined value or more.
The steep waveform change referred to here is, for example, a value in which the amount of change in the peak value at the previous time and the current time exceeds the addition bias value or the subtraction bias value.
In addition, the vertical error width ΔX and the vertical error width ΔT may be determined by visual determination by drawing a large number of reference signal waveforms X on the same graph. It is necessary to grasp the characteristics related to the shape of and to formulate a formula suitable for this.

FIG. 12C shows a monitored signal waveform Z in the case of an inspected product, and a curve connecting the coordinate points (T _i , Z _i ) of the monitored value Z _i corresponding to each time T _i is shown. A monitoring signal waveform Z is obtained.
FIG. 12D shows the band lower limit waveform L and the band upper limit waveform U, and a curve connecting the coordinate points (T _i , L _i ) of the comparison lower limit value L _i corresponding to each time T _i is a band lower limit waveform. The waveform L is a curve that connects the coordinate points (T _i , U _i ) of the comparison upper limit value U _i corresponding to each time T _i, and is the band upper limit waveform U.
In FIG. 12D, the upper limit signal waveform Xu is a waveform obtained by connecting the reference upper limit value Xu _i which is the largest value among the plurality of reference signal waveforms X.
The lower limit signal waveform Xl is a waveform obtained by connecting the reference lower limit value Xl _i is the smallest value among a plurality of reference signal waveform X.
In the second embodiment, the upward correction value A0 _i and the downward correction value B0 _i , which are the peak values at the respective times of the upward movement waveform A0 and the downward movement waveform B0 (not shown), are calculated by Expressions (5) and (6). It is supposed to be.

A0 _i = Xu _i + α _i where α _i = γ × ΔX (5)
B0 _i = Xl _i -β _i where β _i = γ × ΔX (6)

That is, the addition bias value α _i and the subtraction bias value β _i are values proportional to the vertical error width ΔX, and the proportionality coefficient is indicated by γ.

A band upper limit waveform U in FIG. 12D is an upper envelope waveform using a forward movement waveform A1 and a backward movement waveform A2 obtained by moving the upward movement waveform A0 around the time axis.
The band lower limit waveform L in FIG. 12D is a lower envelope waveform using a forward movement waveform B1 and a backward movement waveform B2 obtained by moving the downward movement waveform B0 around the time axis.
However, in actuality, the comparison upper limit value U _i of the band upper limit waveform U and the comparison lower limit value L _i of the band lower limit waveform L are expressed by the equations (7) and (8).

U _i = max (Al _i , A0 _i , A2 _i ) (7)
_{L i = min (Bl i,} B0 i, B2 i) (8)

That corresponds to the largest value of among the comparison the upper limit _{U i} at time _T the forward correction value with the upper correction value A0 _i in _i Al _i and rear correction value A2 _i.
The comparison lower limit value L _i corresponds to the smallest value among the downward correction value B0 _i , the forward correction value Bli _i, and the backward correction value B2 _i at time T _i .

The longitudinal movement time width for the upper limit signal waveform Xu and the lower limit signal waveform Xl is a value proportional to the longitudinal error width ΔT, and the total value of the longitudinal movement time width is larger than the longitudinal error width ΔT. .
Also, ideally, a large number of movement waveforms with a slight time difference between the front correction value A1 _i and the rear correction value A2 _i are assumed, and the maximum values of these many movement waveforms are used. However, in the case of a simple waveform shape, the larger one of the two waveforms of the front correction value A1 _i and the rear correction value A2 _i can be used.
Similarly, for the comparison lower limit value L _i , a large number of movement waveforms with a slight time difference between the forward correction value B1 _i and the rear correction value B2 _i are assumed, and the minimum values of these many movement waveforms are set. Although it is ideal to use, it is also possible to use the smaller one of the two waveforms of the forward correction value B1 _i and the backward correction value B2 _i if the waveform has a simple shape.

Next, FIG. 13 which is an operation flowchart of the reference signal waveform measurement related to the waveform abnormality detection device according to the second embodiment of the present invention will be described.
In FIG. 13, step 1300 is a step in which the microprocessor 110 starts the measurement operation of the reference signal waveform based on the operation command from the setting indicator 130B.
A subsequent step 1301 is a step of updating the setting of the measurement data number N for a plurality of measurements regarding a plurality of standard samples.
In subsequent step 1302, it is determined whether or not the measurement command signal shown in FIG. 12A has been received. If not, NO is determined and the process returns to step 1302, and if received, YES is determined. This is a determination step for shifting to 1303.
Step 1303 is a process block for creating a data table of time T _i vs. reference value X _i . First, data of measurement data number N = 0 is created.
In subsequent step 1304, it is determined whether or not the measurement command signal shown in FIG. 12A has been canceled. If not, NO is determined and the process returns to step 1303. This is a determination step to go to step 1305.

Step 1305 is a step of adding 1 to the measurement data number K, and the subsequent step 1306 determines whether or not all data measurement is completed. If it is not completed, NO is determined and the process returns to Step 1301. If there is, it is a determination step of determining YES and proceeding to step 1307a.
Step 1307a extracts the reference upper limit value Xu _i corresponding to the measurement time T _i and stores it in the RAM memory 112, and the subsequent step 1307b extracts the reference lower limit value Xl _i corresponding to the measurement time T _i and stores it in the RAM. The step of storing in the memory 112, the subsequent step 1307c calculates and stores the vertical error width ΔX, the subsequent step 1307d calculates and stores the front-back error width ΔT, and the subsequent step 1308 is the operation end step.
Incidentally, the upper and lower error width ΔX is calculated as the difference value of the reference upper limit value Xu _i and the reference lower limit value Xl _i in the waveform fluctuation slow time of the upper limit signal waveform Xu and the lower limit signal waveform Xl, error width ΔT longitudinal upper limit signal waveform This is calculated as a time difference between the first time when Xu has a sharp waveform change and the second time when the lower limit signal waveform Xl has a sharp waveform change.

The process block 1310 constituted by the processes 1307a to 1307d serves as a reference width data writing means, and the time during which the measurement command signal is canceled in the process 1304 after the measurement command signal is generated in the process 1302 is generated. Corresponding to T _i , a data table of the measured reference value X _i is generated and stored in the RAM memory 112. From the plurality of reference value data, the reference upper limit value Xu _i , the reference lower limit value Xl _i , and the vertical error width ΔX The front and rear error width ΔT is extracted. The front-rear error width ΔT is a first time that is a time when the upper limit signal waveform Xu has a steep waveform change of a predetermined value or more, and a second time that is the time of the lower limit signal waveform Xl having a steep waveform change of a predetermined value or more. The time is different from the time.
The data table of time T _i vs. reference value X _i is measured by another offline device (not shown), and the reference upper limit value Xu _i , reference lower limit value Xl _i , vertical error width ΔX, front and rear error width as measurement results ΔT may be transferred and written from the setting display 130B to the RAM memory 112.
Further, theoretical values that do not depend on the actual measurement values of the standard sample product can be applied to the reference upper limit value Xu _i , the reference lower limit value Xl _i , the vertical error width ΔX, and the front and rear error width ΔT.

Next, FIG. 14 which is an operation flowchart of the reference band waveform generation related to the waveform abnormality detection device according to the second embodiment of the present invention will be described.
In FIG. 14, a process 1400 is an operation start step of the microprocessor 110 for calculating and generating the comparison upper limit value U _i and the comparison lower limit value L _i that form the reference band waveform U · L of FIG. Is the step of first setting i = 0 as the data number, the subsequent step 1402 is a step of reading the reference upper limit value Xu _i and the reference lower limit value Xl _i from the RAM memory 112, and the subsequent step 1403a is the upper correction value A0 _i according to equation (5). calculating a, a step of extracting a forward correction value Al _i values of upper correction value _{A0 i + k} at time _{T i} of a step 1403b is a time _{T i + k,} the forward corrected time difference and δ _{= T i} + k -T _i It has become.
Subsequent step 1403c is a step of extracting the value of the upper correction value _{A0 i-k} at time _{T i-k} as the rear correction value A2 _i at time _{T i,} the backward correction time difference δ ₌ T _{i -T i-k} It has become.
Subsequent step 1403d is a step of calculating the maximum value among the above correction value A0 _i and forward correction value Al _i and rear correction value A2 _i as compared upper limit _{U i} at time _{T i,} constituted by the steps 1403d from Step 1403a The processed process block 1410 serves as a band upper limit data generating means.

Be a step following step 1404a to be extracted as a forward correction value Bl _i in step, subsequent step 1404b is a time _{T i +} lower correction value in _{k B0 i + k} value the time _{T i} of calculating a lower correction value B0 _i by equation (6) The forward correction time difference is δ = T _{i + k} −T _i .
Subsequent step 1404c is a step of extracting the value of the lower compensation value _{B0 i-k} at time _{T i-k} as the rear correction value B2 _i at time _{T i,} the backward correction time difference δ ₌ T _{i -T i-k} It has become.
Subsequent step 1404d is a step of calculating the minimum value of the downward correction value B0 _i and forward correction value Bl _i and rear correction value B2 _i as compared lower limit _{L i} at time _{T i,} constituted by the steps 1404d from Step 1404a The processed block 1420 serves as a band lower limit data generating means.
A subsequent step 1405 is a step of adding 1 to the time number i, and a subsequent step 1406 determines whether or not the time number i has exceeded the final number n. If not, the determination is NO and the process returns to step 1402. However, if it exceeds, a determination of YES is made and the process proceeds to the operation end step 1407.

As described above, FIG. 14 shows the bandwidth upper limit data stored in the program memory 111B based on the data table of the time T _i vs. the reference upper limit value Xu _i and the reference lower limit value Xli _i stored in the RAM memory 112 in FIG. A data table of the comparison upper limit value U _i and the comparison lower limit value L _i is generated by a control program serving as the generation unit 1410 and the band lower limit data generation unit 1420, and is written and stored in a predetermined address area of the RAM memory 112.
However, if the data table of the comparison upper limit value U _i and the comparison lower limit value L _i is generated by another offline device and there is no need to change this data table, the setting display unit 130B sets the comparison upper limit value U _i . If the data table of the comparison lower limit value L _i is transferred and written to the RAM memory 112, the program memory 111B does not require the control program to be the band upper limit data generation unit 1410 and the band lower limit data generation unit 1420, and the time T _i vs. reference The data table of the upper limit value Xu _i and the reference lower limit value Xl _i is also unnecessary.
However, if the user wants to generate a data table of the comparison upper limit value U _i and the comparison lower limit value L _{i on} the basis of the data table of the time T _i versus the reference upper limit value Xu _i and the reference lower limit value Xl _i , the program It is necessary to provide a control program to be the band upper limit data generation unit 1410 and the band lower limit data generation unit 1420 in the memory 111B or the setting display 130B.
In this case, in order to prevent the comparison determination condition from being inadvertently changed, a measure for enabling a setting change using a specific password is performed.

Next, FIG. 15 which is an operation flowchart of the monitored signal waveform measurement related to the waveform abnormality detection device according to the second embodiment of the present invention will be described.
In FIG. 15, step 1500 is a step in which the microprocessor 110 starts the measurement operation of the reference signal waveform based on the operation command from the setting indicator 130B, and a time counter (not shown) is executed as the step 1500 is executed. The timekeeping operation starts.
In subsequent step 1501, it is determined whether the measurement command signal shown in FIG. 12A has been received. If not, NO is determined and the process returns to step 1501. If received, YES is determined. This is a determination step for shifting to 1504.
Step 1502 is a measurement time interval determination step for waiting for the AD converter 114 to arrive at the generation timing of the AD conversion command ST. If the measurement time arrives, a determination of YES is made and the process proceeds to step 1503. If it is, NO is determined and the process returns to step 1502.
Step 1503 is a step in which the microprocessor 110 generates an AD conversion command ST to the AD converter 114, and a subsequent step 1504 determines whether or not the AD conversion completion signal OK generated by the AD converter 114 has been received. If received, NO is determined and the process returns to step 1504. If received, YES is determined and the process proceeds to step 1505a.

Step 1505a is a step of accessing a predetermined address of the RAM memory 112 in response to the measured time _{T i,} the first time has become _{T 0.}
The subsequent step 1505b is a step of writing the digital conversion input data DATA by the AD converter 114 as the monitoring value Z _i into the RAM memory 112 at the address accessed in step 1505a, and the subsequent step 1505c is a step of adding 1 to the time number i. In subsequent step 1506, it is determined whether or not the measurement command signal shown in FIG. 12A has been canceled. If not, NO is determined and the process returns to step 1502, and if it is canceled, YES is determined. This is a determination step that is performed to move to the operation end step 1507.
Note that the operation required time of a series of steps from step 1502 to step 1506 is sufficiently shorter than the measurement time interval, and the measurement timing at each time is adjusted by the standby time in step 1502 and is measured at regular intervals. Is to be done.
Step 1505a, 1505b, process block 1510 constituted by 1505c is to be a measured data writing means, the monitoring value Z measurement instruction signal according to step 1501 is measured corresponding to the time _{T i} from occurring A data table of _i is generated and stored in the RAM memory 112.

Next, FIG. 16 which is an operation flowchart of the comparison determination relating to the waveform abnormality detection device according to the second embodiment of the present invention will be described.
In FIG. 16, a process 1600 is a step in which the microprocessor 110 starts a comparison determination operation, a subsequent process 1601 is a step in which the time number i is set to zero, and a subsequent process 1602 a is a monitor value Z _i stored in the RAM memory 112. The step of reading, the subsequent step 1602b is the step of reading the comparison upper limit value U _i and the comparison lower limit value L _i stored in the RAM memory 112, and the subsequent step 1603a is the step 1602b where the monitored value Z _i read in step 1602a is the step 1602b. it is determined whether the below read compared lower limit L _i, the process proceeds to step 1604a performs determination of YES is less than, the flow proceeds to step 1603b performs determination of NO be less than This is a judgment step.
In step 1603b, it is determined whether or not the monitoring value Z _i read in step 1602a exceeds the comparison upper limit value U _i read in step 1602b. If it exceeds, a determination of YES is made and step 1604b is performed. If it does not exceed, it is determined that NO is determined and the process proceeds to step 1604c.

Step 1604a is a step of counting the number of underdeterminations by an unillustrated undercounter, step 1604b is a step of counting the number of overdeterminations by an unillustrated overcounter, and step 1604c is a step of counting the number of proper determinations by an appropriate counter (not shown) Subsequent to Steps 1604a, 1604b, and 1604c, the process goes to Step 1605 to add 1 to the time number i.
In subsequent step 1606, it is determined whether or not the time number i has exceeded the final number n. If not, NO is determined and the process returns to step 1602a. If it is exceeded, YES is determined and step 1607 is determined. It is a judgment step to shift to.
In step 1607, after the comprehensive determination output is performed, the process proceeds to the operation end step 1608.
A process block 1610 composed of processes 1603a and 1603b serves as a band comparison unit.

In the above description, after the writing of the measurement data according to FIG. 15 is completed, the comparison determination according to FIG. 16 is performed. However, the measurement time interval is a time interval having a margin compared to the arithmetic processing capability of the microprocessor 110. in some cases, after performing the measurement at time T _i, the until the next measurement can be performed to complete the online determination of the comparison determination of the monitoring value Z _i at the time T _i.

As is apparent from the above description, the waveform abnormality detection device according to the second embodiment of the present invention has a monitoring value Z _i that is a peak value corresponding to the time of the monitored signal waveform Z generated by the signal source 120B to be inspected, and By comparing the reference value Xu _i · Xl _i which is the peak value corresponding to the time of the set reference signal waveform Xu · Xl, it is determined whether or not there is a difference exceeding the allowable error range in the peak value of both waveforms. In this method, the upper and lower limits of the allowable error range are generated by the band upper limit data generation unit 1410 and the band lower limit data generation unit 1420, and the comparison determination is executed by the band comparison unit 1610. It has come to be.
The reference values Xu _i and Xl _i of the reference signal waveforms Xu and Xl are created based on the estimated theoretical value of the peak value of the reference waveform corresponding to the passage of time or the measured value of the standard sample.
The band upper limit data generating means 1410 has an upper correction value A0 _i obtained by adding a predetermined bias value α _i to the reference value Xu _i , and a front obtained by moving the upper correction value A0 _i back and forth on the time axis. of the correction value Al _i and rear correction value A2 _i, using the larger value of at least forward correction value Al _i and rear correction value A2 _i, comparison upper limit overlapping portion of the waveform to form an upper envelope waveform U _i is generated.
Band lower data generation unit 1420, and a lower correction value B0 _i obtained from the reference lower limit value Xl _i by subtracting a predetermined bias value beta _i, obtained by moving the lower correction value B0 _i before and after the time axis using at least smaller value of the forward correction value Bl _i and rear correction value B2 _i, overlapping portions of the waveform to produce a comparison limit value L _i to form a lower envelope waveform.
The band comparison unit 1610 makes an abnormality determination when the monitored value Z _i of the monitored signal waveform Z is outside the band between the comparison upper limit value U _i and the comparison lower limit value L _i .

The reference value of the reference signal waveform X is the upper limit signal waveform Xu based on the estimation theoretical lower limit value of the peak value of the reference waveform corresponding to the passage of time, the actual measurement value of the standard sample, or the actual measurement value of the standard sample. And a reference upper limit value Xu _i and a reference lower limit value Xl _i based on the lower limit signal waveform Xl.
Upper limit bandwidth data generating unit 1410, an upper correction value A0 _i obtained by adding a predetermined bias value alpha _i to the reference upper limit value Xu _i, obtained by moving the upper correction value A0 _i before and after the time axis By using at least the larger value of the forward correction value Al _i and the backward correction value A2 _i , the comparison upper limit value U _{i in} which the overlapping portion of the waveform forms the upper envelope waveform is generated.
Band lower data generation unit 1420, and a lower correction value B0 _i obtained from the reference lower limit value Xl _i by subtracting a predetermined bias value beta _i, obtained by moving the lower correction value B0 _i before and after the time axis using at least smaller value of the forward correction value Bl _i and rear correction value B2 _i, overlapping portions of the waveform to produce a comparison limit value L _i to form a lower envelope waveform.
Thus, the band upper limit waveform and the band lower limit waveform are generated based on the upper limit signal waveform and the lower limit signal waveform related to the reference signal waveform. Accordingly, accurate abnormality determination can be performed without excessively increasing the bandwidth.

A constant value proportional to the vertical error width ΔX is applied to the bias value α _i added in the upper correction value A0 _i or the bias value β _i subtracted in the lower limit correction value B0 _i . This is a deviation value between the reference upper limit value Xu _i and the reference lower limit value Xli _i when the waveform fluctuations of the signal waveform Xu and the lower limit signal waveform Xl are slow.
Thus, the bias value for upward or downward correction is a value proportional to the vertical error width of the reference waveform. Accordingly, accurate abnormality determination can be performed without excessively increasing the allowable error of the peak value.

A constant value proportional to the forward / backward error width ΔT is applied to the moving time for moving back and forth along the time axis in the forward correction values A1 _i and B1 _i or the backward correction values A2 _i and B2 _i .
The front-rear error width ΔT is the first time that is the first time when the upper limit signal waveform Xu has undergone a steep waveform change of a predetermined value or more, and the first time when the lower limit signal waveform Xl has undergone a steep waveform change of a predetermined value or more This is the time difference from the second time, which is the time.
Thus, the moving time width for forward or backward correction is a value proportional to the error width before and after the reference waveform. Therefore, accurate abnormality determination can be performed without allowing the allowable time error to be excessive.

Embodiment 3 FIG.
FIG. 17 which is a circuit block diagram of the waveform abnormality detection device according to the third embodiment of the present invention will be described focusing on differences from the waveform abnormality detection device according to the first embodiment.
In each figure, the same numerals indicate the same or corresponding parts.
In FIG. 17, the waveform abnormality detection device 100 </ b> C mainly includes a microprocessor 110, and a program memory 111 </ b> C storing various control programs (to be described later) and a RAM memory 112 for arithmetic processing are connected to the microprocessor 110 through a bus.
The signal source 120C to be inspected generates a processing power waveform signal in a high-frequency quenching in which processing power changes according to a workpiece rotation angle (not shown) or a laser processing machine or the like as a measurement signal.
The inspected signal source 120C generates a rotation pulse as a time signal at an interval of 10 degrees of the rotation angle of the workpiece and a measurement command signal of one pulse per one rotation of the workpiece, via the interface circuit l16. The clock signal CLK and the trigger signal input TR are supplied to the microprocessor 110.

In response to the generation of the measurement command signal, the signal source 120C to be inspected generates an analog measurement signal, and a digital value is input to the microprocessor 110 as measurement data DATA via the AD converter 114. .
The microprocessor 110 supplies an AD conversion command ST to the AD converter 114 in response to the clock signal CLK which is a rotation pulse signal, and the AD converter 114 returns an AD conversion completion signal OK to the microprocessor 110. At this time, the microprocessor 110 transfers and writes the measurement data DATA to the RAM memory 112.
However, the transfer address of the measurement data DATA to the RAM memory 112 is sequentially updated according to the measurement timing, and the measurement value corresponding to each measurement time is stored.
The setting indicator 130C is connected to the microprocessor 110 via a serial interface circuit (not shown), and performs a man-machine interface for displaying an abnormality detection state.

Next, FIG. 18, which is a characteristic diagram showing the reference signal waveform, monitored signal waveform, and reference band waveform in the waveform abnormality detection device according to the third embodiment of the present invention, will be described.
FIG. 18A shows the waveform of the measurement command signal generated by the signal source 120C to be inspected.
FIG. 18B shows a waveform of the rotation pulse signal generated by the signal source 120C to be inspected.
FIG. 18C shows a reference signal waveform X that responds to the measurement command signal, and a curve connecting the coordinate points (T _i , X _i ) of the reference value X _i corresponding to each time T _i is a reference signal. A waveform X is obtained.
Each time T _i in the third embodiment has a time obtained by counting the rotation pulse signal supplied from the inspection signal source 120C as a clock signal.
The reference signal waveform X is a machining power waveform corresponding to the time (rotation angle) when a standard work sample is used.
FIG. 18D shows reference band waveforms L and U generated as described below, and connects the coordinate points (T _i , L _i ) of the comparison lower limit value L _i corresponding to each time T _i . The curve is the band upper limit waveform L, and the curve connecting the coordinate points (T _i , U _i ) of the comparison upper limit value U _i corresponding to each time T _i is the band upper limit waveform U.

FIG. 18E shows a monitored signal waveform Z that is a machining power waveform related to the workpiece to be inspected, and the coordinate point (T _i , Z _i ) of the monitored value Z _i corresponding to each time T _i (rotation angle). ) Is a monitored signal waveform Z.
FIG. 18 (F) shows the monitored differential waveform ΔZ related to the monitored signal waveform Z shown in FIG. 18 (E). The monitored value Z _i corresponding to the current time T _i (rotation angle), A curve connecting the coordinate points (T _i , ΔZ _i ) of the monitored difference value ΔZ _i = Z _i −Z _i-1 with the monitored value Z _i-1 corresponding to the previous time T _i-1 is the monitored difference waveform. ΔZ.
The band upper limit waveform U in FIG. 18D is an upper envelope waveform using a forward movement waveform Al and a backward movement waveform A2 obtained by moving an upward movement waveform A0 (not shown) before and after the time axis.
The band lower limit waveform L in FIG. 18D is a lower envelope waveform using a forward movement waveform Bl and a backward movement waveform B2 obtained by moving a downward movement waveform B0 (not shown) forward and backward.

In FIG. 18F, time T _j is the time when the monitored difference value ΔZ _i changes from positive to negative, and time T _k is the time when the monitored difference value ΔZ _i changes from negative to positive.
The monitoring values Z _j and Z _{k at} times T _j and T _k are peak values of a convex waveform and a concave waveform as shown in FIG.
And comparing the upper limit value _{U j} for monitoring value _{Z j,} comparing the lower limit value _{L k} for the monitoring value _{Z k} is is as shown in FIG. 18 (D), and comparing the lower limit value _{L j} for monitoring value _{Z j,} monitoring value Z _The comparison upper limit value U _{k for k} is not shown in FIG. 18D, and the dent portion of the waveform is canceled by the envelope. Therefore, in the determination based on the reference band waveforms U and L in FIG. 18D, it can be determined whether the monitoring value Z _j is excessive or the monitoring value Z _k is excessive, but the monitoring value Z _j is There is a problem that it cannot be determined that the monitoring value _Zk is excessively small or the monitoring value _Zk is excessively large.

To solve this problem, a comparison limit value _{L j} at time _{T j,} the comparison the upper limit _{U k} at time _{T k} and calculates by equation (9), equation (l0).

L _j = U _j − (α _j + β _j + | ΔZ _j |) (9)
U _k = L _k + (α _k + β _k + | ΔZ _k |) (10)

Therefore, in the third embodiment, the first band comparison determination based on the reference band waveforms U and L in FIG. 18D and band comparison values (U _j , L _j ), ( Comparison determination is performed by the second band comparison determination by U _k , L _k ).
Note that the descending inversion time T _j is determined at the next time when the previous value of the monitored difference value ΔZ _i is positive and the next value is negative, and the rising inversion time T _k is the monitored difference value ΔZ _i. When the previous value is negative and the next value is positive, it is determined at the next time.
In addition, when the waveform change rate after the peak value Z _i of the monitored signal waveform Z suddenly increases or decreases rapidly is slow over a predetermined period, the descending inversion time T _j or the increasing inversion time T _k is not detected. The second band comparison determination based on the peak value detection is not performed.
However, the slow change rate excluded from the application is the case where the absolute value of the difference value ΔZ _i is equal to or less than the bias values α _j and β _j applied in the upward correction or downward correction, and the slow period excluded from the application is forward correction. And the total time width of the travel time applied in the correction.

  In the above description, the reference band waveforms U and L are determined based on one reference signal waveform X. However, as shown in FIG. 12D, the reference band waveform U is based on the upper limit reference waveform Xu and the lower limit signal waveform Xl. , L can be determined.

Next, FIG. 19 which is an operation flowchart of the reference signal waveform measurement and the reference band waveform generation in the waveform abnormality detection device diagram according to the third embodiment of the present invention will be described.
In FIG. 19, step 1900 is a step in which the microprocessor 110 starts the measurement of the reference signal waveform and the generation operation of the reference band waveform based on the operation command from the setting indicator 130C.
In subsequent step 1901, it is determined whether or not the measurement command signal shown in FIG. 18A has been received. If it has not been received, NO is determined and the process returns to step 1901. If it is received, YES is determined. This is a determination step for shifting to step 1902.
Step 1902 is a block for creating a data table of time T _i vs. reference value X _i , and subsequent step 1903 counts the number of generations of the rotation pulse signal shown in FIG. It is determined whether or not the count value of the time counter has reached a predetermined number of times. If not, NO is determined and the process returns to Step 1902. If it reaches, YES is determined and Step 1904 is determined. This is a determination step for shifting to.

Step 1904 is a block for calculating a band upper limit value U _i based on the reference value X _i measured in step 1902, and subsequent step 1905 is a band lower limit value L _i based on the reference value X _i measured in step 1902. The subsequent block 1906 is the operation end step.
The process block 1902 is a reference data writing unit, the process block 1904 is a band upper limit data generating unit, and the process block 1905 is a band lower limit data generating unit. By these units, the band shown in FIG. An upper limit waveform U and a band lower limit waveform L are generated.

Next, FIG. 20 which is an operation flowchart of monitored signal waveform measurement and comparison determination in the waveform abnormality detection device according to the third embodiment of the present invention will be described.
In FIG. 20, a process 2000 is an operation start step of monitored signal waveform measurement and comparison determination, and a subsequent process 2001 determines whether or not the measurement command signal shown in FIG. A NO determination is made to return to step 2001, and if received, a YES determination is made and a determination step moves to step 2002.
Step 2002 is a block for creating a data table of time T _i vs. monitoring value Z _i , and subsequent step 2003 counts the number of occurrences of the rotation pulse signal shown in FIG. It is determined whether or not the count value of the time counter has reached a predetermined number of times. If not, NO is determined and the process returns to Step 2002. If it has reached, the determination is YES and Step 2004 is performed. This is a determination step for transition.
Note that the step 2003 can be provided as a step 2008 after the step 2007, as will be described later with reference to FIG.

Step 2004 is a block that serves as a first band comparison unit that compares the monitored value Z _i measured in step 2002 with the comparison upper limit value U _i generated in steps 1904 and 1905 and the comparison lower limit value L _i. is there.
A subsequent step 2005a is a block serving as monitoring difference value data generating means for calculating the difference value ΔZ _i shown in FIG. 18F using the monitoring value Z _i measured in step 2002.
The subsequent step 2005b is a block serving as a peak point comparison data generating means for calculating the convex point comparison lower limit value L _j and the concave point comparison upper limit value U _k by the equations (9) and (10).
In the subsequent step 2006, the monitoring value Z _i measured in the step 2002 is compared with the monitoring values Z _j and Z _{k at} the times T _j and T _k and the peak point comparison data calculated in the step 2005b. This block is a block comparing / determining means, and a process block 2010 constituted by steps 2004 to 2006 constitutes a band comparing means.
Subsequent step 2007 is a block serving as a comprehensive determination unit that performs abnormality notification when abnormality determination is performed in either step 2004 or step 2006, and then proceeds to operation end step 2009.

Next, FIG. 21 and FIG. 22 which are characteristic line diagrams showing comparison determination timings in the waveform abnormality detection device according to the third embodiment of the present invention will be described.
In FIG. 21, which is the case of the posterior determination method, FIG. 21A shows a measurement command signal generated by the signal source 120C to be inspected. In this case, the measurement command signal is repeatedly generated. In the flowchart, the operation is returned to the operation start step 2000 following the operation end step 2009.
FIG. 21B shows the timing at which a series of measurement data is written to the RAM memory 112 in step 2002 of FIG.
FIG. 21C shows a timing at which a series of measurement data is compared and determined in step 2010 of FIG. 20 and an abnormality notification is performed in step 2007 if there is an abnormality.

In FIG. 22, which is the case of the simultaneous determination method in which the process 2003 of FIG. 20 is abolished and moved to the position of the process 2008 indicated by the dotted line, FIG. 22A is a measurement command signal generated by the signal source 120C to be inspected. In this case, the measurement command signal is repeatedly generated. In the flowchart of FIG. 20, the operation is returned to the operation start process 2000 following the operation end process 2009.
FIG. 22B shows the timing at which a series of measurement data is sequentially written into the RAM memory 112 in step 2002 of FIG.
FIG. 22C shows a timing at which a series of measurement data is sequentially compared and determined by the process block 2010 in FIG.

As is apparent from the above description, the waveform abnormality detection device according to the third embodiment of the present invention has a monitoring value Z _i that is a peak value corresponding to the time of the monitored signal waveform Z generated by the signal source to be inspected, and By comparing the reference value X _i that is the peak value corresponding to the time of the set reference signal waveform X, it is compared and determined at every predetermined time whether there is a difference between the peak values of both waveforms exceeding the allowable error range. In this waveform abnormality detection method, the upper and lower limit values of the allowable error range are generated by the band upper limit data generation unit 1904 and the band lower limit data generation unit 1905, and the comparison determination is executed by the band comparison unit 2010. Yes.
The reference value X _i of the reference signal waveform X is created based on the estimated theoretical value of the peak value of the reference waveform corresponding to the passage of time or the measured value of the standard sample.
The band upper limit data generating means 1904 has an upper correction value A0 _i obtained by adding a predetermined bias value α _i to the reference value X _i , and a front obtained by moving the upper correction value A0 _i back and forth on the time axis. Using at least the larger value of the correction value A1 _i and the backward correction value A2 _i , the comparison upper limit value U _i is generated in which the overlapping portion of the waveform forms the upper envelope waveform.
Band lower data generation unit 1905, and a lower correction value B0 _i obtained from the reference value X _i by subtracting a predetermined bias value beta _i, forward obtained by moving the lower correction value B0 _i before and after the time axis Using at least the smaller value of the correction value B1 _i and the backward correction value B2 _i , a comparison lower limit value L _{i in} which the overlapping portion of the waveform forms a lower envelope waveform is generated.
The band comparison means 2010 makes an abnormality determination when the monitored value Z _i of the monitored signal waveform Z is outside the band between the comparison upper limit value U _i and the comparison lower limit value L _i .

The band comparison unit 2010 includes a monitoring difference value data generation unit 2005a, a peak point comparison data generation unit 2005b, a first band comparison determination unit 2004, and a second band comparison determination unit 2006.
The monitoring difference value data generation unit 2005a obtains the difference value ΔZ _i = Z _i −Z _i−1 between the monitoring values Z _i and Z _{i−1 at} the time T _i and T _{i−1 that} precede and follow the monitored signal waveform Z. calculated, the difference value [Delta] Z _i is lowered inversion time T _j is inverted from positive to negative or the difference value [Delta] Z _{i detects} the rising inversion time T _k which are positively inverted from the negative.
The peak point comparison data generation unit 2005b calculates an addition value of the upper correction bias value α _j , the lower correction bias value β _j, and the absolute value of the difference value ΔZ _j from the convex point comparison upper limit value U _{j at} the descending inversion time T _j . The convex point comparison lower limit value L _j = U _j − (α _j + β _j + | ΔZ _j |) obtained by subtraction is calculated. Further, the recessed points compared lower limit L _k at elevated inversion time T _k, obtained by adding the sum of the absolute values of the upper correction bias value alpha _k and a lower correction bias value beta _k and the difference value [Delta] Z _k recessed points The comparison upper limit value U _k = L _k + (α _k + β _k + | ΔZ _k |) is calculated.

The first band comparison determination unit 2004 determines that the monitored value Z _i of the monitored signal waveform Z is a band between the comparison upper limit value U _i by the band upper limit data generation unit 1904 and the comparison lower limit value L _i by the band lower limit data generation unit 1905. Anomaly judgment is performed when it is outside.
The second band comparison determination unit 2006 determines whether the convex waveform peak value Z _{j at} the downward inversion time T _j is outside the range between the convex point comparison lower limit value L _j and the convex point comparison upper limit value U _j , or ascending inversion凹波shaped peak value Z _k at time T _k performs an abnormality determination when in outside between the concave point comparison limit value L _k and recessed point comparison upper limit U _k. The monitored signal waveform Z is determined to be abnormal when any one of the first band comparison / determination means 2004 and the second band comparison / determination means 2006 is determined to be abnormal.
As described above, abnormality determination by the first band comparison determination unit and abnormality determination of the waveform peak value by the second band comparison determination unit are performed, and if either one is abnormal, the abnormality determination is performed collectively. It has become. Therefore, when the monitored signal waveform includes a short-time convex waveform or concave waveform, even if it is a band comparison means using the envelope waveform, the peak value of the convex waveform is determined by the waveform peak value abnormality determination. An insufficiency or an inadequate depth of the concave waveform can be detected.
It should be noted that an excessive peak value of the convex waveform or an excessive depth of the concave waveform is detected by the first band comparison determination means, and a more accurate abnormality determination can be performed in total.

The descending inversion time T _j is determined at the next time when the previous value of the difference value ΔZ _i is positive and the next value is negative. The rising inversion time T _k is the previous value of the difference value ΔZ _i. Is negative and the next value is positive, it is determined at the next time.
Thus, the peak time of the monitored signal waveform is detected by the time immediately after the sign of the difference value of the monitored value is inverted. Therefore, the peak point time can be determined immediately without going back to the past time.

When the waveform change rate after the peak value Z _i of the monitored signal waveform Z suddenly increases or decreases rapidly is slow for a predetermined period or more, the descending inversion time T _j or the increasing inversion time T _k is not detected, and the peak The second band comparison determination means 2006 based on value detection is excluded from application.
The slow change rate exempted is the case where the absolute value of the difference value ΔZ _i is equal to or less than the bias values α _j and β _j applied by the upward correction or the downward correction, and the slow period excluded from the application is the forward correction and This is a case where the total time width of the travel time applied in the backward correction is equal to or greater.
As described above, when the period in which the change in the monitored signal waveform is slow continues, the second band comparison / determination means is not applied. Therefore, when a sudden waveform change has occurred with the first band comparison / determination unit as a main body, it is possible to make an accurate abnormality determination by dividing the role depending on the second band comparison / determination unit. it can.

1 is a circuit block diagram of a waveform abnormality detection device according to a first embodiment of the present invention. It is a characteristic diagram of a reference signal waveform, a monitored signal waveform, and a reference band waveform in the waveform abnormality detection device according to the first embodiment of the present invention. It is a figure for demonstrating the production | generation method of the band upper limit waveform and band lower limit waveform in the waveform abnormality detection apparatus by Embodiment 1 which concerns on this invention. FIG. 5 is a characteristic diagram of a band upper limit waveform and a band lower limit waveform generated by the waveform abnormality detection device according to the first embodiment of the present invention. It is the characteristic line figure which showed the reference | standard band waveform and the monitored signal waveform in the waveform abnormality detection apparatus by Embodiment 1 which concerns on this invention. It is a list for comparison judgments in the waveform abnormality detection device according to the first embodiment of the present invention. It is an operation | movement flowchart of the reference signal waveform measurement in the waveform abnormality detection apparatus by Embodiment 1 which concerns on this invention. It is an operation | movement flowchart of the reference | standard band waveform production | generation in the waveform abnormality detection apparatus by Embodiment 1 which concerns on this invention. It is an operation | movement flowchart of the to-be-monitored signal waveform measurement in the waveform abnormality detection apparatus by Embodiment 1 which concerns on this invention. It is an operation | movement flowchart of the comparison determination in the waveform abnormality detection apparatus by Embodiment 1 which concerns on this invention. It is a circuit block diagram of the waveform abnormality detection apparatus by Embodiment 2 which concerns on this invention. It is a characteristic diagram of a reference signal waveform, a monitored signal waveform, and a reference band waveform in the waveform abnormality detection device according to the second embodiment of the present invention. It is an operation | movement flowchart of the reference signal waveform measurement in the waveform abnormality detection apparatus by Embodiment 2 which concerns on this invention. It is an operation | movement flowchart of the reference | standard band waveform production | generation in the waveform abnormality detection apparatus by Embodiment 2 which concerns on this invention. It is an operation | movement flowchart of the to-be-monitored signal waveform measurement in the waveform abnormality detection apparatus by Embodiment 2 which concerns on this invention. It is an operation | movement flowchart of the comparison determination in the waveform abnormality detection apparatus by Embodiment 2 which concerns on this invention. It is a circuit block diagram of the waveform abnormality detection apparatus by Embodiment 3 which concerns on this invention. It is a characteristic diagram of a reference signal waveform, a monitored signal waveform, and a reference band waveform in the waveform abnormality detection device according to the third embodiment of the present invention. It is an operation | movement flowchart of the reference signal waveform measurement in the waveform abnormality detection apparatus by Embodiment 3 which concerns on this invention, and a reference | standard band waveform production | generation. It is an operation | movement flowchart of the to-be-monitored signal waveform measurement in the waveform abnormality detection apparatus by Embodiment 3 which concerns on this invention, and a comparison determination. It is a characteristic diagram which showed the comparison determination timing by the posterior determination in the waveform abnormality detection apparatus by Embodiment 3 which concerns on this invention. It is a characteristic diagram which showed the comparison determination timing by the simultaneous determination in the waveform abnormality detection apparatus by Embodiment 3 which concerns on this invention.

Explanation of symbols

  100A, 100B, 100C Waveform abnormality detection device, 110 microprocessor, 111A, 111B, 111C program memory, 112 RAM memory, 113 interface circuit, 114 AD converter, 115, 116 interface circuit, 120A, 120B, 120C Signal source to be inspected , 130A, 130B, 130C Setting display.

Claims (11)

Compare the peak value corresponding to the time of the monitored signal waveform generated by the signal source to be inspected with the peak value corresponding to the time of the preset reference signal waveform, and the peak value of both waveforms exceeds the allowable error range. A waveform abnormality detection device that compares and determines whether or not every predetermined time,
The reference value of the reference signal waveform is created in advance based on the estimated theoretical value of the peak value of the reference signal waveform corresponding to time, or the measured value of the standard sample of the signal source to be inspected,
A forward correction value obtained by moving an upward correction value obtained by adding a predetermined first bias value to the reference value previously on the time axis by a predetermined movement time or an upward correction value obtained on the time axis. Band upper limit data generating means for generating a comparison upper limit value in which the overlapping portion of the waveform forms an upper envelope waveform, using the larger value of the backward correction value obtained by moving after a predetermined movement time later,
A forward correction value obtained by moving a downward correction value obtained by subtracting a predetermined second bias value from the reference value previously for a predetermined movement time on the time axis, or the downward correction value obtained on the time axis. The band lower limit data generating means for generating a comparison lower limit value in which the overlapping portion of the waveform forms a lower envelope waveform by using the smaller value of the backward correction value obtained by moving after a predetermined movement time later,
Band comparison means for performing abnormality determination when the monitored value of the monitored signal waveform is outside the band between the comparison upper limit value and the comparison lower limit value;
A waveform abnormality detection device comprising:   The first bias value or the second bias value is a constant value at each time of the reference signal waveform or a value proportional to a reference value at each time of the reference signal waveform. Item 4. The waveform abnormality detection device according to Item 1. The comparison upper limit value is a maximum value among the upper correction value, the front correction value related to the upper correction value, or the rear correction value related to the upper correction value,
The comparison lower limit value is a minimum value of the downward correction value, the forward correction value related to the downward correction value, or the backward correction value related to the downward correction value. Waveform abnormality detection device. Compare the peak value corresponding to the time of the monitored signal waveform generated by the signal source to be inspected with the peak value corresponding to the time of the preset reference signal waveform, and the peak value of both waveforms exceeds the allowable error range. A waveform abnormality detection device that compares and determines whether or not every predetermined time,
The reference value of the reference signal waveform is an upper limit based on an estimated theoretical value of the peak value of the reference signal waveform corresponding to time, a plurality of actual measurement values of the standard sample of the signal source to be inspected, or an actual measurement value of the plurality of standard samples. Reference upper limit value and reference lower limit value created in advance from signal waveform and lower limit signal waveform,
The forward correction value obtained by moving the upper correction value obtained by adding the predetermined first bias value to the reference upper limit value on the time axis by the movement time before or the upper correction value obtained later on the time axis. Band upper limit data generating means for generating a comparison upper limit value in which the overlapping portion of the waveform forms an upper envelope waveform using the larger value of the backward correction value obtained by moving only by the moving time;
A forward correction value obtained by moving a downward correction value obtained by subtracting a predetermined second bias value from the reference lower limit value forward on the time axis or a downward correction value obtained by moving the downward correction value later on the time axis. Band lower limit data generating means for generating a comparison lower limit value in which the overlapping portion of the waveform forms a lower envelope waveform using the smaller value of the backward correction value when obtained,
Band comparison means for performing abnormality determination when the monitored value of the monitored signal waveform is outside the band between the comparison upper limit value and the comparison lower limit value;
A waveform abnormality detection device comprising:   The first bias value or the second bias value is a constant value proportional to a deviation value between the reference upper limit value and the reference lower limit value at the time when the waveform fluctuation of the upper limit signal waveform and the lower limit signal waveform is slow. The waveform abnormality detection device according to claim 4, wherein:   The movement time for moving back and forth on the time axis to obtain the forward correction value or the backward correction value is the first time when the upper limit signal waveform changes sharply beyond a predetermined value, and the lower limit signal waveform is a predetermined value. 5. The waveform abnormality detection device according to claim 4, wherein the waveform abnormality detection device is a constant value proportional to a time difference from the second time at which the waveform changes steeply as described above. Compare the peak value corresponding to the time of the monitored signal waveform generated by the signal source to be inspected with the peak value corresponding to the time of the preset reference signal waveform, and the peak value of both waveforms exceeds the allowable error range. A waveform abnormality detection device that compares and determines whether or not every predetermined time,
The reference value of the reference signal waveform is created in advance based on the estimated theoretical value of the peak value of the reference signal waveform corresponding to time, or the measured value of the standard sample of the signal source to be inspected,
A forward correction value obtained by moving an upward correction value obtained by adding a predetermined first bias value to the reference value previously on the time axis by a predetermined movement time or an upward correction value obtained on the time axis. Band upper limit data generating means for generating a comparison upper limit value in which the overlapping portion of the waveform forms an upper envelope waveform, using the larger value of the backward correction value obtained by moving after a predetermined movement time later,
A forward correction value obtained by moving a downward correction value obtained by subtracting a predetermined second bias value from the reference value previously for a predetermined movement time on the time axis, or the downward correction value obtained on the time axis. The band lower limit data generating means for generating a comparison lower limit value in which the overlapping portion of the waveform forms a lower envelope waveform by using the smaller value of the backward correction value obtained by moving after a predetermined movement time later,
Monitoring difference value data generating means for detecting a falling inversion time at which the difference value of the monitoring value at the time before and after the monitored signal waveform is inverted from positive to negative, or an rising inversion time at which the difference value is inverted from negative to positive When,
Convex point comparison lower limit value obtained by subtracting the sum of absolute values of the first bias value, the second bias value and the difference value from the convex point comparison upper limit value at the descending inversion time and the upside inversion Peak point comparison data for calculating a concave point comparison upper limit value obtained by adding the absolute value of the first bias value, the second bias value and the difference value to the concave point comparison lower limit value at the time Generating means;
First band comparison determination means for performing abnormality determination when the monitored value of the monitored signal waveform is outside the band between the comparison upper limit value and the comparison lower limit value;
When the convex waveform peak value at the descending reversal time is outside the range between the convex point comparison lower limit value and the convex point comparison upper limit value, or the concave waveform peak value at the ascent reversal time and the concave point comparison lower limit value A second band comparison determination means for performing an abnormality determination when outside the range between the concave point comparison upper limit,
Have
A waveform abnormality detection device characterized in that, when an abnormality is determined by one of the first band comparison determination means or the second band comparison determination means, the monitored signal waveform is determined to be abnormal. The descending inversion time is determined at the next time when the previous value of the difference value is positive and the next value is negative,
8. The waveform abnormality detection device according to claim 7, wherein the rising inversion time is determined at the next time when the previous value of the difference value is negative and the next value is positive. The monitoring difference value data generation means does not detect the descending inversion time or the ascending inversion time when the absolute value of the difference value is equal to or less than the first bias value or the second bias value,
The second band comparison determination means detects the descending inversion time or the ascending inversion time over a total time of a traveling time for obtaining the forward correction value and a traveling time for obtaining the backward correction value. The waveform abnormality detection device according to claim 7 or 8, wherein abnormality determination is not performed when it is not performed. A RAM memory in which a monitored value of a monitored signal waveform generated by the signal source to be inspected is stored;
A program memory in which a control program is stored in cooperation with the microprocessor;
With
10. The waveform abnormality detection device according to claim 1, wherein the RAM memory or the program memory stores the comparison upper limit value and the comparison lower limit value corresponding to time. The above microprocessor is connected to a setting display which is a man-machine interface,
The setting indicator notifies the abnormality of the monitored signal waveform, and the comparison upper limit value and the comparison lower limit value stored in the program memory or the RAM memory, a reference value for the reference signal waveform, The waveform abnormality detection device according to claim 10, wherein the setting of the movement time is changed to correct the bias value of 1 and the second bias value, or forward / backward movement correction.

Images