JP2978995B2 - Error detection method using instrument signals - Google Patents

Description

Description: BACKGROUND OF THE INVENTION The present invention relates to a method for detecting an abnormality in each part forming a process control system based on process data appearing on an instrument, and particularly to a method for detecting an abnormality at an early stage. The present invention relates to an abnormality detection method using an instrument signal that enables detection.

[Related Art] In process control for controlling various plants, a control target (process) forming a control system, a detector,
When an abnormality occurs in any of the transmitter, the controller, the operation unit, and the like, an abnormality signal appears in the control system. This abnormal signal is
Normally, it is confirmed when the upper and lower thresholds for abnormality confirmation set in the controller are exceeded, and only when this confirmation is made, the control system is notified that an abnormality has occurred.

However, there are some abnormal signals that do not reach the upper and lower thresholds for abnormality confirmation.In such a case, the controller cannot detect this abnormality even though the control system is abnormal, and The same control was performed as at the time, which could cause an unexpected accident.

Therefore, it is desired to detect a small abnormal signal at an early stage. For example, a threshold is set at a lower limit of a signal at a lower limit than a normal amplitude, and the number of unchangeable times that does not reach the small threshold is set. To detect abnormalities by using
No. 271418).

[Problems to be Solved by the Invention] However, the above-mentioned Japanese Patent Application Laid-Open No. 61-271418 has the following problems. That is, first, since the waveform of the abnormal signal is unknown, it is difficult to determine a threshold value smaller than the amplitude, and second, a signal whose amplitude has increased as a precursor of the abnormality is detected. There was a problem that it was not possible.

As a method of solving this problem, a method of performing a test by analysis of variance between the data at normal time and the measured data can be considered. In this case, firstly, the calculation becomes very troublesome, and secondly, Depending on the method of collecting data (population) during normal times, accuracy may be lacking. Thirdly, since statistical processing is performed, it is difficult to understand the results.

The present invention has been made in view of these problems, and an abnormality detection method based on an instrument signal that allows an abnormality occurring in a control system to be detected early based on process data represented as an instrument signal. The purpose is to provide.

[Means for Solving the Problems] In order to achieve the above object, the abnormality detection method using an instrument signal according to the present invention uses an increase / decrease tendency conversion point of process data represented as an instrument signal as a peak data, and a method for detecting abnormalities at normal times. At the same time as the number of occurrences of the peak data within a certain period of time, a pseudo amplitude is obtained from the peak data of the process data within a certain period of normal time, and an integrated value of the pseudo amplitude is obtained from the pseudo amplitude and the peak data. The average pseudo-amplitude is obtained from the sum of the number of appearances of the peak data and the pseudo-amplitude, and if the average pseudo-amplitude is outside a predetermined normal range, an abnormality detection signal is output. is there.

[Operation] According to the above method, the average pseudo amplitude approaches zero when the movement of the instrument signal decreases, and increases when the movement of the instrument signal increases. By determining the threshold value, the occurrence of an abnormal signal can be detected at an early stage.

[Embodiment] Hereinafter, an embodiment of the abnormality detection method using an instrument signal according to the present invention will be described.

FIG. 1 shows a block diagram of a system to which one embodiment of the method of the present invention is applied. In FIG. 1, reference numeral 1 denotes a data collection unit, which collects process data. This data collection unit 1
May be not only a detector for directly detecting data, but also a signal via a transmitter or a controller. That is, any instrument capable of confirming the process data as a waveform signal may be used.

Reference numeral 2 denotes a peak data counting unit, which counts the number of occurrences of peak data in process data during a certain period of time in a normal state based on a signal from the data collection unit 1.

Reference numeral 3 denotes a pseudo-amplitude integrating unit that obtains an integrated value of pseudo-amplitude of process data in a certain period of time based on a signal from the data collecting unit 1.

Reference numeral 4 denotes an average pseudo-amplitude calculation unit which calculates the number of peak data counted by the peak data counting unit 2 and the pseudo-amplitude integration unit 3
The average pseudo-amplitude is calculated based on the integrated value of the pseudo-amplitude obtained in (1).

Reference numeral 5 denotes a comparison unit that determines whether or not the average pseudo amplitude value calculated by the average pseudo amplitude calculation unit 4 is within a normal range.

Reference numeral 6 denotes a notifying unit, and when the comparing unit 5 determines that the average pseudo amplitude value is out of the normal range, the notifying unit 6 notifies an abnormality by voice, sound, blinking light, or the like.

Next, an embodiment method of the present invention will be described with reference to the flowchart of FIG.

The data collection unit 1 collects data such as level, pressure, temperature, and flow rate as process data, and outputs a part or all of the collected process data as an instrument signal as necessary.

The peak data counting unit 2 counts the number N ₀ of peak data in the process data (instrument signal) within a certain fixed time T ₀ in a normal state.

Here, the peak data, as shown by P ₁ to P ₈ in FIG. 3, the downward trend data from increasing, also refers to data P _i of point to be converted from decreasing to increasing.

In the normal state, for example, several times at a certain time interval, statistical processing of the pseudo amplitude data to determine the average value and standard deviation,
This means that there is no large difference between the average value and the standard deviation.

In the pseudo amplitude integrating unit 3 determines the integrated value S _SUM of the pseudo amplitude S _i in within a predetermined time T _0.

Here the pseudo amplitude S _{i is, S i = | P i -P} i-1 | determined by the formula, also the pseudo amplitude integrated value S _SUM is It is calculated by the following equation.

The average pseudo amplitude calculator 4 calculates the average pseudo amplitude E from the number N _{0 of} peak data and the integrated pseudo amplitude S _SUM.
Ask for.

Here, the average pseudo amplitude E is It is calculated by the following equation.

The comparison unit 5 compares the average pseudo-amplitude E with an upper threshold value and a lower threshold value of the amplitude which are determined in advance based on the process data in a normal state. It is determined whether or not.

L _MIN <E <L _MAX (L _MIN : upper threshold, L _MAX : lower threshold) Here, the upper threshold and the lower threshold are obtained by statistically processing the data at normal time and Determine the deviation. For example, the upper threshold value is calculated by the following equation: upper threshold value = average amplitude value at normal time + standard deviation × 3.

It should be noted that as normal data, fluctuation data such as charge-up and stable data without changing conditions are used.

As a result of the comparison, when the average pseudo amplitude E is out of the normal range, that is, when L _MIN ≧ E or L _MAX ≦ E, the notification unit 6 has caused an abnormality in the process data (instrument signal). Announce that.

According to such an abnormality detection method, the average pseudo amplitude E approaches 0 when the movement of the process data (instrument signal) decreases, and the average pseudo amplitude E approaches when the movement of the process data increases. growing. Therefore, an upper threshold value and a lower threshold value are determined based on the normal value of the average pseudo amplitude E, and by comparing these values with the average pseudo amplitude, abnormal signals such as hold and hunting can be detected. It can be detected reliably.

This makes it possible to detect a small abnormality that could not be detected by the conventional detection method.

Also, since the average pseudo amplitude indicates the pseudo amplitude of the instrument signal, it is easy to understand the state of the signal, and it is also easy to understand the upper and lower limits.

In addition, the signal detected by the above-described abnormality detection method can be used not only to notify the abnormality of the signal, but also as an emergency operation signal to an operating device or another device.

[Effects of the Invention] As described above, according to the present invention, it is possible to early detect an abnormality in each unit constituting the process control system based on an instrument signal.

[Brief description of the drawings]

FIG. 1 shows a block diagram of a system to which an embodiment of the method of the present invention is applied, FIG. 2 shows a flowchart for explaining an embodiment of the method of the present invention, and FIG. 3 explains peak data. FIG. 1: Data collection unit, 2: Peak data counting unit 3: Pseudo amplitude integration unit, 4: Average pseudo amplitude calculation unit 5: Comparison unit, 6: Notification unit

Claims (1)

(57) [Claims] A peak data is used as a conversion point of an increase / decrease tendency of process data expressed as an instrument signal, and the number of appearances of the peak data within a certain period of normal time is obtained. The pseudo-amplitude is obtained from the peak data of the process data at the time of, and the integrated value of the pseudo-amplitude is obtained from the pseudo-amplitude and the number of appearances of the peak data. An abnormality detection method using an instrument signal, wherein an amplitude is obtained, and an abnormality detection signal is output when the average pseudo amplitude deviates from a predetermined normal range.