JPH08166820A - Method and device for detecting abnormality of measuring instrument - Google Patents

Abstract

PURPOSE: To improve the detection precision even for the distribution in which the frequency distributions of measured values of respective measuring instruments and measured value variation quantities are not in normal distribution by detecting abnormality of a measured value on the basis of a reference range determined on the basis of the distances of 1st - 3rd quartile points of the relative cumulative frequency of stored measured values. CONSTITUTION: A decision reference value arithmetic unit 20 sets a decision reference value (weighted quartile standard) on the basis of hinge width 1 (distance from 1st quartile point to 2nd quartile point of stored measured values) and hinge width 2 (distance from 2nd quartile point to 3rd quartile point) obtained by a hinge width arithmetic unit 19, and the 1st quartile point and 3rd quartile point. An abnormality decision device 22 compares a measured value stored newly in a measured value storage device 18 with the decision reference value stored in a decision reference value storage device 21 to decide that a measured value meeting a condition of 1st quartile point - m × hinge width 1 < measured value < 3rd quartile point + m ×hinge width 2 (m: positive optional number) is normal. Namely, the standard is weighted so that the normal range is wide in the direction wherein the frequency distribution is trailing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormality detecting method and apparatus for detecting an abnormality in a measuring instrument used in a plant such as a water and sewage treatment equipment.

[0002]

2. Description of the Related Art FIG. 4 is a block diagram showing a sewage treatment system and a conventional measuring device abnormality detecting device (standard deviation standard). In all the drawings, the same reference numerals indicate the same or corresponding members. The sewage containing pollutants in the sewage treatment system is first introduced into the settling tank 2 from the pipeline 1, and the pollutants that easily settle are separated by sedimentation, and the supernatant water is aerated in the aeration tank 3.
Leaked to In the aeration tank, the returned sludge extracted from the final settling tank 4 is returned via the pipe line 5 and the diffuser pipe
Air sent under pressure from a blower (not shown) is supplied to the aeration tank from (not shown), and contaminants are adsorbed by activated sludge,
It is decomposed and led to the final settling basin. The activated sludge is settled and separated in the final settling basin, the clear water is discharged as treated water from the pipe 6 through a sterilization tank (not shown), and the sludge that has settled is returned to the aeration tank, part of which is the sludge treatment process. (Not shown)
I will send it to you for processing. 7 is an inflow sewage flowmeter, 8 is an inflow water pH meter, 9 is an inflow water suspended matter concentration meter, 10 is an inflow water organic matter concentration meter, 11 is a dissolved oxygen concentration meter, 12 is an MLSS (mixed li-quor). suspended solid)
Concentration meter, 13 water temperature meter, 14 redox potentiometer, 15 pH meter for treated water, 16 concentration meter for suspended solids in treated water, 17
Is an organic matter concentration meter for treated water. As described above, in a sewage treatment plant, various kinds of water quality data are usually measured and collected. Abnormality detection of these measuring instruments is often performed using a judgment reference value set based on the experience of the operator. However, the determination reference value must be set for each measuring instrument according to the season, and changing the setting of the determination reference value has been a complicated task for the operator. Therefore, it has become common to use a computer to set the judgment reference value of the measuring instrument. In general, the setting of the judgment reference value by the computer is often decided by using the standard deviation or the like. A method of setting the judgment reference value using the standard deviation will be described below as a first conventional example. [First Conventional Example] A block diagram of an abnormality detecting device (standard deviation standard) of a measuring instrument in a sewage treatment system of the first conventional example shown in FIG. 4 will be described. 18 is a measurement value storage device that stores the measurement values obtained from each measuring device, 24 is a standard deviation calculation device for obtaining the standard deviation of each measuring device, and 25 is a standard deviation calculation device 24 Judgment reference value calculation device (standard deviation reference) for obtaining the judgment reference value, 2
Reference numeral 1 is a judgment reference value storage device for storing the result obtained by the judgment reference value calculation device (standard deviation reference) 25, and reference numeral 22 is a measurement value newly stored in the measurement value storage device 18 and is stored in the judgment reference value storage device 21. An abnormality determination device that compares the determination reference values that have been determined to determine whether or not there is an abnormality, and the determination result is displayed on the display device 2
3 is output. The standard deviation is given by the following [Formula 1].

[Equation 1] FIG. 6A shows the frequency distribution of the measured values in a certain period, and FIG. 6B shows the setting range of the judgment reference value based on the standard deviation. However, the frequency distribution of the measured values collected by the sewage treatment system is shown in Figure 6.
As shown in (a), since it does not always show a normal distribution, there is a difficulty in setting the judgment reference value based on the standard deviation. Therefore, using the quartile, which is one of the so-called exploratory data analysis methods proposed as a method for handling data whose frequency distribution is not a normal distribution, It is possible to make settings. A method of setting the judgment reference value using the quartile will be described below as a second conventional example. [Second Conventional Example] FIG. 5 is a block diagram showing an abnormality detecting device (quartile quantile standard) of a measuring instrument in a sewage treatment system. Reference numeral 19 is a hinge width calculation device for obtaining the hinge width of each measuring instrument, and 26 is a determination reference value calculation device for obtaining a determination reference value from the hinge width obtained from the hinge width calculation device 19 (quartile standard). Is. The idea of the quartile is explained below. The quartile is 25% of the relative cumulative frequency
Points, 50% points, and 75% points. The 25% points are called the first quartile, the 50% points are called the second quartile, and the 75% points are called the third quartile. The distance H from the 1st quartile to the 3rd quartile is used as the hinge width, and the judgment reference value is set based on the hinge width. The distance from each hinge (1st quartile, 3rd quartile) by 3 times or more of the hinge width is called far out and is regarded as an abnormal value. That is, in the determination method based on the quartile, a measured value that satisfies the following equation is determined to be normal. 1st quartile -3 x hinge width <measured value <3rd quartile + 3 x hinge width (4) For the period shown as an example in Fig. 6 (c) above. It shows the frequency distribution of measured values and the setting range of the judgment reference value by the quartile.
This is the setting range of the judgment reference value by the quartile. H
Indicates the hinge width.

[0003]

The abnormality detecting devices for measuring instruments of the first conventional example and the second conventional example are, for example, the standard deviation or quadrant of each measured value accumulated in the measured value accumulating device 18. The upper and lower limits of the measured value and the amount of change in the measured value were set using the orders. In particular, the second standard quartile judgment reference value calculation device 26 can handle the measured values and the measured value change amount whose frequency distribution is not a normal distribution. However, since this method is based on the assumption that the shape of the frequency distribution is symmetrical, no consideration is given to the distribution in which the frequency distribution is tailed to the left or right. Here, the present invention is a distribution in which the frequency distribution of the measurement value of each measuring instrument and the variation of the measurement value is not a normal distribution,
It is an object of the present invention to provide a method for setting a criterion value that can be applied to a distribution in which the frequency distribution is tailed to the left or right.

[0004]

In order to achieve the above object, the present invention is an abnormality detection method for detecting an abnormality of a measuring instrument for measuring a state of a process, in which measured values by the measuring instrument are accumulated and accumulated. The distance H1 from the first quartile to the second quartile of the relative cumulative frequency of measured values is the hinge width 1, the second
Hinge width 2 for the distance H2 from the quartile to the third quartile
Then, the hinge width 1 and the hinge width 2 are obtained from the past measured values, and the abnormality detection of the measuring instrument that detects the abnormality of the newly accumulated measured value based on the reference range determined based on the hinge width 1 and the hinge width 2 When the newly accumulated measured value deviates from the following equation, the range of the measured value represented by the following equation is used as a reference range, and the first quartile-m × hinge width 1 <measured value <Third quartile + m × hinge width 2 [where m is a positive arbitrary number] The method for detecting an abnormality of a measuring instrument according to the preceding paragraph, further measuring the state of the process In an abnormality detection device for detecting an abnormality of a measuring instrument, a measuring value accumulating device for accumulating the measured value by the measuring instrument, and a second quartile from the first quartile of the relative cumulative frequency of the accumulated measured values. From the hinge width 1 and the second quartile, which is the distance H1 to the point 3 and the hinge width computing unit for obtaining a hinge width 2, which is a distance H2 to the quartile,
A determination reference value calculation device that determines a reference value for determining abnormality of the measurement value based on the values determined based on the hinge width 1 and the hinge width 2 and the values of the first and third quartiles. An abnormality of the measuring instrument is detected by comparing the determination reference value storage device that stores the determination reference value with the determination reference value stored in the determination reference value storage device and the measurement value newly stored in the measurement value storage device. An abnormality detection device for a measuring instrument, comprising:

[0005]

With the above-mentioned means, the present invention provides a measurement value and a measurement value change amount in which the frequency distribution is not a normal distribution, a measurement value in which the frequency distribution is tailed to the left or right, and a criterion for determining the measurement value change amount. Since the value can be set, the accuracy of detecting abnormality of the measuring instrument is improved.

[0006]

Embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing an abnormality detection device for a measuring instrument of a sewage treatment system according to an embodiment of the present invention. 20 is the hinge width 1 obtained by the hinge width calculation device 19
(Distance H1 from first quartile to second quartile) and hinge width 2 (distance H2 from second quartile to third quartile) and first quartile, It is a criterion value calculator (weighted quartile standard) for setting a criterion value based on the second quartile. The concept of weighted quartiles will be described. The quartile is the relative cumulative frequency
25% point, 50% point, and 75% point. The 25% point is called the first quartile, the 50% point is called the second quartile, and the 75% point is called the third quartile. , The distance H1 from the first quartile to the second quartile
Is set as the hinge width 1, and the distance H2 from the second quartile to the third quartile is set as the hinge width 2, and the determination reference value is set based on the two hinge widths.

Each hinge (first quartile, third
The distance from the quartile point) to 6 times the hinge width 1 and 6 times the hinge width 2 is called the far-out and is an abnormal value as described above. That is, in the determination method based on the weighted quartile, the measured value that satisfies the following equation is determined to be normal. 1st quartile -6 x hinge width 1 <measured value <3rd quartile + 6 x hinge width 2 (5) Weighting based on equation (5) 4 The setting range of the determination reference value based on the quantiles and the setting range of the determination reference value based on the quartiles are the same, and it means that the weight is shifted in the weighted direction (the direction in which the tail is drawn).

FIG. 2A shows the frequency distribution of measured values and FIG. 2D shows the setting range of the judgment reference value of the present invention. That is, FIG. 2 (d)
Is the setting range of the judgment reference value (weighted quartile standard). H1 represents the hinge width 1, H2 represents the hinge width 2, and the sum of the hinge width 1 and the hinge width 2 represents the hinge width H [FIGS. 2 (b) and 2 (c) are conventional examples 1, 2 of FIG. 6 (b),
It is the same as FIG. 6 (c) and is transferred for comparison]. The present invention is based on the tailed direction of the frequency distribution (upper or lower)
In addition, weighting is performed so that the normal range is wide,
As shown in Fig. 2 (d), the hinge width is changed from the first quartile to the second.
It is divided into a section to the quartile and a section from the second quartile to the third quartile, and the lower distance is the distance H1 from the first quartile to the second quartile. As for the distance on the upper side, the judgment reference value is set based on the distance H2 from the second quartile to the third quartile.

Here, in the conventional method shown in FIGS. 2 (b) and 2 (c) and the method according to the present invention, the distance between the upper and lower limit values is the same, and the tail is drawn (upward in this case). Will be shifted to. Then, the detection result when weighting the judgment criteria of FIG. 2 (d) is only the points judged by the operator to be abnormal, and the actual value and the judgment value are completely coincident with each other. It is considered to be a reference value.

[0010]

As described above, the abnormality detection result of the measuring instrument when the judgment standard value of the standard deviation standard of the first conventional example is used is different from the result judged by the operator to be a normal value. Is detected as an abnormal value. Next, the abnormality detection result of the measuring instrument when using the determination standard value of the quartile of the second conventional example is considerably improved from the abnormality detection result of the determination standard value by the standard deviation, but is still normal. The value is judged as an abnormal value. Finally, the abnormality detection result of the measuring instrument when the quartile is weighted to the criterion value according to the present invention is:
This is in agreement with the point that the operator determined that there was no error, and there was no false detection. From this, according to the present invention, even in the case of the measured value whose frequency distribution does not show a normal distribution, the measured value change amount and the measured value in which the frequency distribution is tailed to the left or right, and the measured value change amount, The determination reference value can be automatically set, and a special effect of improving the accuracy of the abnormality detection of the measuring instrument can be achieved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a circuit configuration of an embodiment of the present invention.

FIG. 2 is a diagram showing a comparison between a frequency distribution of measured values, a determination reference value setting range of the present invention, and a determination reference value setting range of a conventional example.

FIG. 3 shows the present invention (when the judgment reference value is weighted), the first conventional example (when the standard deviation is used as the judgment reference value), and the second conventional example (the quartile for the judgment reference value). Figure comparing each abnormality detection result of the measuring instrument of the sewage treatment system (when using the number)

FIG. 4 is a block diagram showing a circuit configuration in a first conventional example (when a standard deviation is used as a judgment reference value).

FIG. 5 is a block diagram showing a circuit configuration in a second conventional example (when a quartile is used as a criterion value).

FIG. 6 is a diagram showing a comparison of frequency distributions of measured values and setting ranges of determination reference values of the first conventional example and the second conventional example.

[Explanation of reference symbols]

1, 5, 6 Pipeline 2 First settling tank 3 Aeration tank 4 Final settling tank 7 Inflowing sewage flowmeter 8 Inflowing water pH meter 9 Inflowing water suspended matter 10 Inflowing water organic matter concentration meter 11 Dissolved oxygen concentration meter 12 MLSS Concentration meter 13 Water temperature meter 14 Redox potential meter 15 pH meter of treated water 16 Suspended substance of treated water 17 Concentration meter of organic matter of treated water 18 Measured value accumulator 19 Hinge width calculator 20 Judgment standard value calculator (weighted Quantile standard performed) 21 Judgment reference value storage device 22 Abnormal value judgment device 23 Display device 24 Standard deviation calculation device 25 Judgment reference value calculation device (standard deviation reference) 26 Judgment reference value calculation device (quartile number reference) )

Claims (3)

[Claims] 1. An abnormality detecting method for detecting an abnormality of a measuring instrument for measuring a process state, wherein a measurement value of the measuring instrument is accumulated, and a first quartile of relative accumulated frequency of the accumulated measurement values. From the second quartile to the hinge width 1, the distance H2 from the second quartile to the third quartile is hinge width 2, and the hinge width 1 and the hinge width from the past measurement values. 2. A method for detecting abnormality of a measuring instrument, which comprises obtaining 2 and detecting an abnormality in a newly accumulated measurement value based on a reference range determined based on the hinge width 1 and the hinge width 2. 2. The measuring instrument is determined to be abnormal when the newly accumulated measured value deviates from the following equation with a range of the measured value represented by the following equation as a reference range.
Abnormality detection method of measuring instrument described. First quartile-m x hinge width 1 <measured value <third quartile + m x hinge width 2 where m is a positive arbitrary number. 3. An abnormality detection device for detecting an abnormality of a measuring instrument for measuring the state of a process, comprising: a measurement value accumulating device for accumulating the measurement value by the measuring instrument; and a relative cumulative frequency of the accumulated measurement values. Hinge width calculation to obtain the hinge width 1 which is the distance H1 from the 1st quartile to the 2nd quartile and the hinge width 2 which is the distance H2 from the 2nd quartile to the 3rd quartile Device and determination criteria for determining a reference value for determining abnormality of the measured value based on the values determined based on the hinge width 1 and the hinge width 2 and the values of the first and third quartiles A value calculation device, a determination reference value storage device that stores the determination reference value, a determination reference value stored in the determination reference value storage device, and a measurement value newly stored in the measurement value storage device are compared. An abnormality determination device that detects abnormalities in the measuring instrument Abnormality detecting device of the instrument to symptoms.