JPH05203544A - Method and apparatus for automatic construction of abnormality diagnosing algorithm - Google Patents

Abstract

(57) [Summary] [Purpose] To construct an abnormality diagnosis algorithm that performs highly accurate diagnosis on various types of diagnosis targets. [Structure] Values relating to a plurality of diagnostic parameters are measured for diagnostic groups that are divided into those with normal properties and those with abnormal properties, and statistical processing is performed on the measured values for each diagnostic parameter. A plurality of diagnostic parameters that are predicted to be effective parameters are extracted from the diagnostic parameter group based on the result, and the quality of the diagnosis target is judged based on the measured values of the extracted effective diagnostic parameters. When determining the diagnostic level for normal products and constructing an algorithm to discriminate between normal and abnormal products based on the diagnostic level, if the average deviation value of normal products in the normal group is less than the standard value, a predetermined value for the normal product average value is set. The value of
Recalculate each measured value of the normal product as a virtual measured value. When the individual measured values are smaller than the average value of normal products, the deviation amount calculation means forces the calculated value of the deviation amount to zero.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a diagnostic algorithm construction method and apparatus, and more particularly, to a diagnostic algorithm for automatically diagnosing abnormalities of various devices used in automobiles, home electric appliances, power generation equipment and the like. The present invention relates to a method and apparatus for automatically constructing an abnormality diagnosis algorithm suitable for construction.

[0002]

2. Description of the Related Art An acoustic emission (AE) method is known as a method for diagnosing an abnormality in equipment.
As a diagnostic device using this AE method, for example, Japanese Patent Laid-Open Nos. 62-28258 and 62-1975.
The one described in Japanese Patent Publication No. 5 is known.

In the former device, a diagnosis is made for a diagnosis target by using a unique diagnosis algorithm, and in the latter device, a diagnosis algorithm is constructed using specific AE characteristic parameters. It has become.

[0004]

SUMMARY OF THE INVENTION The above-mentioned prior art is capable of making highly accurate diagnoses on various diagnostic objects. However, no consideration is given to versatility. Therefore,
Even if these methods are adopted, it is necessary to develop a diagnostic algorithm every time the device to be diagnosed changes, and there is a problem that the development efficiency is low. That is, in the case of the former device, the signal that has passed the band of the specific wavelength of the AE signal is taken out, this signal is compared with the reference value, and the damage of the bearing is predicted from the comparison result. However, since the reference value differs depending on the diagnosis target, it cannot be used for devices other than the diagnosis target. Further, in the case of the latter device, the nature of the abnormality is mainly classified, and it is difficult to accurately determine the abnormality determination level of the diagnosis target,
It is impossible to make a diagnosis using the optimum AE characteristic value for the diagnosis target. Therefore, in order to perform highly accurate diagnosis by the method applied to both devices, there is a problem that a diagnostic device must be developed for each device to be diagnosed.

An object of the present invention is to provide a method and an apparatus for automatically constructing an abnormality diagnosis algorithm capable of performing highly accurate diagnosis on various kinds of objects to be diagnosed.

[0006]

In order to achieve the above-mentioned object, the first method of the present invention relates to a plurality of diagnostic parameters for a diagnostic object having a normal property and an abnormal property. The value is measured, the measured value for each diagnostic parameter is subjected to statistical processing, and a plurality of diagnostic parameters predicted to be effective parameters are extracted from the diagnostic parameter group based on the processing result and extracted. Determine the diagnostic level for determining the quality of the property of the diagnostic target based on the measured values for the effective diagnostic parameters,
In the algorithm construction method for distinguishing between normal and abnormal products based on the diagnostic level, among the valid diagnostic parameters, when the average deviation amount of the normal products is below the reference value, the measured values of individual normal groups Is a virtual value having a predetermined value with respect to the average value of the normal group as a virtual measurement value of each normal group, and thereafter, the deviation amount, the standard deviation, the individual confidence degree, and the final overall confidence degree by the virtual measurement value of the normal group. The quality of the diagnosis is judged based on the diagnosis level determined according to the calculated value of.

In the second method, in the process of calculating the deviation amount on the basis of the valid diagnostic parameters, the deviation is calculated when the value of each of the normal product and the abnormal product is smaller than the average value of the normal group. Force the amount to zero.

The third method has both the first method and the second method.

In the first device of the present invention which achieves the above object, measuring means for measuring values relating to a plurality of diagnostic parameters with respect to a diagnostic group including normal and abnormal properties, Among the measured values of the means, an average value calculating means for calculating an average value of each diagnostic parameter obtained from the normal group, a first variance calculating means for calculating a variance for each normal parameter for each diagnostic parameter, and a measurement A second variance calculating means for calculating the variance for the abnormal group for each diagnostic parameter from the measured value of the means and the calculated value of the average value calculating means, and for each diagnostic from the ratio or difference of the calculated value of each variance calculating means A dispersion ratio calculating means for calculating the dispersion ratio of the parameters, and a parameter extracting means for extracting a plurality of effective diagnostic parameters from the diagnostic parameter group based on the calculated value of the dispersion ratio calculating means. Means, and the deviation amount for calculating the deviation amount between the measurement value and the average value for each diagnosis target based on the calculated value of the measurement value and the average value calculation means, for each effective diagnostic parameter extracted by the parameter extraction means Calculating means, calculating means for calculating the deviation amount average value and standard deviation of the normal diagnosis target group from the calculated values of the deviation amount calculating means, and the individual confidence factor for each valid diagnostic parameter from the calculated value of the calculating means. Individual confidence factor calculation means for calculating, final total confidence factor calculation means for calculating the final total confidence factor based on the calculated value of the individual confidence factor calculation means, and the property of the diagnosis target according to the calculated value of the final total confidence factor calculation means In the automatic algorithm construction device consisting of the diagnostic level determining means for determining the diagnostic level for judging the quality, if the deviation amount average value of the normal group falls below the reference value, the The deviation value calculation means for calculating the deviation amount average value based on the hypothetical measurement value of each normal group with a virtual value having a predetermined value as the average value of the normal group It is added to the means and the calculation means.

In the second device, when the value of each of the normal product and the abnormal product of the valid diagnostic parameters becomes smaller than the average value of the normal group, the deviation amount calculation means forces the deviation amount to zero. A deviation amount calculating means is provided.

The third device has both the first device and the second device.

[0012]

[Function] When the average deviation amount becomes equal to or less than the reference value, the individual confidence factor calculation means cannot determine the range of the individual confidence factor. Therefore, the measured value of each normal group is compared with the average value of the normal group. By using a virtual measurement value having a predetermined value and recalculating the deviation amount, it is possible to determine the range of the individual confidence factor.

Further, in the parameter which can judge the quality of the property by the level difference of the output voltage and the like, when the originally measured value level is extremely smaller than the normal product average value level,
Should be judged as normal. However, when the deviation amount is calculated by the deviation amount calculating means, the calculated value becomes large with respect to the data extremely close to the normal product average value, and as a result of the subsequent calculation, the individual certainty factor may be greatly calculated. Therefore, when the measured value of each normal group is smaller than the average value of the normal group, if the deviation amount is forced to be zero, the sample that should be a normal product can be properly evaluated.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a flowchart showing a processing procedure of an abnormality diagnosis algorithm automatic construction method according to an embodiment of the present invention, and FIG. 2 shows an abnormality diagnosis algorithm automatic construction apparatus for executing the abnormality diagnosis algorithm automatic construction method shown in FIG. It is a whole block diagram. In FIG. 2, the diagnosis target is, for example, a rotating body such as a steam turbine, a generator, a water turbine, a rolling mill, or a motor, and each diagnosis target 10 is classified into a normal property and an abnormal property. Then, each diagnostic target 10
The detector 12 detects values relating to various diagnostic parameters. The output signal of the detector 12 is sent to the computer 18 via the amplifier 14 and the A / D converter 16. The computer 18 includes a waveform analysis unit 20, a frequency analysis unit 22, and an arithmetic processing unit 24, and the arithmetic processing unit 24 is connected to a display unit 26 and an input / output unit 28. Computer 1
Reference numeral 8 constitutes a measuring means together with the detector 12, the amplifier 14, and the A / D converter 16, and the mean value calculating means, the first dispersion calculating means, the second dispersion calculating means, the dispersion ratio calculating means, the parameter extracting means, It comprises a deviation amount calculating means, a calculating means, an individual certainty factor calculating means, a final total certainty factor calculating means, and a diagnostic level determining means. And the computer 18
Is configured to statistically process the values relating to a plurality of diagnostic parameters based on the detection signal of the detector 12, and to determine the diagnostic level based on the processing result.

As diagnostic parameters, diagnostic parameters relating to waveform characteristics and diagnostic parameters relating to frequency characteristics are used. As shown in Table 1, the AE parameters relating to the waveform characteristics are as follows.

[0016]

[Table 1]

There are an average value, an average peak value, an energy value, etc. As the AE parameters relating to frequency characteristics,
As shown in Table 2,

[0018]

[Table 2]

There are a first peak frequency and its voltage, a rotation primary frequency and its voltage, a rotation secondary frequency and its voltage, and the like.

For example, to evaluate the detection waveform of the AE signal, the AE parameters relating to the waveform characteristics are measured by the method shown in Table 3 below based on the detection waveform shown in FIG.

[0021]

[Table 3]

The AE parameters relating to the frequency characteristics are determined by the method shown in Table 4 below based on the result of the frequency analysis (for example, FFT processing) of the AE detection waveform shown in FIG. 3 (one example is shown in FIG. 4). To measure by.

[0023]

[Table 4]

Before specifically describing the embodiments of the present invention, for comparison, an algorithm is constructed by a conventional statistical processing method and fuzzy processing method by the apparatus configuration shown in FIG. The method will be described.

FIG. 5 shows a diagnosis flowchart. First, the plurality of diagnostic objects 10 are divided into normal products and abnormal products, the number of normal products is n, and the number of abnormal products is m (step 100), and measurement of various AE parameters is performed on each diagnostic product 10. (Step 102). That is, the detector 12 is installed in each diagnosis target 10, the detection signal from each diagnosis target 10 is transferred to the computer 18, and the computer 18
The waveform analysis and the frequency analysis of the detection signal are executed in and the analysis result is processed in the arithmetic processing unit 24. That is, the measurement value of the AE parameter relating to the waveform characteristic and the measurement value of the AE parameter relating to the frequency characteristic are processed by the arithmetic processing unit 24. When performing this arithmetic processing, the following equation (1)
Equation (12) is used.

[0026]

[Equation 1]

[0027]

[Equation 2]

[0028]

[Equation 3]

[0029]

[Equation 4]

[0030]

[Equation 5]

[0031]

[Equation 6]

[0032]

[Equation 7]

[0033]

[Equation 8]

[0034]

[Equation 9]

[0035]

[Equation 10]

[0036]

[Equation 11]

[0037]

[Equation 12]

First, the average value of each AE parameter of a normal product is calculated according to the equation (1) (step 104).
Next, the variance of the normal product is calculated from the measured value of the normal product and the average value of the normal product according to the equation (2) (step 10).
6). Furthermore, from the measured values of abnormal products and the average value of normal products,
The variance of the abnormal product is calculated according to the equation (3) (step 108). Then, the dispersion ratio is calculated according to the equation (4) as the ratio or difference between the dispersion of the normal product and the dispersion of the abnormal product (step 110). These processes are performed for each AE parameter, and a plurality of effective parameters are extracted from the dispersion ratio of each AE parameter (step 112). For example, four AE parameters are selected in descending order of dispersion ratio.

FIG. 6 shows an example of extraction of effective parameters and an example of graphic display of the dispersion ratio of each AE parameter. After extracting, for example, four AE parameters as effective AE parameters, for each effective parameter, the deviation amount of the normal product and the abnormal product from the average value of the normal product (5),
It is calculated for each diagnosis target 10 according to the equation (6) (step 114). Then, based on the deviation amount from the average value of the normal product, the average deviation amount of the normal product and the standard deviation are calculated according to the equations (7) and (8). Then, the individual certainty factor of each effective AE parameter is determined based on these calculated values (step 116). In determining the individual certainty factor, in the present embodiment, the criterion for the individual certainty factor R is (13)
Set like an expression.

[0040]

[Equation 13]

FIG. 7 shows the deviation amount and the individual certainty factor for each of the normal and abnormal samples. The value within the value obtained by the equation (13) is set to "0.1", and its multiple is set to "0.2". Similarly, the individual certainty factor is determined by setting "0.3" and "0.4".

When calculating the final total confidence Ms,
As a fuzzy reasoning method, the final total confidence Ms is calculated according to the combine function method. For calculating the final total confidence Ms, the above formulas (9) to (12) are used. First, the first total confidence M1 is calculated using the individual confidence R1 of the first effective parameter. Ask according to. Next, using the individual confidence factor R2 of the second effective parameter, the second comprehensive confidence factor M2 is obtained according to the equation (10). Similarly, the third comprehensive certainty factor M3 is obtained by the equation (11), and the final comprehensive certainty factor Ms is obtained by the equation (12).

After the final total confidence Ms is calculated, FIG.
As shown in, the final comprehensive certainty factor Ms is plotted on the horizontal axis,
In the case of the separation type in which the frequency is plotted on the vertical axis and the normal product and the abnormal product are separated, the center of the maximum value of the normal product Ms and the minimum value of the abnormal product Ms is set as the determination level. Also,
In the case of the mixed type in which the normal product and the abnormal product overlap, the intersection of the normal product and the abnormal product is set as the diagnostic level, and the normal product and the abnormal product are determined according to this diagnostic level (step 120).
This makes it possible to construct a diagnostic algorithm.
The above is the outline of the conventional algorithm construction method.

Conventional algorithm construction method 5 shown in FIG.
When only 0 is used, if the measured values of the valid normal parameters of the valid diagnostic parameters are, for example, the same, naturally the average value of the normal products is also the same, and the value shown in equation (5) is calculated by the deviation amount calculating means. It becomes zero. Therefore, since the calculated values of the equations (7) and (8) become zero thereafter, the individual confidence factor calculation means cannot determine the range of the individual confidence factor. For example, the following case applies to the example. FIG. 9 shows the measured values of the valid diagnostic parameters that have been validated. First, when the average value of normal products is calculated in the equation (1), normal product sample No. 1-No. The measured value up to 5 is the same as "2.0". Next, when the normal product deviation amount is calculated by the equation (5), all of them are “0” and the deviation amount average value
The deviation standard deviation is also “0”. Furthermore, when the individual confidence factor range is calculated in the equation (13), it becomes "0". As a result, the individual certainty factor range as shown in FIG. 7 cannot be determined originally.

Next, a specific embodiment of the present invention will be described with reference to the flowchart shown in FIG. In the flowchart shown in FIG. 1, an algorithm 60 is added to the algorithm construction method 50 described in FIG. That is,
In this embodiment, in order not to perform the above-mentioned calculation,
In step 200, it is determined whether or not the average deviation amount is less than or equal to the reference value k, and if it is less than or equal to the reference value k, a virtual value having a predetermined value is set for each measured value of normal products. (Step 202). Then, the following (14), (1
The virtual deviation amount, the virtual deviation amount average value, and the virtual deviation amount standard deviation are calculated using the equations (5) and (16),

[0046]

[Equation 14]

[0047]

[Equation 15]

[0048]

[Equation 16]

Further, in the following equation (17)

[0050]

[Equation 17]

Determine the range of the virtual individual certainty factor (step 2
04) and calculate the final total confidence (step 118)
Then, the normal product and the abnormal product are determined according to the diagnostic level.
Note that the standard value k = 0 is generally set, but the value of k is arbitrarily changed depending on the situation.

A specific method will be described with reference to the data shown in FIG. When the reference value k = 0 is set, the deviation amount average value obtained from the measurement values of FIG. 9 is “0”, and therefore, as shown in FIG. 10, for example, a virtual measurement that is 1.2 times the normal product average value. The value "2.4" is set as each measured value. as a result,
Sample No. 1-No. 5, the virtual deviation amount average value is “0.2”, and the virtual deviation amount standard deviation is “0.
2 ", and (17) in the virtual individual confidence factor calculation means.
In the formula, for example, if L = 1/3, the individual confidence factor range can be calculated as “0.267”.

Next, an example in which the function of the abnormality diagnosis algorithm constructing means is further improved will be described. In the abnormal sound diagnosis of rotating equipment by the AE method, the parameters whose unit is voltage display, such as the W system parameters Vmean, Vpeak, Vmin and the F system parameters F1V, F2V, R1V, R2V, and FT shown in FIG. In many cases, a device having a low output voltage level is determined to be normal due to the AE phenomenon.

However, when the calculated values are calculated in the measured value data as shown in FIG. 11, the normal sample No. having a smaller voltage level than the normal product average value of the normal group. The individual certainty factor of 3 becomes as large as "0.3". In order to eliminate this contradiction, as shown in FIG. 12, effective diagnostic parameters are extracted and step 30
It is confirmed whether or not a parameter whose unit is voltage display is included in 0, and whether the individual measured values of the normal product and the abnormal product are smaller than the average value of the normal product (step 30).
2) Then, for the corresponding one, the deviation amount is forcibly set to zero in step 304.

By combining the abnormality diagnosis algorithm constructing means shown in FIGS. 1 and 12 to obtain the processing procedure shown in FIG. 13, it is possible to further improve the diagnosis accuracy.

[0056]

According to the present invention, the values relating to a plurality of diagnostic parameters are measured for the diagnostic groups divided into those with normal properties and those with abnormal properties, and the measured values are subjected to statistical processing. , A parameter for diagnosis that is predicted to be an effective parameter from this processing result is extracted, and the deviation amount calculation means,
The calculation means for calculating the average deviation amount and the standard deviation, the individual confidence factor calculation means, the final total confidence factor calculation means, etc. are calculated,
When determining the diagnostic level based on the measured value and constructing the algorithm, it is impossible to set the individual confidence factor range that occurs in effective parameters, and the individual confidence factor of the extremely normal measured value is greatly calculated. It is possible to prevent erroneous diagnosis due to inadequate calculation. Therefore, it is possible to construct an extremely accurate abnormality diagnosis algorithm.

[Brief description of drawings]

FIG. 1 is a flowchart showing a processing procedure according to an embodiment of the present invention.

FIG. 2 is an overall configuration diagram of an apparatus to which the present invention is applied.

FIG. 3 is an AE detection waveform diagram.

FIG. 4 is a frequency characteristic diagram of an AE detection waveform.

FIG. 5 is a supplementary diagram for explaining the present invention.

FIG. 6 is a diagram showing a graphic display example of calculated dispersion ratios and an extraction example of effective parameters.

FIG. 7 is a diagram showing an example of data of deviation amount and individual confidence factor of each sample of normal product and abnormal product.

FIG. 8 is a diagram for explaining a method of determining a diagnostic level.

FIG. 9 is a supplementary diagram for explaining the operation of the present invention.

FIG. 10 is a supplementary diagram for explaining the operation of the present invention.

FIG. 11 is a supplementary diagram for explaining the operation of the present invention.

FIG. 12 is a flow chart for explaining another operation of the present invention.

FIG. 13 is a flow chart for explaining another operation of the present invention.

[Explanation of symbols]

10 ... Diagnostic target, 12 ... Detector, 14 ... Amplifier, 16 ...
A / D converter, 18 ... Computer, 20 ... Waveform analysis section, 22 ... Frequency analysis section, 24 ... Arithmetic processing section, 26 ... Display section, 28 ... Input / output section.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor, Kaya Sato, 1168 Moriyama-cho, Hitachi, Hitachi, Ibaraki Energy Research Laboratory, Hitachi, Ltd. (72) Yuji Taguchi, 3-2-2, Saiwai-cho, Hitachi, Ibaraki Stock company Hitachi Engineering Service

Claims (10)

[Claims] 1. A plurality of diagnostic objects are divided into a normal group and an abnormal group, values relating to a plurality of diagnostic parameters are measured for each diagnostic object of each group, and normal values are measured among the measured values for each diagnostic parameter. The average value of the group was calculated, the variance of the normal group was calculated for each diagnostic parameter from the average value and the measured value, the variance of the abnormal group was calculated for each diagnostic parameter, and the variance was calculated for the normal group. Calculate the variance of each diagnostic parameter from the ratio or difference between the variance and the variance calculated for the anomaly group, and extract multiple effective diagnostic parameters from the diagnostic parameter group based on the calculated variance. , The deviation amount between the measured value for each diagnostic object and the average value for the normal group is calculated for the extracted diagnostic parameters, and the calculated deviation amount The average deviation amount and standard deviation amount of the normal group are calculated from, the individual confidence factor for each valid diagnostic parameter is calculated from these calculated values, and the final total confidence factor is calculated based on each individual confidence factor. , In the algorithm construction method for judging the quality of the property of the diagnosis object by the diagnosis level determined according to this calculated value, among the valid diagnostic parameters, the average deviation amount of the normal group is below the reference value. In this case, a virtual value having a predetermined value with respect to the average value of the normal group is set as a virtual measurement value of each normal group, and thereafter, a deviation amount, a standard deviation, an individual confidence factor, and a final total confidence by the virtual measurement value of the normal group. A method of automatically constructing an abnormality diagnosis algorithm, characterized in that the quality of diagnosis is judged from a diagnosis level determined according to a calculated value of the degree. 2. A plurality of diagnostic objects are divided into a normal group and an abnormal group, values for a plurality of diagnostic parameters are measured for each diagnostic object in each group, and normal values are measured among the measured values for each diagnostic parameter. The average value of the group was calculated, the variance of the normal group was calculated for each diagnostic parameter from the average value and the measured value, the variance of the abnormal group was calculated for each diagnostic parameter, and the variance was calculated for the normal group. Calculate the variance of each diagnostic parameter from the ratio or difference between the variance and the variance calculated for the anomaly group, and extract multiple effective diagnostic parameters from the diagnostic parameter group based on the calculated variance. , The deviation amount between the measured value for each diagnostic object and the average value for the normal group is calculated for the extracted diagnostic parameters, and the calculated deviation amount The average deviation amount and standard deviation amount of the normal group are calculated from, the individual confidence factor for each valid diagnostic parameter is calculated from these calculated values, and the final total confidence factor is calculated based on each individual confidence factor. In the algorithm construction method for determining the quality of the property of the diagnosis object by the diagnosis level determined according to the calculated value, in the process of calculating the deviation amount based on the valid diagnostic parameter, the normal product and the abnormal product A method for automatically constructing an abnormality diagnosis algorithm characterized by forcibly setting the deviation amount to zero when the individual value is smaller than the average value of the normal group. 3. A plurality of diagnostic objects are divided into a normal group and an abnormal group, values for a plurality of diagnostic parameters are measured for each diagnostic object in each group, and normal values are measured among the measured values for each diagnostic parameter. The average value of the group was calculated, the variance of the normal group was calculated for each diagnostic parameter from the average value and the measured value, the variance of the abnormal group was calculated for each diagnostic parameter, and the variance was calculated for the normal group. Calculate the variance of each diagnostic parameter from the ratio or difference between the variance and the variance calculated for the anomaly group, and extract multiple effective diagnostic parameters from the diagnostic parameter group based on the calculated variance. , The deviation amount between the measured value for each diagnostic object and the average value for the normal group is calculated for the extracted diagnostic parameters, and the calculated deviation amount The average deviation amount and standard deviation amount of the normal group are calculated from, the individual confidence factor for each valid diagnostic parameter is calculated from these calculated values, and the final total confidence factor is calculated based on each individual confidence factor. , In the algorithm construction method for judging the quality of the property of the diagnosis object by the diagnosis level determined according to this calculated value, among the valid diagnostic parameters, the average deviation amount of the normal group is below the reference value. In this case, a virtual value having a predetermined value with respect to the average value of the normal group is set as a virtual measurement value of each normal group, and thereafter, a deviation amount, a standard deviation, an individual confidence factor, and a final total confidence by the virtual measurement value of the normal group. Whether the quality of the diagnosis is good or bad is determined from the diagnostic level determined according to the calculated value of the degree, and the deviation amount is calculated based on the valid diagnostic parameters. In, for those individual values normal product and abnormal products is less than the average value of the normal group, the abnormality diagnostic algorithm automatic construction method, characterized by forced to zero deviation. 4. The abnormality diagnosis algorithm automatic construction method according to claim 1 or 3, wherein the reference value is changed according to the situation. 5. A measuring means for measuring values relating to a plurality of diagnostic parameters for a diagnostic group including normal and abnormal properties, and each of the measured values of the measuring means obtained from the normal group. An average value calculating means for calculating an average value of the diagnostic parameters, a first variance calculating means for calculating a variance for a normal group for each diagnostic parameter, a measured value of the measuring means and a calculated value of the average value calculating means. A second variance calculating means for calculating the variance of the abnormality group for each diagnostic parameter, and a variance calculating means for calculating the variance of each diagnostic parameter from the ratio or difference of the calculated values of each variance calculating means, A parameter extracting means for extracting a plurality of effective diagnostic parameters from the diagnostic parameter group based on the calculated value of the dispersion ratio calculating means, and a measured value and a calculated value of the average value calculating means are also included. And the deviation amount of the measured value and the average value for each diagnostic group, the deviation amount calculation means for calculating each valid diagnostic parameter extracted by the parameter extraction means, and the calculated value of the deviation amount calculation means of the normal group Computation means for calculating the deviation amount average value and standard deviation, individual confidence factor calculation means for calculating the individual confidence factor for each effective diagnostic parameter from the calculation value of the calculation means,
A final total confidence factor calculating means for calculating the final total confidence factor based on the calculated value of the individual confidence factor calculating means, and a diagnostic level for judging the quality of the property of the diagnosis object according to the calculated value of the final total confidence factor calculating means. When the average value of the deviation amount of the normal group becomes less than or equal to the reference value in the algorithm construction device including the determination level determining means for determining, the measured value of each normal group is a virtual value having a predetermined value with respect to the average value of the normal group. A device for automatically constructing an abnormality diagnosis algorithm, wherein a value is set as a virtual measurement value for each normal group, and means for calculating an average deviation amount based on the virtual measurement value is added to the deviation amount calculating means and the calculating means. 6. A measuring unit for measuring values relating to a plurality of diagnostic parameters for a diagnostic group including normal and abnormal properties, and each of the measured values of the measuring unit obtained from the normal group. An average value calculating means for calculating an average value of the diagnostic parameters, a first variance calculating means for calculating a variance for a normal group for each diagnostic parameter, a measured value of the measuring means and a calculated value of the average value calculating means. A second variance calculating means for calculating the variance of the abnormality group for each diagnostic parameter, and a variance ratio calculating means for computing the variance ratio of each diagnostic parameter from the ratio or difference of the calculated values of each variance calculating means, A parameter extracting means for extracting a plurality of effective diagnostic parameters from the diagnostic parameter group based on the calculated value of the dispersion ratio calculating means, and a measured value and a calculated value of the average value calculating means are also included. And the deviation amount of the measured value and the average value for each diagnostic group, the deviation amount calculation means for calculating each valid diagnostic parameter extracted by the parameter extraction means, and the calculated value of the deviation amount calculation means of the normal group Computation means for calculating the deviation amount average value and standard deviation, individual confidence factor calculation means for calculating the individual confidence factor for each effective diagnostic parameter from the calculation value of the calculation means,
A final total confidence factor calculating means for calculating the final total confidence factor based on the calculated value of the individual confidence factor calculating means, and a diagnostic level for judging the quality of the property of the diagnosis object according to the calculated value of the final total confidence factor calculating means. In the algorithm construction device comprising the determination level determination means for determining, in the process of calculating the deviation amount based on the valid diagnostic parameters, the value of each normal product and abnormal product is smaller than the average value of the normal group. With regard to the above, the abnormality diagnosis algorithm automatic construction apparatus is provided with a deviation amount calculation means for forcibly making the deviation amount zero. 7. A measuring unit for measuring values relating to a plurality of diagnostic parameters for a diagnostic group including normal and abnormal properties, and each of the measured values of the measuring unit obtained from the normal group. An average value calculating means for calculating an average value of the diagnostic parameters, a first variance calculating means for calculating a variance for a normal group for each diagnostic parameter, a measured value of the measuring means and a calculated value of the average value calculating means. A second variance calculating means for calculating the variance of the abnormality group for each diagnostic parameter, and a variance calculating means for calculating the variance of each diagnostic parameter from the ratio or difference of the calculated values of each variance calculating means, A parameter extracting means for extracting a plurality of effective diagnostic parameters from the diagnostic parameter group based on the calculated value of the dispersion ratio calculating means, and a measured value and a calculated value of the average value calculating means are also included. And the deviation amount of the measured value and the average value for each diagnostic group, the deviation amount calculation means for calculating each effective diagnostic parameter extracted by the parameter extraction means, and the calculated value of the deviation amount calculation means of the normal group Computation means for calculating the deviation amount average value and standard deviation, individual confidence factor calculation means for calculating the individual confidence factor for each effective diagnostic parameter from the calculation value of the calculation means,
A final total confidence factor calculating means for calculating the final total confidence factor based on the calculated value of the individual confidence factor calculating means, and a diagnostic level for judging the quality of the property of the diagnosis object according to the calculated value of the final total confidence factor calculating means. When the average value of the deviation amount of the normal group becomes less than or equal to the reference value in the algorithm construction device including the determination level determining means for determining, the measured value of each normal group is a virtual value having a predetermined value with respect to the average value of the normal group A value is set as a virtual measurement value for each normal group, and means for calculating an average deviation value based on the virtual measurement value is added to the deviation amount calculation means and the calculation means. In the process of calculating the deviation amount,
An apparatus for automatically constructing an abnormality diagnosis algorithm, comprising deviation amount calculation means for forcibly setting the deviation amount to zero when the value of each of the normal product and the abnormal product is smaller than the average value of the normal group. 8. The abnormality diagnostic algorithm automatic construction apparatus according to claim 5 or 7, further comprising means for changing the value of the reference value according to the situation. 9. An abnormality characterized by performing abnormality diagnosis based on data obtained from a diagnosis target according to an algorithm created by the abnormality diagnosis algorithm automatic construction method according to any one of claims 1 to 4. Diagnostic method. 10. A means for inputting data to be diagnosed,
An abnormality diagnosis algorithm automatic construction apparatus according to any one of claims 5 to 8, and means for diagnosing data to be diagnosed by an algorithm constructed by the abnormality diagnosis algorithm automatic construction apparatus. Abnormality diagnosis device.