Equipment deterioration diagnostic device and method

Abstract

To provide an equipment deterioration diagnosing apparatus and method capable of easily and accurately detecting deterioration of equipment in which operating conditions fluctuate greatly.SOLUTION: The equipment deterioration diagnosing apparatus is provided with an operation database for storing operation data measured in equipment, characteristic quantity calculating means for obtaining an operation condition characteristic quantity related to an operation condition and a deterioration characteristic quantity affecting the deterioration from the operation data, deterioration degree calculating means for calculating a deterioration degree by comparing a value of the deterioration characteristic quantity in the state of no deterioration with a current deterioration characteristic quantity under the condition that the operation condition characteristic quantity can be regarded as the same, and deterioration degree display means for displaying the deterioration degree obtained by the deterioration degree calculating means.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an equipment deterioration diagnosis apparatus and method.

Background Art

[0002] In recent years, from the viewpoints of the decrease in veteran maintenance staff and the reduction of maintenance costs, the need for improving the efficiency of equipment maintenance has been increasing.

[0003] As an abnormal diagnosis technology for equipment, there is an abnormal omen diagnosis technology shown in Patent Document 1. This technology includes "a classification unit that classifies measurement values obtained from a plant into a plurality of categories based on the similarity of the measurement values; a calculation unit that calculates a feature quantity that is the result of comparing the measurement values belonging to a premonition category other than the normal category to which the measurement values obtained during a period when the plant is known to be normal belong, with the measurement values belonging to the normal category; a prediction unit that predicts the calculated feature quantity in a future time series; and an estimation unit that estimates a future time when the plant becomes abnormal by applying a predetermined threshold value to the predicted feature quantity". That is, it is a technology that uses measurement values obtained from existing sensors and can diagnose deterioration without installing additional sensors.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, on the other hand, some equipment has characteristics such that the change in operating conditions is large and the fluctuation of the obtained operation data is also large. For such equipment, when trying to apply the technology of Patent Document 1, it may be difficult to detect deterioration because the change in operation data due to deterioration is small compared to the change in operation data due to the fluctuation of operating conditions.

[0006] Therefore, an object of the present invention is to provide an apparatus and method for diagnosing equipment deterioration that can easily and accurately detect the deterioration of equipment whose operating conditions vary greatly.

Means for Solving the Problems

[0007] From the above, in the present invention, “an operation data database for storing operation data measured by equipment, a feature quantity calculation means for obtaining an operation condition feature quantity related to an operation condition and a deterioration feature quantity affecting deterioration from the operation data, and under conditions where the operation condition feature quantities can be regarded as the same, a deterioration degree calculation means for calculating the deterioration degree by comparing the value of the deterioration feature quantity in a non-deteriorated state with the value of the current deterioration feature quantity, and a deterioration degree display means for displaying the deterioration degree obtained by the deterioration degree calculation means The system includes a principal component analysis means that performs principal component analysis on the operational condition features and degradation features obtained by the feature calculation means. The degradation degree calculation means uses the first and second principal components output by the principal component analysis means, and, under conditions where the first principal component can be considered identical, compares the value of the second principal component in a state without degradation with the current value of the second principal component to calculate the degradation degree. An apparatus for diagnosing equipment deterioration, characterized by the above.”

[0008] Also, in the present invention, “from the operation data measured by equipment, an operation condition feature quantity related to an operation condition and a deterioration feature quantity affecting deterioration are obtained, and under conditions where the operation condition feature quantities can be regarded as the same, the value of the deterioration feature quantity in a non-deteriorated state is compared with the value of the current deterioration feature quantity Principal component analysis is performed on the operational condition features and degradation features. For the first and second principal components obtained through principal component analysis, the value of the second principal component under conditions where the first principal component can be considered identical and there is no degradation is compared with the current value of the second principal component. A method for diagnosing equipment deterioration, characterized by calculating the deterioration degree.”

Effects of the Invention

[0009] According to the present invention, it is possible to easily and accurately detect the deterioration of equipment whose operating conditions vary greatly.

Brief Description of the Drawings

[0010] [Figure 1] A diagram showing a configuration example of an apparatus for diagnosing equipment deterioration according to Example 1 of the present invention. [Figure 2] A diagram showing an example of operation data. [Figure 3] A diagram showing the algorithm of the deterioration degree calculation means 4. [Figure 4]A diagram showing the outline of the processing in step S42 of the deterioration degree calculation means 4. [Figure 5] A diagram showing the outline of the processing in step S43 of the deterioration degree calculation means 4. [Figure 6] A diagram showing an example of the display of the deterioration degree display means. [Figure 7] A diagram showing a configuration example of the equipment deterioration diagnosis device according to Example 2 of the present invention. [Figure 8] A diagram showing an example of the result of the principal component analysis means. [Figure 9] A diagram showing an example of the deterioration degree display / prediction means. [Figure 10] A diagram showing a configuration example of the equipment deterioration diagnosis device according to Example 3 of the present invention. [Figure 11] A diagram showing an example of the result of the data classification means 8. [Figure 12] A diagram showing the outline of the processing of the deterioration degree calculation means 4C.

Mode for Carrying Out the Invention

[0011] Hereinafter, examples will be described with reference to the drawings.

Example

[0012] In this example, an example in which the deterioration diagnosis device of the present invention is applied to a boiler facility that supplies steam to a factory or the like will be described. The deterioration diagnosis device 10 of the present invention is configured as shown in FIG. 1, for example. The deterioration diagnosis device 10 in FIG. 1 includes an operation data database DB that stores various operation data from the boiler facility 1, a feature quantity calculation means 3, a deterioration degree calculation means 4, and a deterioration degree display means 5. The deterioration diagnosis device 10 is preferably configured using a computer device, but the parts other than the operation data database DB and the deterioration degree display means 5 are functions realized by processing in the arithmetic unit (CPU) of the computer device.

[0013] Among these, as the operation data D stored in the operation data database DB, data on fuel and air, which are input conditions of the boiler, and data on generated steam are stored as time-series data. In the example of FIG. 2 showing an example of the operation data D, the operation data D is time-series data with a period of 1 minute. The items being measured are the fuel flow rate F, the combustion air flow rate G, the steam flow rate Fs, the steam pressure Ps, the outside air temperature T, and the like.

[0014] Next, the feature quantity calculation means 3 calculates a feature quantity D1 related to the operation condition and a feature quantity D2 related to deterioration from the operation data D stored in the operation data database DB. Here, the operation condition is measurement data (operation data) that is not affected by the deterioration of the equipment, such as the fuel flow rate F, the combustion air flow rate G, and the outside air temperature T, and the feature quantity D1 related to the operation condition is obtained from these. Further, the deterioration feature quantity D2 is a feature quantity using measurement data (operation data) whose value changes even under the same operation condition due to the deterioration of the equipment. For example, the steam flow rate F s 、steam pressure force P s and the like are applicable, and the feature quantity D2 is obtained from these. Note that the deterioration feature quantity D2 does not necessarily need to use only measurement data whose value changes due to deterioration, and a feature quantity obtained by dividing measurement data whose value changes due to deterioration by measurement data related to the operation condition may also be used.

[0015] Here, the feature quantities D1 and D2 may be the measurement data itself, processed data or combined data of the measurement data, or data related to components obtained from the measurement data. Here, these are collectively referred to as feature quantities.

[0016] In this embodiment, the operation condition feature quantity D1 is obtained by the formula (1) as a function of the fuel flow rate F and the outside air temperature T. (Equation 1) D1 = F + K × T (1) Here, K is a coefficient for correcting the operation condition according to the outside air temperature T. Also, the deterioration feature quantity D2 was set as the steam flow rate Fs. This is because in this boiler facility, the steam pressure Ps is controlled to be constant.

[0017] In the deterioration degree calculation means 4 of FIG. 1, using the two feature quantities D1 and D2 obtained by the feature quantity calculation means 3, the deterioration degree of the facility corrected by the operation conditions is calculated.

[0018] The specific algorithm will be described using the processing flow of FIG. 3. In this embodiment, the case of diagnosing the deterioration of the boiler facility using the current measurement data is taken as an example.

[0019] In the processing flow of FIG. 3, first, in processing step S41, the value of the operation condition feature quantity D1 at the time to be diagnosed is calculated. In this embodiment, from the operation data D stored in the database DB of FIG. 2, the values of the fuel flow rate F and the outside air temperature T are acquired from the operation data D stored in the database DB at the current time (for example, 2020 / 3 / 3 10:03:00), and the value of D1 is calculated.

[0020] In processing step S42, the deterioration feature quantity under the operation conditions obtained in processing step S41 is obtained from the operation data at the time when there is no deterioration. The specific method will be described using FIG. 4. The horizontal axis of FIG. 4 is the operation condition feature quantity D1, the vertical axis is the deterioration feature quantity D2, and the operation data when there is no deterioration is plotted. If the current operation condition is D1c, the data included in the interval from (D1c - ΔD1 / 2) to (D1c + ΔD1 / 2) in the vicinity thereof is acquired, and the average value of the deterioration feature quantity D2 at that time is taken as the value of the deterioration feature quantity in the normal state. Note that the time when there is no deterioration is the data within a certain period from the start of operation.

[0021] In processing step S43, the deterioration feature quantity D1 at the current time is calculated. The data at the current time may be a single point of data at the current time, or may be an average value using a plurality of data under the same operation conditions. For example, in the example of FIG. 5 showing the time on the horizontal axis and the magnitudes of the operation condition feature quantity D1 and the deterioration feature quantity D2 on the vertical axis, the average value of the data in section 1 close to the current operation condition may be used, or the average value of the data including the data in section 2 may also be used.

[0022] In processing step S44, the degree of deterioration G is obtained from the deterioration feature amount D1ave at the time when there is no deterioration obtained in processing step S42 and the current feature amount D1c by the following formula. (Equation 2) G = (D2ave - D2c) / D2ave (if G < 0, then G = 0) (2) Since there is variation in the data, the degree of deterioration may take a negative value. However, in this embodiment, when G < 0, G is defined as 0. Through the above steps, the degree of deterioration G can be calculated.

[0023] Returning to FIG. 1, in the degree-of-deterioration display means 5, the degree of deterioration G obtained by the degree-of-deterioration calculation means 4 is displayed on the display. An example of the display screen is shown in FIG. 6. The horizontal axis in FIG. 6 is time, and the vertical axis is the degree of deterioration G. In this embodiment, the historical data when the degree of deterioration was calculated in the past is included in the display. In this way, by displaying not only the current degree of deterioration but also the past degree of deterioration, the state of change of the degree of deterioration can be confirmed.

[0024] As described above, by using the feature amount related to the operating conditions and the feature amount related to the degree of deterioration from the operation data to obtain the degree of deterioration G, comparison can be made under the same operating conditions. Therefore, even when the operating conditions vary greatly, the deterioration of the equipment can be accurately detected.

[0025] According to Example 1, the deterioration of the boiler equipment with greatly varying operating conditions can be simply and accurately detected.

Example

[0026] Next, the equipment deterioration diagnosis apparatus according to Example 2 of the present invention will be described with reference to FIG. 7. The differences between Example 2 and Example 1 are that the principal component analysis means 6 is added and the degree-of-deterioration display means 5 is changed to the degree-of-deterioration display and prediction means 7. Note that with the addition of the principal component analysis 6, the subsequent degree-of-deterioration calculation means 4B also changes. The differences from Example 1 will be described below.

[0027] The deterioration diagnostic device 10 in Figure 7 is preferably configured using a computer, but all parts other than the operation data database DB and the deterioration degree display means 5 are functions that are realized by processing in the calculation unit (CPU) of the computer.

[0028] The principal component analysis (PCA) means 6 uses PCA to organize the feature quantities D1 and D2 obtained by the feature calculation means 3 into axes corresponding to the operating conditions and the degree of degradation. In PCA, the direction in which the data variance is maximized is defined as the first principal component, and the axis orthogonal to the first principal component and in which the variance is maximized is defined as the second principal component. Therefore, when PCA is performed on the data shown in Figure 4, as shown in Figure 8, which is an example of the results of the PCA, the first principal component (horizontal axis) corresponds to the operating conditions, and the second principal component (vertical axis) represents the components that cannot be explained by the variability due to the operating conditions.

[0029] In the degradation degree calculation means 4B, the degradation degree is calculated using the values ​​of the first principal component and the second principal component. The method for calculating the degradation degree is the same as in Example 1, but while Example 1 used the operational condition feature D1 and the degradation feature D2, Example 2 uses the first principal component score PCA1 and the second principal component score PCA2 obtained from the operational condition feature D1 and the degradation feature D2. Specifically, the degradation degree is calculated using equation (3). (Math 3) G = PCA2ave - PCA2c (If G < 0, then G = 0) (3) Here, PCA2ave is the PCA2 value at the point when there was no degradation, and PCA2c is the current PCA2 value.

[0030] The degradation level display and prediction means 7 displays the current degradation level and predicts the future degradation level, and displays the predicted degradation level. Figure 9 shows an example of the output of the degradation level display and prediction means 7, with time on the horizontal axis and degradation level G on the vertical axis. The difference from Figure 6 shown in Example 1 is that it predicts and displays the future degradation level. The predicted value was obtained using a linear regression equation with data from the present to a certain period enclosed by circles.

[0031] Thus, according to Example 2, by more clearly separating the operating conditions as the first principal component, equipment deterioration can be detected with greater accuracy. Furthermore, by displaying the predicted degree of deterioration, this information can be used to determine the timing of future maintenance.

[0032] In this embodiment, a linear regression equation was used as the regression equation, but a nonlinear regression equation may also be used to predict the future degree of degradation. [Examples]

[0033] Finally, an example of the configuration of a boiler equipment deterioration diagnosis device according to Embodiment 3 of the present invention will be described with reference to Figure 10. The difference between Embodiment 3 and Embodiment 2 is that a data classification means 8 is added after the principal component analysis means 6. In addition, the calculation method of the deterioration degree calculation means 4C has also been changed. The data classification means 8 and the deterioration degree calculation means 4C will be described below.

[0034] The deterioration diagnostic device 10 in Figure 10 is preferably configured using a computer, but all parts other than the operation data database DB and the deterioration degree display means 5 are functions that are realized by processing in the calculation unit (CPU) of the computer.

[0035] The method used in data classification means 8 is Adaptive Resonance Theory (ART), as described in Patent Document 2. Adaptive Resonance Theory classifies multidimensional data into multiple categories according to their similarity.

[0036] Figure 11 shows an example of the results of classifying the principal component scores obtained by the principal component analysis means 6 using the data classification means 8. In this embodiment, the data is classified into six categories from C1 to C6. The categories are arranged in the direction of the first principal component, and there is no overlap in the direction of the second principal component. This is because the data has large fluctuations in the direction of the first principal component, and the classification is achieved by adjusting the parameter that determines the size of the categories.

[0037] In the degradation degree calculation method 4C, the degradation degree is determined for each data item classified into each category. The specific method of calculation is explained using Figure 12. The left side of Figure 12 is a schematic diagram showing data classified into a certain category. The gray circles represent data, and the diamonds represent the centroid of the data classified into this category. The degradation degree is calculated using ΔPCA2, which is the deviation in the direction of the second principal component from the centroid, as shown on the right side of Figure 12. The value of ΔPCA2 that serves as the standard for a state without degradation is determined for each category as the average value of ΔPCA2 of the data included in that category, ΔPCA2ave. Since ΔPCA2 is calculated based on the centroid of each category, the average value ΔPCA2ave will be close to zero.

[0038] Next, the current ΔPCA2c is calculated based on the centroid of the category closest to the current data, and the degree of degradation G is calculated using equation (4). (Math 4) G = ΔPCA2ave - ΔPCA2c (If G < 0, then G = 0) (4) Thus, according to this embodiment, since the operating data can be classified into multiple categories, the current operating conditions can be intuitively grasped from the currently classified categories, making it easier to understand the relationship between operating conditions and the degree of deterioration. [Explanation of Symbols]

[0039] 1: Boiler equipment 3: Feature calculation method 4, 4B, 4C: Deterioration degree calculation means 5: Deterioration degree display means 6: Principal component analysis means 7: Degradation level display and prediction methods 8: Data classification methods D: Driving data DB: Operation Data Database

Claims

1. The system comprises: an operating data database for storing operating data measured by the equipment; a feature calculation means for obtaining operating condition features related to operating conditions and degradation features that affect degradation from the operating data; a degradation degree calculation means for calculating the degree of degradation by comparing the values ​​of the degradation features in a no-degradation state and the current degradation features under conditions where the operating condition features can be considered the same; a degradation degree display means for displaying the degree of degradation obtained by the degradation degree calculation means; and a principal component analysis means for performing principal component analysis on the operating condition features and degradation features obtained by the feature calculation means. The equipment deterioration diagnostic device is characterized in that the deterioration degree calculation means uses the first principal component and the second principal component output by the principal component analysis means, and, under conditions where the first principal component can be considered identical, compares the value of the second principal component in a state without deterioration with the current value of the second principal component to calculate the degree of deterioration.

2. A device for diagnosing equipment deterioration according to claim 1, The equipment deterioration diagnostic device is characterized in that the deterioration degree display means predicts and displays the future deterioration degree from the time-series change of the obtained deterioration degree.

3. A device for diagnosing equipment deterioration according to claim 1, The system includes a data classification means that classifies the first and second principal components output by the principal component analysis means using adaptive resonance theory and categorizes them into multiple categories. The equipment deterioration diagnostic device is characterized in that the deterioration degree calculation means determines the deviation in the direction of the second principal component of the centroid of the category closest to each of the classified data, compares the deviation value of a no-deterioration state with the current deviation value, and calculates the degree of deterioration.

4. A method for diagnosing deterioration of equipment, characterized by obtaining operational condition features related to operating conditions and deterioration features that affect deterioration from operating data measured by the equipment, comparing the value of the deterioration features in a state without deterioration with the value of the current deterioration features under conditions where the operational condition features can be considered the same, performing principal component analysis on the operational condition features and the deterioration features, and comparing the value of the second principal component in a state without deterioration with the value of the current second principal component under conditions where the first principal component can be considered the same for the first principal component and the second principal component obtained in the principal component analysis, thereby calculating the degree of deterioration.

5. A method for diagnosing equipment deterioration according to Claim 4, A method for diagnosing equipment deterioration, characterized by predicting and displaying the future degree of deterioration based on the time-series changes in the determined degree of deterioration.

6. A method for diagnosing equipment deterioration according to claim 4, The first principal component and the second principal component are classified according to adaptive resonance theory and then classified into multiple categories. A method for diagnosing equipment deterioration, characterized by determining the deviation in the direction of the second principal component of the centroid of the category closest to each of the classified data, comparing the deviation value of a state without deterioration with the current deviation value, and calculating the degree of deterioration.

Images