JP2014026327A - Device state diagnostic apparatus using actual operation data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve statistical reliability of a reference model and early know a symptom of abnormality in a device to be diagnosed.SOLUTION: The device state diagnostic apparatus generates a reference model (an industry model (a change characteristic I)) on the basis of actual operation data collected from multiple devices that are of the same type as a specified device to be diagnosed and are used in the same industry. The apparatus generates a diagnostic object model (a change characteristic V) on the basis of the actual operation data collected from the specified device to be diagnosed. The apparatus compares the diagnostic object model (the change characteristic V) with an acceptable range model (change characteristics II and III) to the industry model (the change characteristic I), and diagnoses a state of the device to be diagnosed, on the basis of a result of the comparison. Specifically, the apparatus obtains a model coefficient AX1 of the diagnostic object model, compares it with an acceptable range ±α to a coefficient AG1 in a model formula for the industry model G1, and determines that the device to be diagnosed is in a special state when the AX1 is outside the acceptable range ±α.

Description

  The present invention relates to a device state diagnosis apparatus based on actual operation data for diagnosing the state of a device designated as a diagnosis target device based on actual operation data from the device.

  2. Description of the Related Art Conventionally, there is a device for diagnosing the state of a fuel cell power generator used in a factory or a building as a device state diagnosing device based on this type of actual operation data.

For example, in the apparatus for diagnosing the state of the fuel cell power generator disclosed in Patent Document 1, the initial voltage of the fuel cell that generates power from the fuel gas and the oxidant gas is E0, and the integrated power generation time is T. In this case, the initial voltage E0 and the coefficient A of the model equation are calculated from the cell voltage E (Tn) measured at a plurality of accumulated power generation times Tn using the following model equation (Equation (1)). When the coefficient A changes beyond a predetermined range, it is determined that the apparatus is not in an appropriate state.
E (Tn) = E0−Aln (Tn) (1)

  For example, using the change characteristic of the cell voltage shown in FIG. 6, the voltage is measured as E (Tn) = 0.75 V at the point t1 where the integrated power generation time is Tn = 1000 hours. At the point t2 where the integrated power generation time is Tn = 2000 hours, the voltage is measured as E (Tn) = 0.745V. When these numerical values are substituted into the above equation (1) and calculated, the initial voltage E0 is calculated to be 0.7998 V, and the coefficient A of the model equation is calculated to be 0.0072.

  Here, if the voltage at the point t3 when the integrated power generation time is Tn = 3000 hours is E (Tn) = 0.742 V, the numerical value when Tn = 1000 hours and the numerical value when Tn = 3000 hours are used. The initial voltage E0 calculated in this way is 0.8003V, and the coefficient A of the model formula is 0.0073. In this case, the coefficient A of the model formula is almost the same as the value calculated when Tn = 2000 hours, and it is determined that the apparatus is in an appropriate state (normal state).

  On the other hand, if the voltage at the point t3 when the integrated power generation time is Tn = 3000 hours is E (Tn) = 0.74V, the numerical value when Tn = 1000 hours and the numerical value when Tn = 3000 hours The initial voltage E0 calculated using the above is 0.8118V, and the coefficient A of the model formula is 0.0089. In this case, the coefficient A of the model formula is greatly changed with respect to the value calculated when Tn = 2000 hours, and it is determined that the apparatus is not in an appropriate state (in a special state).

  Further, as another example of a device state diagnosis apparatus based on actual operation data, for example, Patent Document 2 discloses an abnormality monitoring apparatus for a plurality of apparatuses. In the abnormality monitoring device disclosed in Patent Document 2, a plurality of similar (K + 1) devices (devices of the same type) are processed, and each of the devices obtained by measuring the state of each device with a plurality of sensors is processed. Based on the plurality of data items, an abnormality in the state of at least one of the plurality (K + 1) devices is monitored. Among a plurality of (K + 1) devices, one first device is to be monitored, and other similar second devices generate a first model for monitoring the first device. Based on a plurality of (K) models dedicated to individual second devices created based on a plurality of normal data items in each of the plurality (K) of second devices in the first step, A first model dedicated to the first device is generated, and in the second step, a plurality of data items from the first device are input in a predetermined time unit, and the first model is used to Monitor device status abnormalities.

  In this abnormality monitoring apparatus, the first step is a step of statistically classifying a plurality of data items of each of the plurality of (K) apparatuses into an objective variable in regression analysis and one or more other explanatory variables; A step of creating a plurality of (K) regression models as individual models of each of the plurality (K) of second devices, and coefficients and intercepts of the plurality of (K) regression models from the feature item values of each of the devices. A step of creating a meta prediction model common to similar devices to be predicted, and a feature item value of the first device is input to the meta prediction model to generate coefficients and intercepts of a regression model as a first model dedicated to the device A step of performing.

  The second step includes a step of inputting an explanatory variable in a plurality of data items of the first device to a first model dedicated to the first device to calculate a predicted value of the target variable, an actual value of the target variable, A step of calculating a deviation degree between the predicted value and a step of detecting an abnormality of the first device by comparing the deviation degree and a threshold value.

JP 2008-98076 A JP 2010-287011 A

  In the device disclosed in Patent Document 1, the fuel cell power generation device is a diagnosis target device. However, it is possible to diagnose the state of a heat source device or the like using a similar diagnosis method. In this case as well, as with the fuel cell power generation device, irreversible damage is caused by determining whether or not the diagnosis target device is in a special state from the change in the coefficient of the model formula indicating the change characteristic of the device state. It is possible to detect an abnormality of the device at an early stage. That is, it is possible to know early signs of device abnormality.

  In the diagnostic method disclosed in Patent Document 1, a reference model indicating a change characteristic of a basic state of a device is generated based on experimental data, and an allowable range is determined for a coefficient of a model formula of the reference model. Whether or not the device to be diagnosed is in a special state is determined based on whether or not the coefficient of the model formula obtained from the actual operation data is within an allowable range determined with respect to the coefficient of the model formula of the reference model It is considered that

  However, large-scale and low-volume production equipment, such as fuel cell power generation equipment and heat source equipment used in factories and buildings, is difficult to prepare an experimental environment with conditions close to the practical environment. It is also difficult to prepare abundant experimental samples themselves. For this reason, abundant experimental data cannot be collected and the statistical reliability of the reference model cannot be increased.

  The diagnostic method disclosed in Patent Document 1 can also be applied to field devices such as valves and transmitters, but such field devices are originally used for a long period (about 15 years). Therefore, no clear signs of abnormality due to deterioration or failure appear in a short period of time. For this reason, even when the device to be diagnosed is a field device, it is difficult to collect abundant experimental data in a short period of time using limited experimental samples, and the statistical reliability of the reference model Can not increase.

  Further, in the device disclosed in Patent Document 2, one of the plurality (K + 1) devices is set as a monitoring target (diagnosis target device), and a plurality of other second devices similar to the first device are used. The first model is generated from normal data from the target as a target for generating a first model for monitoring the first device. That is, the reference model is generated from normal data from a plurality of other devices similar to the diagnosis target device. This increases the number of data when generating the reference model and increases the statistical reliability of the reference model.

  However, in the apparatus disclosed in Patent Document 2, explanatory variables in a plurality of data items of a diagnosis target device are input to a reference model to calculate a predicted value of an objective variable, and the calculated predicted value and measured value are calculated. The degree of divergence is calculated, and an abnormality of the diagnosis target device is detected by comparing the calculated degree of divergence with a threshold value. For this reason, it is possible to detect the presence or absence of abnormality at the stage of obtaining the actual measurement value, but it is not possible to know the sign of abnormality early.

  Moreover, in the apparatus shown in Patent Document 2, a reference model is generated from data from a plurality of other devices similar to the diagnosis target device. However, a plurality of other devices similar to the diagnosis target device are used. Usage conditions are not considered. In other words, even if the device is similar to the device to be diagnosed, the operation pattern of the device differs depending on the industrial system in which the device is placed and the industry in which the facility is operated. In spite of the influence, since the reference model is generated using data from all the other similar devices, the statistical reliability of the reference model is lowered, and the accuracy of diagnosis is lowered.

  The present invention has been made to solve such a problem, and the object of the present invention is to improve the statistical reliability of the reference model and to know early signs of abnormality of the diagnostic target device. Another object of the present invention is to provide a device state diagnosis apparatus based on actual operation data.

  In order to achieve such an object, the present invention provides a device state diagnosis apparatus based on actual operation data for diagnosing the state of a device designated as a diagnosis target device based on actual operation data from the device. The actual data collected from multiple devices, and the device to be diagnosed based on the actual data collected from multiple devices of the same type and used in the same industry as the device to be diagnosed. A reference model generation unit that generates a reference model that indicates a change characteristic of an overall state of a device group of the same type and the same industry, and a range that is permitted as a normal state with respect to the reference model generated by the reference model generation unit Tolerance model generation means for generating the indicated tolerance model and the diagnosis target device based on the actual operation data collected from the specified diagnosis target device A diagnostic target model generating means for generating a diagnostic target model indicating a change characteristic of the current state, a diagnostic target model generated by the diagnostic target model generating means, and an allowable range model for the reference model generated by the allowable range model generating means; And device state diagnosis means for diagnosing the state of the device to be diagnosed based on the comparison result.

  According to the present invention, a reference model is generated based on actual operation data collected from a plurality of devices of the same type and used in the same industry as the designated diagnosis target device, and an allowable range for the reference model. A model is generated. In addition, a diagnosis target model is generated based on actual operation data collected from the designated diagnosis target device. Then, the generated diagnosis target model is compared with an allowable range model for the reference model, and the state of the diagnosis target device is diagnosed based on the comparison result.

  For example, in the present invention, the reference model generation means generates a coefficient of the model expression of the reference model, the allowable range model generation means generates an allowable range for the coefficient of the model expression of the reference model, and the diagnosis target model generation means The model condition coefficient of the diagnosis target model is generated, and the device state diagnosis means generates the model expression coefficient of the diagnosis target model generated by the diagnosis target model generation means and the model of the reference model generated by the allowable range model generation means. The allowable range for the coefficient of the equation is compared, and the state of the diagnosis target device is diagnosed based on the comparison result.

According to the present invention, since the reference model is generated based on the actual operation data collected from a plurality of devices of the same type and used in the same industry as the diagnosis target device, the operation pattern is similar from the operation data. A reference model is generated, and the statistical reliability of the reference model is increased.
Further, according to the present invention, by comparing the diagnosis target model with the tolerance model for the reference model, if the coefficient of the model expression of the diagnosis object model is out of the tolerance for the coefficient of the model equation of the reference model, a special It is possible to know at an early stage a sign of abnormality of the diagnosis target device, for example, by determining that it is in a proper state.

It is a figure which shows the principal part of one Embodiment of the diagnostic system using the apparatus state diagnostic apparatus by the actual operation data which concerns on this invention. It is a functional block diagram of a device state diagnosis device (device state diagnosis device) based on actual operation data in this diagnosis system. It is a figure which illustrates the industry model G1 (change characteristic I) produced | generated from the actual operation data of apparatus X2, X3, and the tolerance | permissible_range model (change characteristics II, III) with respect to this industry model G1. It is a figure which illustrates the diagnostic object model (change characteristic IV) at the time of the integrated electric power generation time Tn of the diagnostic object apparatus X1 having reached 700 hours. It is a figure which illustrates the diagnostic object model (change characteristic V) at the time of the integrated electric power generation time Tn of the diagnostic object apparatus X1 reaching 1500 hours. It is a figure explaining the diagnostic method of the state of the fuel cell electric power generating apparatus in patent document 1. FIG.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing a main part of an embodiment of a diagnosis system using an apparatus state diagnosis apparatus based on actual operation data according to the present invention.

  In FIG. 1, reference numeral 100 denotes a device state diagnosis device (hereinafter referred to as a device state diagnosis device) based on actual operation data according to the present invention, and 200 denotes a device diagnosed by the device state diagnosis device 100. The device 200 is connected to the device state diagnosis apparatus 100 via the network 300.

  In FIG. 1, the device 200 is assumed to be the same type of device. In this embodiment, it is assumed that the device 200 is a fuel cell power generator and the model numbers are all the same. 1 shows only the devices X1 to X6 as the device 200, but there are actually more devices X.

  In this embodiment, for the sake of simplicity, the industry G1 is the semiconductor industry, the industry G2 is the commercial building industry, the devices X1, X2, and X3 are used in the industry G1, and the devices X4 and X5 are used in the industry G2. , X6 are used. In this example, the semiconductor industry is cited as the industry G1, and the commercial building industry is cited as the industry G2. However, there are many more industries G such as the public equipment industry.

  The device state diagnosis apparatus 100 is realized by hardware including a processor and a storage device, and a program that realizes various functions in cooperation with the hardware. The program is provided in a state where it is recorded on a recording medium such as a CD-ROM, and is read from the recording medium and installed in a storage device such as a hard disk.

  FIG. 2 shows a functional block diagram of the device state diagnosis apparatus 100. The device state diagnosis apparatus 100 includes an actual operation data collection unit 1, a reference model generation unit 2, an allowable range model generation unit 3, a diagnosis target model generation unit 4, a device state diagnosis unit 5, and a database unit 6. I have.

  In the device state diagnosis apparatus 100, the actual operation data collection unit 1 periodically collects actual operation data from the devices 200 that are actually in operation via the network 300 and stores them in the database unit 6. In this example, the accumulated power generation time Tn and the cell voltage E (Tn) are collected as the actual operation data of the device 200, and are classified by industry and stored in the database unit 6.

  The reference model generation unit 2 accesses the database unit 6 and, based on actual operation data collected from a plurality of devices 200 of the same type and used in the same industry as the specified diagnosis target device, Generate a reference model that shows the overall state change characteristics of devices of the same type and industry.

The allowable range model generation unit 3 generates an allowable range model indicating a range that is allowed as a normal state with respect to the reference model generated by the reference model generation unit 2.
The diagnosis target model generation unit 4 accesses the database unit 6 and generates a diagnosis target model indicating the change characteristics of the current state of the diagnosis target device based on the actual operation data collected from the designated diagnosis target device. .

  The device state diagnosis unit 5 receives the specification of the diagnosis target device from the operator, transmits the specified diagnosis target device to the reference model generation unit 2 and the diagnosis target model generation unit 4, and generates the diagnosis target model generation unit 4. The diagnosis target model is compared with an allowable range model for the reference model generated by the allowable range model generation unit 3, and the state of the diagnosis target device is diagnosed based on the comparison result.

  Next, detailed functions of the reference model generation unit 2, the allowable range model generation unit 3, the diagnosis target model generation unit 4, and the device state diagnosis unit 5 will be described with specific examples.

  For example, when the device condition diagnosis unit 5 receives a designation from the operator with the diagnosis target device as X1, the device state diagnosis unit 5 transmits the designated diagnosis target device X1 to the reference model generation unit 2 and the diagnosis target model generation unit 4.

  When receiving the designation of the diagnosis target device X1 from the device state diagnosis unit 5, the reference model generation unit 2 accesses the database unit 6 and other devices 200 excluding the diagnosis target device X1 of the industry G1 to which the diagnosis target device X1 belongs. Get actual production data. In this example, actual operation data of the devices X2 and X3 is acquired.

  Then, the reference model generation unit 2 generates a reference model indicating the change characteristics of the overall state of the device group of the same type and the same industry as the diagnosis target device X1 from the acquired actual operation data of the devices X2 and X3. In this example, a model formula represented by the above-described formula (1) is created from a plurality of accumulated power generation times Tn and cell voltages E (Tn) acquired as actual operation data of the devices X2 and X3, and the created model formula Is the model equation of the reference model, and the coefficient A of this model equation is obtained as AG1. Hereinafter, the reference model generated from the actual operation data of the devices X2 and X3 is referred to as an industry model G1.

  FIG. 3 illustrates an industry model G1 (change characteristics I) generated from actual operation data of the devices X2 and X3. The actual operation data of the device X2 is accumulated for a relatively short period of time, and the actual operation data of the device X3 is accumulated for a relatively long period of time. The devices X2 and X3 belong to the same industry G1 as the device X1, and the actual operation data has similar operation patterns. By using this similar operational data as the original data of the reference model and performing statistical processing, the effects of lack of accumulation period and data variation are improved, the statistical reliability of the reference model is increased, and accurate diagnosis is performed. Can be done.

  The coefficient AG1 of the model formula of the industry model G1 generated by the reference model generation unit 2 is sent to the allowable range model generation unit 3. Based on the coefficient AG1 of the model formula of the industry model G1 from the reference model generation unit 2, the allowable range model generation unit 3 generates an allowable range model indicating an allowable range allowed as a normal state with respect to the industry model G1. In this example, an allowable range ± α is determined for the coefficient AG1 of the model formula of the industry model G1. FIG. 3 illustrates an allowable range model (change characteristics II and III) generated for the industry model G1. The coefficient AG1 and the allowable range ± α of the model formula of this industry model G1 are sent to the device state diagnosis unit 5.

  When receiving the designation of the diagnosis target device X1 from the device state diagnosis unit 5, the diagnosis target model generation unit 4 accesses the database unit 6, acquires the actual operation data of the diagnosis target device X1, and acquires the acquired diagnosis target device. A diagnosis target model indicating a change characteristic of the current state of the diagnosis target device X1 is generated from the actual operation data of X1. In this example, a model formula shown by the above-mentioned formula (1) is created from a plurality of accumulated power generation times Tn and cell voltage E (Tn) acquired as actual operation data of the device X1, and the created model formula is diagnosed. The model formula of the target model is used, and the coefficient A of this model formula is obtained as AX1. The coefficient AX1 of the model formula of this diagnosis target model is sent to the device state diagnosis unit 5.

  The device state diagnosis unit 5 receives the coefficient AX1 of the model formula of the diagnosis target model from the diagnosis target model generation unit 4, the coefficient AG1 of the model formula of the industry model G1 from the allowable range model generation unit 3 and the allowable range ± α. The coefficient AX1 of the model formula of the diagnosis target model is compared with the allowable range ± α for the coefficient AG1 of the model formula of the industry model G1.

  Here, if the coefficient AX1 of the model formula of the diagnosis target model is within an allowable range ± α with respect to the coefficient AG1 of the model formula of the industry model G1, the device state diagnosis unit 5 is in a normal state of the diagnosis target device X1. And the determination result (diagnosis result) is output. On the other hand, if the coefficient AX1 of the model formula of the diagnosis target model is out of the allowable range ± α with respect to the coefficient AG1 of the model formula of the industry model G1, the device state diagnosis unit 5 determines that the diagnosis target device X1 is not in a normal state. Is judged to be in a special state (deterioration, damage, etc.), and the judgment result (diagnosis result) is output.

  FIG. 4 shows a diagnosis target model (change characteristic IV) when the accumulated power generation time Tn of the diagnosis target device X1 reaches 700 hours. Since this diagnosis target model (change characteristic IV) is within the range (confidence interval) of the tolerance model (change characteristics II and III) for the industry model G1, the coefficient AX1 of the model expression of the diagnosis target model is the industry model G1. It falls within the allowable range ± α for the coefficient AG1 of the model formula, and is determined to be in a normal state.

  FIG. 5 shows a diagnosis target model (change characteristic V) when the accumulated power generation time Tn of the diagnosis target device X1 reaches 1500 hours. Since this diagnosis target model (change characteristic V) is outside the allowable range model (change characteristics II and III) for the industry model G1 (outside the confidence interval), the coefficient AX1 of the model expression of the diagnosis target model is the industry model G1. It is determined that this is a special state that is out of the allowable range ± α with respect to the coefficient AG1 of the model formula and may be deteriorated or damaged.

  In this manner, in the present embodiment, by diagnosing whether or not the state is a special state using the coefficient A of the model formula, an abnormality sign of the diagnosis target device X1 is obtained before irreversible damage. It becomes possible to know.

  For reference, FIGS. 4 and 5 show an industry model G2 (change characteristic VI) generated from actual operation data of the devices X5 and X6 used in the industry G2. The industry G1 and G2 have different use conditions even in the same type of device 200, and therefore, the reference models generated from the actual operation data, that is, the industry models G1 and G2, are different. When the device X4 in the industry G2 is the device to be diagnosed, diagnosis is performed in the same manner as the device X1 in the industry G1 described above, using the tolerance range model defined for the industry model G2 (change characteristic VI). Is called.

  In the above-described embodiment, the device 200 is a fuel cell power generator, but a facility device such as a field device such as a valve or a transmitter or a heat source device may be used as the device 200. For such devices, remote monitoring systems are beginning to become popular as equipment maintenance targets. In many cases, the remote monitoring operator is not a user of a field device or facility device, but is an operator of a company corresponding to an external organization from the viewpoint of a user company. As a background to this, equipment is not a main process equipment such as a semiconductor manufacturing apparatus, and field equipment is often attached to equipment, so that operational confidentiality of operation data is not high. In addition, because a certain remote supervisor can collect data using a certain algorithm for many devices used by many different companies, it is possible to collect abundant data that only increases the statistical reliability. It is. By utilizing such features of field devices and equipment and utilizing remote monitoring data, the number of data when generating a reference model can be increased and the statistical reliability of the reference model can be increased. it can.

  In addition, not all devices used by many different companies are used under almost the same conditions. However, in relation to the characteristics of field devices and equipment that are not highly confidential, these devices are rarely used under conditions and conditions unique to individual companies. Easily adjusted to equivalent conditions. For example, the so-called “industry standard” is basically used for the tendency of the operating rate to rise and fall under the influence of the economy, and the purpose of use (use and target) of equipment. Therefore, even if the devices are of the same type, the statistical reliability of the reference model can be increased by classifying operation data by industry and using the data as statistical analysis method data by industry. For example, in the case of field equipment, there are classifications such as petroleum, chemical, steel, and water (water purification plant). For example, in the case of equipment, there are classifications such as semiconductors, foods, plastics, and pharmaceuticals.

  In the above-described embodiment, for example, the coefficient AG1 of the model equation of the industry model G1 is obtained, the allowable range ± α for the coefficient AG1 of the model equation of the industry model G1 is determined, and the coefficient AX1 of the model equation of the model to be diagnosed The condition of the diagnosis target device X1 is diagnosed by comparing the allowable value ± α with respect to the coefficient AX1 of the model formula of the diagnosis target model and the coefficient AG1 of the model formula of the industry model G1. It is not necessary to take a method using such a coefficient of the model formula. That is, the state of the diagnosis target device X1 is diagnosed by comparing on the screen the tolerance range model (change characteristics II and III) for the industry model G1 and the change characteristics (change characteristics IV and V) of the diagnosis target model. You may do it.

  In the above-described embodiment, only the diagnosis target device is specified to the device state diagnosis unit 5, but the industry may be specified together with the diagnosis target device.

[Extension of the embodiment]
The present invention has been described above with reference to the embodiment. However, the present invention is not limited to the above embodiment. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the technical idea of the present invention. Each embodiment can be implemented in any combination within a consistent range.

  DESCRIPTION OF SYMBOLS 1 ... Actual operation data collection part, 2 ... Reference | standard model production | generation part, 3 ... Permissible range model production | generation part, 4 ... Diagnosis object model production | generation part, 5 ... Device state diagnosis part, 100 ... Device state diagnosis apparatus by operation data ( Device state diagnosis apparatus), 200 (X1 to X6)... Device, 300.

Claims (2)

In the device status diagnosis device based on the actual operation data for diagnosing the status of the device designated as the diagnosis target device based on the actual operation data from the device,
Production data collection means for collecting production data from devices in production;
Based on the operation data collected from a plurality of devices of the same type and used in the same industry as the designated device to be diagnosed, the overall state of the device group of the same type and the same industry as the device to be diagnosed A reference model generation means for generating a reference model indicating a change characteristic;
An allowable range model generating means for generating an allowable range model indicating a range allowed as a normal state with respect to the reference model generated by the reference model generating means;
A diagnostic target model generating means for generating a diagnostic target model indicating a change characteristic of a current state of the diagnostic target device based on actual operation data collected from the specified diagnostic target device;
The diagnosis object model generated by the diagnosis object model generation means is compared with the tolerance model for the reference model generated by the tolerance range model generation means, and the state of the diagnosis object device is diagnosed based on the comparison result A device state diagnosis device using actual operation data. In the device status diagnosis apparatus according to the operation data according to claim 1,
The reference model generation means includes
Generating a coefficient of a model formula of the reference model;
The tolerance model generation means includes:
Generating tolerances for the coefficients of the model equation of the reference model;
The diagnostic object model generation means includes
Generating a coefficient of a model formula of the model to be diagnosed;
The device status diagnosis means
The coefficient of the model expression of the diagnosis target model generated by the diagnosis target model generation means is compared with the allowable range for the coefficient of the model expression of the reference model generated by the allowable range model generation means, and based on the comparison result And diagnosing the status of the device to be diagnosed.

Images