JP6856443B2 - Equipment abnormality diagnosis system - Google Patents

Description

The present invention relates to an abnormality diagnosis system for equipment that uses measurement data of equipment as input and performs abnormality diagnosis based on learning data prepared in advance.

Detecting abnormalities at the predictive stage by analyzing measurement data of equipment and taking countermeasures before a serious failure occurs is effective in avoiding unplanned outages due to failures and reducing loss costs. .. The processing function of the system that performs abnormality diagnosis for such equipment is roughly divided into the following two.

The first processing function is anomaly detection processing. Anomaly detection is to detect deviation from the normal state. In this case, if there is measurement data during normal operation, it is possible to construct a judgment logic for abnormality detection. The change pattern of the measurement data during normal operation may be learned in advance, and the pattern of the measurement data and the learning data may be compared. The learning content required for this processing includes a causal relationship between a plurality of measurement data during normal operation.

There is a physical relationship between the measurement data from multiple sensors installed in the equipment according to the design conditions. For example, in the case of a heat exchanger that raises the temperature of a low-temperature fluid by heat transfer from a high-temperature fluid, if the operating conditions such as the temperature and flow rate of the high-temperature fluid are constant, the rising temperature will decrease as the flow rate of the low-temperature fluid increases. There is a physical causal relationship such as By learning such a causal relationship in advance using the measurement data of a plurality of sensors measured during normal operation and monitoring whether the relationship is maintained, the normality / abnormality of the device can be determined.

In addition, the measurement data of equipment usually fluctuates even when the operating conditions such as output are constant. If the fluctuation characteristics of the data in the normal state, for example, the fluctuation range are learned, normal / abnormal can be determined by comparing with the fluctuation range of the measured data. The normal / abnormal discrimination process as described above is carried out based on the measurement data during normal operation.

On the other hand, the second processing function is diagnostic processing. The diagnostic process here is to acquire more detailed information on the abnormality than the normal / abnormal discrimination. Examples of the information provided in the diagnostic process include the cause of the abnormality and the location of the abnormality. In order to carry out the diagnostic processing, measurement data when an abnormality has occurred in the equipment is required. By collating the change pattern of the measurement data captured online with the change pattern at the time of an abnormality prepared in advance, the cause and the abnormal part of the abnormality currently occurring in the equipment are identified.

When an abnormality occurs in equipment, if the abnormality diagnosis system can also provide information on the location of the abnormality and the cause of the abnormality, it will be a guideline for the operator to decide how to deal with the abnormality. For example, if it is determined that the cause of the abnormality identified by the system is a minor event and is not serious enough to interfere with the operation of the device, it is possible to take measures such as continuing the operation. In addition, if the system provides information on the abnormal location, it is possible to quickly respond to the abnormality, such as arranging defective parts. As described above, it is useful for the operator to determine the response to the abnormality if the system provides information on the location of the abnormality and the cause of the abnormality.

As an example of a system for performing such an abnormality diagnosis, for example, Patent Document 1 discloses an abnormality based on a past operation history with respect to a correlation value representing a mutual correlation of a plurality of measurement parameters measured in a nuclear power plant. Abnormal events are identified by comparing them with the judgment parameters associated with the event.

Japanese Unexamined Patent Publication No. 2017-62730

As described above, the abnormal location and the cause of the abnormality can be identified by collating the measurement data of the equipment to be diagnosed with the change pattern of the measurement data prepared in advance at the time of abnormality.

However, depending on the equipment, the frequency of abnormality occurrence is low, and a situation may occur in which abnormality data cannot be accumulated. As described above, normal / abnormal discrimination requires only data during normal operation, so it is easy to accumulate data.

On the other hand, it is not easy to accumulate data because data at the time of occurrence of an abnormality is required to identify an abnormal event or an abnormal location. In addition, for abnormal data, it is necessary to prepare data for each cause of the abnormality. This is because if the cause of the abnormality is different, the change pattern of the measurement data is also different. When an abnormality occurs in equipment, the parts that caused it are replaced with new ones. For this reason, abnormalities mainly caused by material deterioration of parts rarely occur at the same location.

From the above points, it is difficult to prepare a set of abnormality data that covers various causes and locations of abnormalities that occur in equipment. Therefore, a processing method that enables abnormality diagnosis even when the acquired abnormality data is small has been desired.

From the above, the equipment abnormality diagnosis system according to the present invention has an input unit for inputting measurement data of the equipment to be monitored together with an input value, and a pattern for determining normality or abnormality of the measurement data using a diagnosis pattern. A diagnostic pattern is provided with a collating unit, a first database that stores measurement data of normal or abnormal results determined by the pattern collating unit, and a second database that stores simulated abnormality data that simulates abnormalities in equipment. The pattern generation unit is composed of a pattern generation unit that generates simulated abnormal data for a physical model or mathematical model constructed from normal measurement data, taking into account the assumed abnormal conditions, and stores it in a second database. At the same time, it is characterized in that a simulated diagnostic pattern is obtained by using the simulated abnormality data stored in the second database.

According to the equipment abnormality diagnosis system according to the present invention, even when the abnormality data measured by the equipment equipment is insufficient, it is possible to realize a process for identifying the cause of the abnormality and the location of the abnormality.

The figure which shows the outline of the whole structure of the abnormality diagnosis system which concerns on this invention. The figure which shows an example of the processing flow of specifying an abnormal part which concerns on Example 1. FIG. The figure which shows an example of the processing flow of specifying an abnormal part which concerns on Example 2. FIG. The figure which shows an example of the processing flow of specifying an abnormal part which concerns on Example 3. FIG. The figure which shows an example of the processing flow of specifying an abnormal part which concerns on Example 4. FIG. The figure which shows an example of the processing flow of specifying an abnormal part which concerns on Example 5. The figure which shows an example of the processing flow of specifying an abnormal part which concerns on Example 6. The figure which shows the schematic structure when the equipment which is the application example of the abnormality diagnosis system of this invention is a heat exchanger. The figure which shows the example of the diagnostic pattern list of the simulated abnormality data when the heat exchanger is implemented as an application example. The figure which shows the example of the database for pattern collation when the heat exchanger is implemented as an application example. A flowchart showing the operation contents of the pattern collation unit and the abnormality evaluation unit. The figure which shows an example of the display monitor screen of the abnormality diagnosis system of this invention.

The configuration of the abnormality diagnosis system for equipment according to the present invention will be described below with reference to the drawings.

Before going into the description of the examples, the overall configuration of the outline of the abnormality diagnosis system of the equipment according to the present invention will be clarified with reference to FIG. The following examples can be performed using part or all of the overall configuration shown in FIG.

As shown in FIG. 1, the abnormality diagnosis system 103 temporarily measures the measurement data D1 acquired by the sensor 102 installed in the equipment 101 and the input value D2 acquired from the input medium 110 in the operation unit 100. It is stored in the signal database DB1 and then acquired as a pair of data.

Further, the measurement data D1 is the data of the equipment 101 that can be directly measured by the sensor 102, but the information that cannot be directly measured or cannot be directly measured but can be estimated or identified from the measurement data D1 or the like is also available. Therefore, the signal processing unit 105 performs estimation and identification processing, and the estimated value and the identification value are used as signal processing result data D3, which is abnormal as data paired with the input value D2 acquired from the input medium 110 in the operation unit 100. Acquire to the diagnostic system 103.

Here, the reason why the measurement data D1 and the signal processing result data D3 are acquired as paired data with the input value D2 is that the measurement data D1 and the signal processing result data D3 are values indicating that the temperature is, for example, 30 degrees. However, this value is due to the fact that it is impossible to distinguish between normal and abnormal unless the magnitude of the input value D2 (or load value) of the equipment 101 is taken into consideration. Even if 30 degrees is a normal value at the rated input, it cannot always be said to be a normal value at 50% input. Therefore, the input value D2 (or load) is used to judge whether the measurement data D1 or the signal processing result data D3 is normal or abnormal. It is necessary to obtain the magnitude of (value) as a pair.

The input data (measurement data D1, input value D2, signal processing result data D3) input to the abnormality diagnosis system 103 is input to the pattern collation unit 107. In the pattern collating unit 107, the database of the diagnostic pattern generation unit 104 and the paired information of the current measurement data D1 and the input value D2 of the equipment 101 (or the paired information of the current signal processing result data D3 and the input value D2). Is used to identify a diagnostic pattern PT that matches the current measurement data D1, and to identify the occurrence of an abnormality and the location of the abnormality.

The diagnostic pattern generation unit 104 includes a plurality of databases, and generates diagnostic pattern PTs for normal and abnormal. The database in the diagnostic pattern generation unit 104 stores the normal database DB2 that stores the normal determination data D4 determined to be normal by the pattern matching unit 107, and the abnormality determined data D5 that is determined to be abnormal by the pattern matching unit 107. An abnormality case database DB4 and a simulated abnormality database DB3 for storing assumed simulated abnormality data are provided.

The normal database DB2 and the abnormal case database DB4 are stored in a table format as paired information with the input value D2 at that time. Further, the diagnostic pattern PT generated from these is also generated corresponding to the input value D2. Therefore, the diagnostic pattern PT for evaluating the current measurement data D1 is extracted based on the past measurement data D1 when the past input value D2 is the same as the current input value D2. It is the subject of comparison.

In this way, the database in the diagnostic pattern generation unit 104 stores data (past data) after discrimination between normal and abnormal, and is based on the past data and the diagnostic pattern PT obtained from the assumed abnormality as the past data. The pattern collation unit 107 will determine whether the current measurement data D1 or the current signal processing result data D3 is normal or abnormal.

Further, the abnormality diagnosis system 103 displays the identified abnormality and the occurrence location as the diagnosis result data D6 on the monitor screen 111 of the operation unit 100. The type, installation location, and number of sensors 102 are not limited to one, and it is preferable to install a plurality of sensors 102. Further, a signal processing unit 105 and an abnormality evaluation unit 108 can be added as a pre-processing unit or a post-processing unit of the abnormality diagnosis system 103. The abnormality evaluation unit 108 outputs the abnormality evaluation result data D8 and displays it on the monitor screen 111.

FIG. 1 shows the configuration of an abnormality diagnosis system for equipment and devices common to the examples described below in the present invention. Some embodiments of the present invention performed with reference to FIG. 1 will be described in detail with reference to the drawings of FIGS. 2 and subsequent sections. The components in the following embodiments can be combined as appropriate, and when there are a plurality of embodiments, each embodiment can be combined.

In the first embodiment, an example of the concept of identifying an abnormal portion will be described. FIG. 2 shows an example of a processing flow for identifying an abnormal location by the abnormality diagnosis system 103. Hereinafter, as an example, an abnormality determination method will be described according to the processing flow of FIG. Note that, in FIG. 2, among the databases in the diagnostic pattern generation unit 104 of FIG. 1, the normal database DB2 and the simulated abnormality database DB3 are used.

The abnormality diagnosis system 103 of FIG. 2 acquires the measurement data D1 and the input value D2 from the measurement signal database DB1, and the pattern collation unit 107 collates with the diagnosis pattern PT generated by the diagnosis pattern generation unit 104.

The diagnostic pattern generation unit 104 uses the normality-determined data D4 (hereinafter referred to as actual data) that has been verified to be normal and stored in the normal database DB2 to generate the actual diagnosis pattern PTR, and the actual diagnosis pattern generation unit 104R and simulated data. It is provided with a simulated diagnostic pattern generation unit 104V that generates a simulated diagnostic pattern PTV by using it.

The performance diagnosis pattern generation unit 104R is mainly composed of the normal database DB2. The normal database DB2 stores the measurement data value D1 and the operating state (input value D2, etc.) when the equipment 101 is operating normally, and the performance diagnosis pattern PTR generated from these performance data is stored. Obtained from the performance diagnosis pattern generation unit 104R. The normal database DB2 has the same or similar specifications when the history of measurement data when the equipment 101 has been operated in the past or when the equipment has just been operated and the past operation record of the equipment is small. The history of operating other equipment can be used.

The simulated diagnosis pattern generation unit 104V is mainly composed of the simulated abnormality database DB3. The simulated abnormality database DB3 is a simulation of measurement data estimated using a physical model or a mathematical model 301 based on a hypothetical abnormality condition 208, which assumes that an arbitrary abnormality occurs at an arbitrary location in the equipment 101. It is a database about values, and this is stored in the simulated diagnosis pattern generation unit 104V. In forming the physical model or the mathematical model 301, the normality-determined data D4 in the normal database DB2 is referred to. A plurality of simulated abnormality databases DB3 are prepared according to the assumed abnormality and the abnormality location. The simulated abnormality database DB3 is generated as a simulated diagnostic pattern PTV by combining the assumed abnormalities and abnormal locations with the simulated values of the measurement data.

As described above, the diagnostic pattern generation unit 104 stores the actual diagnosis pattern PTR generated from the actual data for the normal database DB2 and the simulated diagnostic pattern PTV generated from the simulated data for the simulated abnormality database DB3. When it is not necessary to distinguish between the actual diagnostic pattern PTR and the simulated diagnostic pattern PTV, it is simply referred to as the diagnostic pattern PT.

The pattern collation unit 107 collates the measurement data D1 with the diagnostic pattern PT and identifies the closest diagnostic pattern PT. As the collation method, for example, a comparison of changes over time of data, a decision tree, a clustering process, or a neural network can be used.

When performing comparison and collation data by specific data changes over time as the matching method, the variation of the measurement data D1 with respect to the time difference Δt of any measurement period including time t ₁ an error occurs Calculate and identify the diagnostic pattern PT with the closest change to Δt. When a plurality of measurement data D1s are acquired, a diagnostic pattern PT in which the positive and negative of each measurement data match may be adopted, or the measurement data items may be determined according to the priority order.

When diagnosing with a decision tree as a specific collation method, in addition to dividing the data according to the value range of the measurement data D1, the amount of change in the measurement data D1 with respect to the time difference Δt and the physical quantity calculated based on the measurement data D1. , The division using the state quantity and the like may be performed.

When collating data by clustering processing, for example, vector quantization, support vector machine, adaptive resonance theory, etc. can be used. Before the clustering process, the measurement data D1 may be subjected to preprocessing such as normalization.

According to this embodiment, even when the data measured by the equipment 101 is insufficient, the type of abnormality and the abnormality location of the equipment 101 can be specified.

FIG. 3 shows an example different from that of the first embodiment with respect to the processing flow for identifying the abnormal location by the abnormality diagnosis system 103. In FIG. 3, the databases (normal database DB2, simulated abnormality database DB3, abnormality case database DB4) in the diagnostic pattern generation unit 104 of FIG. 1 are used.

The diagnostic pattern generation unit 104 of the second embodiment is different from the first embodiment in that the diagnostic pattern PT is generated by using the abnormal case database DB4 in addition to the normal database DB2 and the simulated abnormality database DB3. The abnormality case database DB4 is stored in the performance diagnosis pattern generation unit 104R. The performance diagnosis pattern PTR generated from the normal database DB2 is shown as PTR1, and the performance diagnosis pattern PTR generated from the abnormality case database DB4 is shown as PTR2 in FIG.

The abnormality case database DB 4 holds the measurement data value (abnormality determined data D5) when the equipment 101 actually causes an abnormality, the type of abnormality, and the abnormality location as data. The accident case database DB4 contains specifications for the history of measurement data when an abnormality occurs in the past operation of the equipment 101, or when the equipment has just been operated and the past operation record of the equipment is small. Uses the history of when other equipment of the same or similar operation is operated and an abnormality occurs.

According to this embodiment, by storing the function of generating the performance diagnosis pattern PTR2 using the abnormality case database DB4 in the performance diagnosis pattern generation unit 104R, the abnormality that has occurred in the past is based on the actual situation. You can refer to the measured data. Therefore, in addition to the effect of the first embodiment, it is possible to improve the accuracy in identifying the abnormality and the abnormal portion for the abnormality in which the case exists.

FIG. 4 is different from the first embodiment in that the processing flow for identifying the abnormal location by the abnormality diagnosis system 103 is provided with the actual normal database DB 5 and the parameter adjustment unit 112.

In FIG. 4, the normal data D4 verified as normal by the diagnostic pattern collation by the diagnostic pattern collating unit 107 is stored in the actual normal database DB 5 as the actual normal data D10. In the parameter adjustment unit 112, the accumulation result of the actual normal database DB5 and the accumulation result of the normal database DB2 are given to the physical model or the mathematical model 302 for updating the parameters, and the information of the physical model or the mathematical model 301 is updated. Obtain parameter information D9.

In this case, the normal data D4, which is the accumulation result of the normal database DB2, is the data set or obtained in a relatively old age. It is the data obtained in. On the other hand, the actual result normal data D10 accumulated in the actual result normal database DB5 is an experience value in recent years. Generally, as the aging deteriorates, the physical model or mathematical model 301 that should have been optimally set by us also tends not to reflect the latest state. Therefore, the physical model or mathematical model 301 is compared with the old and new property data and the difference is used. The intention is to adjust the parameters of.

According to the configuration of the third embodiment, when the equipment 101 is not abnormal and the operation is performed normally, the abnormality diagnosis system 103 accumulates the operation history as the actual normal data D10. The parameter adjustment unit 112 uses the normal data D4 and the actual normal data D10 accumulated in the normal database DB2 to fit the physical model or the mathematical model 301, and optimizes the parameters included in the physical model or the mathematical model 301. Parameter information D9 can be obtained.

In this embodiment, the accuracy of the physical model or the mathematical model 301 can be improved by updating the parameters of the physical model or the mathematical model 301 using the actual normal data D10 accumulated during operation. In addition to the effect of the first embodiment, it is possible to create simulated abnormality data with higher accuracy.

FIG. 5 relates to another processing flow for identifying an abnormality location by the abnormality diagnosis system 103. Here, in Example 2 of FIG. 3, the normal data D4 and the abnormal data D5 obtained by the diagnostic pattern collation in the pattern collating unit 107 are accumulated in the actual normal database DB5 and the actual abnormal database DB6, and the actual normal data D10 and the abnormal data D5 are stored, respectively. As the actual abnormality data D11, the data is transferred to the normal database DB2 or the abnormality case database DB4 of the diagnosis pattern generation unit 104, and the diagnosis pattern PTR is generated from the data of the normal database DB2 or the abnormality case database DB4.

In this case, the normal data D4 and the abnormal data D5, which are the accumulation results of the normal database DB2 or the abnormal case database DB4 of the diagnostic pattern generation unit 104, are data set or obtained in a relatively old age, for example. This data was initially set at the time of designing the abnormality diagnosis system or was obtained in an era when the driving experience was not sufficient. On the other hand, the actual normal data D10 and the actual abnormal data D11 stored in the actual normal database DB5 or the actual abnormal database DB6 are recent experience values. Generally, as the deterioration over time, the diagnostic pattern PTR that should have been optimally set at our office tends not to reflect the latest state. Therefore, by considering the latest actual normal data D10 and actual abnormal data D11, the diagnostic pattern PTR1 It is intended to generate PTR2.

Here, when the measurement data D1 is determined to be normal, the data is stored in the normal database DB2 as the actual normal data D10, and this is added to the diagnostic pattern as normal data. If the measurement data D1 matches the diagnostic pattern of the simulated abnormality data, the data is stored in the actual abnormality database DB6 as the actual abnormality data D11, and the abnormality case database is combined with the abnormality type and abnormality location specified by the diagnostic pattern. Add to DB4.

According to this embodiment, the database can be expanded while operating the equipment 101 by adding the actual normal data D10 and the actual abnormal data D11 that identify the abnormality and the abnormal portion to the diagnostic pattern generation unit 104. it can. Therefore, in addition to the effect of the first embodiment, the diagnostic pattern PTR of the diagnostic pattern generation unit 104 to be collated while operating the equipment 101 can be subdivided, and the accuracy of normal / abnormal determination and identification of the abnormal portion can be improved. Can be done.

FIG. 6 relates to another example of a processing flow for identifying an abnormal location by the abnormality diagnosis system 103. In Example 1 of FIG. 2, the configuration in the case where the signal processing result data D3 from the signal processing unit 105 is input to the pattern collating unit 107 and used for specifying the diagnostic pattern PTR is illustrated. The point where the signal processing result data D3 from the signal processing unit 105 is input to the abnormality diagnosis system 103 and the diagnostic signal processing result data D13 output by the diagnostic signal processing unit 106 are held in the diagnostic pattern generation unit 104. The point is different.

The signal processing unit 105 performs signal processing on the input measurement data D1 and outputs the signal processing result data D3. The diagnostic signal processing unit 106 performs signal processing on the normal data D4 in the normal database DB2, the simulated abnormality data D12 in the simulated abnormality database DB3, and the abnormality data D5 in the abnormality case database DB4, and the signal processing unit 105 performs signal processing. The diagnostic signal processing result D13 corresponding to the signal processing result data D3 of is output. The diagnostic signal processing unit 106 may be replaced by the signal processing unit 105. Further, the diagnostic signal processing unit 106 may not be provided, and the diagnostic signal processing result D13 may be calculated in advance and stored in the diagnostic database.

As the signal processing, for example, a clustering method is used. The signal processing may use either a single method or a combination of a plurality of methods. In addition to the measurement data D1, the values input to the signal processing unit 105 and the diagnostic signal processing unit 106 are the normal data D4 in the normal database DB2, the simulated abnormality data D12 in the simulated abnormality database DB3, and the abnormality case database DB4. Part or all of the anomalous data D5 can be used. Further, these data may be divided, a plurality of signal processings may be performed in parallel, and a plurality of signal processing result data D3 and a plurality of diagnostic signal processing results D13 may be created.

According to this embodiment, by inputting the signal processing result data D3 into the abnormality diagnosis system in addition to the measurement data D1, the diagnosis accuracy can be improved in addition to the effect of the first embodiment.

FIG. 7 relates to another example of a processing flow for identifying an abnormal location by the abnormality diagnosis system 103. In addition to the first embodiment of FIG. 2, an abnormality evaluation unit 108 for estimating non-measurement data is provided, and the estimated non-measurement data is used to evaluate the cause of abnormality, the location of abnormality, and the degree of abnormality (degree of influence).

The anomaly evaluation unit 108 is activated and functions in some cases. One of them is a case where the pattern collation result in the pattern collation unit 107 becomes abnormal, and it is a scene in which it is desired to grasp the degree of the abnormality, that is, the magnitude of the influence of the occurrence of the abnormality on the operation of the equipment 101.

Another activation case is when the pattern matching in the pattern matching unit 107 becomes undecidable. The pattern collation in the pattern collation unit 107 determines whether the measurement data is normal or abnormal, but in reality, the pattern collation in the pattern collation unit 107 may not be discriminating. The diagnostic pattern generation unit 104 is created assuming as many diagnostic patterns as possible, but if a case that does not correspond to any of the assumed diagnostic patterns occurs, it becomes impossible to determine. Can be considered.

In the former case where the pattern collation result is abnormal, in the abnormality diagnosis system 103, when the measurement data D1 matches the diagnosis pattern of the simulated abnormality data (collation result, determination as abnormality), the abnormality evaluation unit 108 is input. The measurement data D1, the input value D2, and the diagnostic pattern PTR from the normal database DB2 are input to the physical model or the mathematical model 301, and the estimated non-measurement data D15 is output. From the difference between the estimated non-measurement data D15 calculated from the measurement data D1 and the estimated non-measurement data D16 in the normal state, the magnitude of the influence caused by the occurrence of an abnormality is calculated.

In the latter case where the pattern matching result cannot be determined, the abnormality diagnosis system 103 determines that the measurement data D1 does not match the diagnosis pattern PT of the diagnosis pattern generation unit 104, and therefore is not normal, that is, cannot be determined. The abnormality evaluation unit 108 is started.

At this time, the abnormality evaluation unit 108 refers to the abnormality cause database DB 7, identifies the abnormality location in the location / cause estimation unit 305, and calculates the magnitude of the impact in the impact assessment unit 306. For example, the abnormal location / cause estimation unit 305 compares the estimated non-measurement data D15 calculated from the measurement data D1 with the estimated non-measurement data D16 at the normal time, and identifies the data item having a large difference in value to identify the abnormal location. Make a specific. From the abnormality location / cause estimation unit 305, the abnormality information D17 and the amount of change D18 from the non-estimated measurement data at the normal time are obtained as outputs. The impact evaluation unit 306 calculates the magnitude of the impact caused by the occurrence of an abnormality from the difference between the estimated non-measurement data D15 calculated from the measurement data D1 and the estimated non-measurement data D16 in the normal state. Specific judgment methods of the abnormality location / cause estimation unit 305 and the impact evaluation unit 306 will be described later with an example of a heat exchanger.

According to this embodiment, in addition to the effect of the first embodiment, even when the measurement data D1 does not match the diagnostic pattern included in the normal database DB2 and the simulated abnormality database DB3, the type, location, and cause of the abnormality It can be identified. In addition, the effects caused by abnormalities can be quantitatively evaluated.

Example 7 shows an example in which the processing flow for identifying an abnormal location by the abnormality diagnosis system 103 is applied to equipment including a heat exchanger, and a heat transfer model of the heat exchanger is used as a physical model. FIG. 8 shows the equipment 101 including the assumed heat exchanger.

The heat exchanger HE as an example of the equipment shown in FIG. 8 is composed of two heat exchangers HEa and HEb connected in series, whereas the high temperature fluid FA (FAi on the inlet side and FAi on the outlet side) is used. A case where a low-temperature fluid FB (referred to as FAo) and a low-temperature fluid FB (referred to as FBi on the entry side and FBo on the exit side) is flowed in a countercurrent manner will be described.

The sensor 102 installed in the heat exchanger HE uses the measurement data D1 as the heat exchange of the heat exchanger HEa inlet temperature TA _{, in} , the heat exchanger HEb outlet temperature TA _{, out} , and the low temperature fluid FB of the high temperature fluid FA. Heat exchanger HEb inlet temperature TB _{, in} , flow rate _{FB, in} , pressure P _{B, in} , heat exchanger HEa outlet temperature TB _{, out} , pressure P _{B, out} , heat exchanger HEa and heat exchanger HEb The temperatures TB _{and m} between them are obtained.

For each measurement data D1, the values during normal operation are _{TAN, in} , _{TAN, out} , T _{BN, in} , T _{BN, out} , F _{BN, in} , P _{BN, in} , P _{BN, out} , T _{BN. , M.}

Here, as an example of the abnormal pattern of the heat exchanger HE, it is assumed that the heat passage rate is lowered due to the adhesion of dirt, the flow path is abnormal due to dirt or foreign matter, and the fluid leaks, and the physical model or the mathematical model 301 is used. , A simulated abnormal database DB3 was created. As the heat transfer model of the heat exchanger, the heat transfer formula of convective heat transfer shown below was used.

In the equations (1), (2) and (3), Q is the heat transfer amount (Js ^-1 ), A is the heat transfer area (m ² ), and U is the heat transfer efficiency (Jm ^-2 s ^-1 K ^-1 ). ΔT is the logarithmic mean temperature difference (K), H _{B and in} are the enthalpies of the low temperature fluid FB at the heat exchanger inlet (J), and H _{B and out} are the enthalpies of the low temperature fluid FB at the heat exchanger inlet (J). The subscripts i and o indicate the inflow side and the outflow side with respect to the heat exchangers HEa and HEb. The heat transfer area A is determined by the specifications of the equipment.

FIG. 8 shows the assumed abnormal conditions. As a hypothetical anomalous condition
1. 1. Decrease in heat transfer rate due to dirt adhesion, etc. (208A),
2. Flow path abnormality (208B),
3. 3. Fluid leakage (208C)
Occurs alone and
4. When the heat transfer rate decreases and the flow path becomes abnormal due to the adhesion of dirt (208D)
A total of 4 cases were assumed.

The creation of the simulated abnormality database DB3 will be described below, taking as an example the abnormality of "1. Decrease in heat transfer rate due to adhesion of dirt, etc." Since the calculation result of the simulated abnormality database DB3 changes depending on the abnormal location and the operating conditions of the heat exchanger, the simulated abnormality database DB3 has a plurality of diagnostic patterns PT for one type of abnormality.

"1. Decrease in heat transfer rate due to dirt adhesion, etc.":
Decrease in heat transfer rate due to dirt adhesion, etc .:
In FIG. 8, it is assumed that dirt adheres to a part of the heat exchanger HEa and the heat transfer rate decreases. Since the conditions under which the fluid FA and the fluid FB flow into the heat exchanger HEa and the heat exchanger HEb, respectively, do not change depending on the presence or absence of the abnormality in "1. Decrease in heat transfer rate due to adhesion of dirt, etc.", TA _{, in} = T _{AN, in} , TB _{, in} = T _{BN, in} , _{FB, in} = F _{BN, in} , P _{B, in} = P _{BN, in} . Further, it was assumed that the adhesion of dirt did not affect the flow of the fluid FB in the heat exchanger, and it was assumed that P _{B, out} = P _{B, in} = P _{BN, in} . Assuming that the heat passage rate U of the heat exchanger HEa decreases due to dirt adhesion, the outlet temperatures TA _{and out} of the high temperature fluid FA of the heat exchanger HEb become higher than those during normal operation, and the heat exchanger HEa The outlet temperatures TB _{and out} of the low temperature fluid FB are lower than those during normal operation. _{The combination of TA, out} and TB _{, out} is calculated from the equations (1), (2), and (3). FIG. 9 shows an example of simulated abnormality data stored in the diagnostic database created above.

In FIG. 9, the vertical axis illustrates dirt adhesion as a hypothetical abnormal condition, and heat exchanger HEa as an abnormal part. On the horizontal axis, the measurement data, the change tendency, and the diagnostic signal processing result data D13 are displayed by specific numerical values for each consistent case number.

_{At this time, the temperature (TB, m} ) of the low-temperature fluid FB between the heat exchanger HEa and the heat exchanger HEb is the same as during normal operation when the heat transfer rate of a part of the heat exchanger HEa decreases. Although it is about the same, when the heat transfer rate of the heat exchanger HEb is lowered, it becomes a lower value than in the normal state. As a result, it is possible to identify the abnormal portion by creating simulated abnormality data for the heat exchanger HEa and the heat exchanger HEb, assuming that an abnormality occurs in each of them as an abnormal condition 208.

Regarding "2. Flow path abnormality", "3. Fluid leakage", and "4. Heat transfer rate decrease due to dirt adhesion and flow path abnormality", "1. Heat transfer rate decrease due to dirt adhesion, etc." ], Each measurement data D1 is calculated using the equations (1), (2), and (3).

"2. Abnormal flow path due to dirt or foreign matter":
An example will be described in which dirt, foreign matter, etc. are accumulated in the flow path and an abnormality occurs in which the pressure loss in the flow path is increased. From FIG. 8, it is assumed that the pressure loss becomes high in the middle of the pipe through which the low temperature fluid FB inside the heat exchanger HEb is flowing. _{The temperatures (TA, in} , TB _{, in} ) when the fluid FA and the fluid FB flow into the heat exchanger HEa and the heat exchanger HEb, respectively, are the same as during normal operation because they do not depend on the presence or absence of abnormalities. Take the value of (TA _{, in} = T _{AN, in} , TB _{, in} = T _{BN, in} ). As the pressure loss in the flow path of the pipe through which the low temperature fluid FB flows increases, the pressure (P BB _{, out} ) of the low temperature fluid FB at the inlet of the heat exchanger HEb becomes higher than in normal operation.

"3. Fluid leakage":
The case where the flow path is damaged due to corrosion or other causes and the fluid leaks will be described as an example. From FIG. 8, it is assumed that the fluid leaks in the middle of the pipe through which the low temperature fluid FB inside the heat exchanger HEb is flowing, and the low temperature fluid FB enters the flow path of the high temperature fluid FA. The temperature conditions under which the fluid FA and the fluid FB flow into the heat exchanger HEa and the heat exchanger HEb, respectively, do not depend on the presence or absence of abnormalities, and therefore take the same values as during normal operation (TA _{, in} =). T _{AN, in} , TB _{, in} = T _{BN, in} ). Since the low-temperature fluid FB flows into the flow path of the high-temperature fluid FA inside the heat exchanger HEb, the outlet temperature (TA _{, out} ) of the high-temperature fluid FA of the heat exchanger HEb becomes lower than during normal operation. Since the low-temperature fluid FB partially leaks into the flow path of the high-temperature fluid FA, the pressure of the low-temperature fluid FB at the inlet of the heat exchanger HEb (PB _{, out} ) and the pressure of the low-temperature fluid FB at the outlet of the heat exchanger HEa ( P _{B, out} ) decreases. Due to leakage, in the subsequent leak point, the flow rate of the cryogen FB decreases, the temperature rise between the inlet of the heat exchanger HEb and heat exchanger HEa to the outlet ((T _{B, m} -T B _{, In} ), (TB _{, out} - _{TB, m} )) show different values from normal operation.

"4. Decreased heat transfer rate and abnormal flow path due to dirt adhesion":
The above-mentioned "1. Decrease in heat transfer rate due to adhesion of dirt" and "2. When abnormalities in the flow path occur at the same time" will be described as examples. The location where the abnormality occurs is the same as the assumptions in "1. Decrease in heat transfer rate due to adhesion of dirt" and "2. When abnormalities in the flow path occur at the same time". From "1. Decrease in heat transfer rate due to adhesion of dirt", the outlet temperature (TA _{, out} ) of the high temperature fluid FA of the heat exchanger HEb is higher than that during normal operation, and the outlet of the low temperature fluid FB of the heat exchanger HEa. The temperature (TB _{, out} ) becomes low. From "2. When the flow path abnormality occurs at the same time", when the flow path abnormality occurs, the pressure (PB _{, out} ) of the low temperature fluid FB at the inlet of the heat exchanger HEb becomes higher than that during normal operation. From the above, in the hypothesized "4. Decrease in heat transfer rate and abnormal flow path due to adhesion of dirt", the outlet temperature (TA _{, out} ) of the high-temperature fluid FA and the inlet pressure of the low-temperature fluid FB are compared to those during normal operation. (P _{B, in} ) is high, and the outlet temperature (TB _{, out} ) of the low-temperature fluid FB of the heat exchanger HEa is low.

"1. Decrease in heat transfer rate due to dirt adhesion", "2. Flow path abnormality", "3. Fluid leakage", "4. Heat transfer rate decrease due to dirt adhesion and flow path abnormality occur" The simulated abnormality database DB3 created for "Case" is associated with the simulated abnormality type and abnormality location, and is stored in the diagnostic pattern generation unit 104 as a diagnostic pattern PT. At this time, the characteristics of the simulated abnormality database DB3 for each type of abnormality or for each type of abnormality and the abnormality portion may be extracted in advance to create the pattern matching database DB8.

As the pattern collation database DB8, for example, a database in which the change tendency of each measurement data D1 and the diagnostic signal processing result are associated with each other for the type and abnormality location of the abnormality shown in FIG. 10 can be used. In addition to the change tendency and the diagnostic signal processing result, the value of the measurement data D1 may be added to the diagnostic pattern generation unit 104 and the pattern collation database DB 8 by processing the value of the measurement data D1 by a mathematical formula, four arithmetic operations, or the like. The pattern matching database can be expected to reduce the calculation load on the pattern matching unit 107.

The pattern collation unit 107 collates the normal database DB2, the simulated abnormality database DB3, and the measurement data D1 to identify a diagnostic pattern PT that matches the measurement data D1.

When the identified diagnostic pattern PT belongs to the normal database DB2 in the pattern collating unit 107, it indicates that the heat exchanger is operating normally on the monitor screen 111. On the other hand, when the specified diagnostic pattern PT belongs to the simulated abnormality database DB3, the type and location of the abnormality simulated by the simulated abnormality database DB3 are shown on the monitor screen 111.

A case of updating the parameters of the physical / mathematical model 301 using the parameter adjusting unit 112 described in the third embodiment will be described with respect to the present embodiment. In this embodiment, the heat transfer model of the heat exchanger is used as the physical / mathematical model 301. Among the parameters included in the heat transfer model, for example, by updating the parameters related to heat transfer efficiency, it is possible to take into account the fluctuation of the measurement data D1 due to the influence of changes in operating conditions that are not classified abnormally and minute deterioration of equipment over time. Can be done.

An example in which the processing flow for identifying the abnormality location by the abnormality diagnosis system 103 and the abnormality evaluation unit 106 is applied to a heat exchanger that exchanges heat between the high-temperature fluid FA and the low-temperature fluid FB will be described with reference to FIG. The sensor 102 installed in the heat exchanger has acquired the same items as the measurement data D1 shown in the seventh embodiment.

The measurement data D1 is input to the pattern matching unit 107 via the measurement signal database DB1. The pattern collation unit 107 collates the normal database DB2, the simulated abnormality database DB3, and the abnormality case database DB4 in the diagnosis pattern generation unit 104, and identifies the diagnosis pattern PT. The operation of the pattern collation unit 107 and the abnormality evaluation unit 108 is shown in the flowchart of FIG.

The process of the first process step S1 in the flowchart of FIG. 11 corresponds to the process of the pattern collating unit 107. Here, the measurement data D1 is compared with the diagnosis pattern of the normal database DB2, and if they match, it is determined to be "normal", the process proceeds to the process step S6, and the diagnosis result data D6 is displayed on the monitor screen 111.

If the measurement data D1 matches the diagnostic pattern of the simulated abnormality database DB3 or the abnormality case database DB4 in the processing of the processing step S1 of the pattern collating unit 107, it is determined as "abnormal", and the pattern collating unit 107 determines the abnormality evaluation unit 108. The diagnosis result data D6 is output to the processing step S3 of.

In the process of the process step S3 of the abnormality evaluation unit 108, the diagnosis result specified in the process of the process step S1 of the pattern collation unit 107 and the measurement data D1 are input, and the degree of influence due to the generated abnormality is output.

Specifically, the process of the process step S3 of the abnormality evaluation unit 108 is described. The diagnosis result data D6 and the measurement data D1 input from the measurement signal database DB1 are transferred to the convective heat transfer of the equations (1), (2) and (3). The value of the estimated non-measurement data D15 is calculated by inputting the heat formula into the converted physical model or the formula of the mathematical model 301. Similarly, the normal database DB2 is also input into the formula of the physical model or the mathematical model 301 and calculated as the estimated non-measurement data D15 at the normal time. The estimated non-measurement data D15 includes, for example, the heat transfer amount and heat transfer efficiency of the heat exchanger HEa and the heat exchanger HEb, the fluid flow rate of the high temperature fluid FA at the inlet and outlet of the heat exchanger HEa and the heat exchanger HEb, and the heat exchanger. Examples include the fluid flow rate of the low temperature fluid FB at the inlet and outlet of HEa, the amount of change in the fluid flow rate between the heat exchanger HEa and the heat exchanger HEb, and the like.

In the processing of the processing step S4 of the abnormality evaluation unit 108, the abnormality cause is estimated by referring to the abnormality cause database DB7 from the item in which the difference between the calculated estimated non-measurement data D15 and the value of the estimated non-measurement data D16 at the normal time is large. .. The process of the process step S4 of the abnormality evaluation unit 108 corresponds to the process of the abnormality location / cause estimation unit 305 of FIG. 7. The abnormality cause database DB 7 stores the correspondence relationship of the abnormality causes that can be physically estimated from each estimated non-measurement data D15. Specifically, the items in which the difference between the estimated non-measured data D15 and the estimated non-measured data D16 in the normal state are large, the physical phenomenon assumed according to the magnitude relationship thereof, and the cause of the abnormality are stored as data. In addition, information on items whose values are expected to change from normal operation, items that are expected to change, and non-measurement data is stored.

An example of a sensor failure will be described as an item in which the difference between the estimated non-measured data D15 and the estimated non-measured data D16 in the normal state is large, and as a possible cause of abnormality depending on the magnitude relationship thereof.

Sensor failure:
When a sensor failure occurs, only the value acquired by one sensor in the measurement data D1 has a large difference from the value during normal operation. _{For example, a case where the outlet temperature (TB, out} ) of the low-temperature fluid FB of the heat exchanger HEa shows a value significantly lower than that during normal operation will be described. In the above cases, in _{addition to TB and out , the inlet temperature (TA, in} ) of the high-temperature fluid FA of the _{heat exchanger HEa, the outlet pressure (P B, out} ) of the low-temperature fluid FB of the heat exchanger HEa, and non-measurement It is estimated that there is a change in the value of any one or more of the high temperature fluid FA, the low temperature fluid FB flow rate, and the heat transfer efficiency in the heat exchanger HEa, which is the data. However, if the above tendency is not seen from the measurement data D1 and the estimated non-measurement data D15, a sensor failure is suspected as the cause of the abnormality.

In the processing steps S5 and S6 of the abnormality evaluation unit 108, the abnormality information D17 including the estimated abnormality location and cause and the amount of change D18 of the calculated estimated non-measurement data D15 from the estimated non-measurement data D16 at the time of normal are obtained. Evaluate the degree of impact. The processing of the processing steps S5 and S6 of the abnormality evaluation unit 108 corresponds to the processing of the influence evaluation unit 306 of FIG.

The degree of influence is output as information useful for the magnitude of the abnormality of the heat exchanger due to the occurrence of the abnormality, the assumed secondary damage, and the planning of countermeasures. For example, when the abnormality information D17 is a leakage of the low temperature fluid FB in the heat exchanger HEb, the leakage amount of the low temperature fluid FB is calculated as the degree of influence. In addition, when it is estimated that the secondary damage is, for example, mixing of high temperature fluid FA and low temperature fluid FB, or expansion or increase of leakage points due to abnormal flow path of low temperature fluid FB, based on the operating conditions and the amount of leakage at the time of leakage. Is output as an assumed secondary damage. In the processing step S5 of the abnormality evaluation unit 108, it is determined whether or not there is an estimated secondary damage, and if there is damage, the process proceeds to the processing step S6 to evaluate the time at which the degree of influence of the secondary damage occurs. If there is no damage in the process of the process step S5, the process step S6 is not executed.

Returning to the processing step S1 of FIG. 11, when the pattern collation unit 107 cannot identify the measurement data D1 in any of the diagnostic patterns of the normal database DB2, the simulated abnormality database DB3, and the abnormality case database DB4, "abnormality" is displayed. Assuming that there is a possibility, the abnormality evaluation unit 108 evaluates the degree of influence. In evaluating the degree of influence, in the processing step S2, the abnormal portion is identified from the diagnostic pattern, and thereafter, the degree of impact evaluation processing is performed in the processing step S3. The abnormality evaluation unit 108 calculates the estimated non-measurement data D15 and the estimated non-measurement data D16 at the normal time in the same manner as when it matches the diagnosis pattern of the simulated abnormality database DB3 or the abnormality case database DB4, and sets the error cause database DB7. Estimate the location of the abnormality and the cause of the abnormality by collation. The degree of influence is evaluated from the abnormality information D17 including the estimated abnormality location and cause and the amount of change D18 of the calculated estimated non-measurement data D15 from the estimated non-measurement data D16 at the normal time.

The identified diagnostic pattern 104, the degree of impact, and the estimated secondary damage are displayed on the monitor screen 111, and the identified diagnostic result data D6, abnormality evaluation result data D8, and measurement data D1 are registered in the actual normal data D4 or the actual abnormal data D5. , Stored in the diagnostic pattern generation unit 104.

At this time, a mechanism may be provided in which the user can determine whether or not the registration in the actual result normal data D4 and the actual result abnormality data D5 and the storage in the diagnostic pattern generation unit 104 are possible. The process step S8 of FIG. 11 indicates a process in which the user determines whether or not the case can be registered, and the process step S9 indicates a process of registering in the diagnostic database 104.

FIG. 12 is an example of displaying the diagnosis result data D6 output by the abnormality diagnosis system 103 and the abnormality evaluation result data D8 output by the abnormality evaluation unit 108 on the monitor screen 111.

The monitor screen 111 displays the input value D2, the diagnosis result data D6, and the abnormality evaluation result data D8 on the operation status display unit 500, the equipment schematic diagram 501, and the result display unit 502. In the operation status display unit 500, the presence / absence, type, and abnormal location of the abnormality specified based on the diagnosis pattern PT of the diagnosis result data D6 are estimated as the abnormal condition based on the impact evaluation result data D8, and the secondary damage is assumed. Is output. In the equipment schematic diagram, the configuration of the equipment is illustrated based on the equipment information input by the input value D2. From the diagnosis result data D6, the abnormality information is displayed at the corresponding part of the equipment schematic diagram. The result display unit 502 displays the details of the diagnosis result data D6. Alternatively, the result display unit may display a data monitoring unit 503 showing the transition of the measurement data D1 and the signal processing result data D3.

100: Operation unit 101: Equipment 102: Sensor 103: Abnormality diagnosis system 104: Diagnosis database 105: Signal processing unit 106: Diagnostic signal processing unit 107: Pattern matching unit 108: Abnormality evaluation unit 110: Input medium 111: Monitor screen 112: Parameter adjustment unit 208: Assumed abnormal condition D1: Measurement data D2: Input value D3: Signal processing result data D4: Normal judgment data D5: Abnormality judgment data D6: Diagnosis result data D8: Abnormal evaluation result data D9: Parameter information D10: Actual normal data D11: Actual abnormal data D12: Simulated abnormal data D13: Diagnostic signal processing result D15: Estimated non-measured data D16: Estimated non-measured data at normal time D17: Abnormal information D18: Estimated non-measured data D15 Normal time estimated amount of change from non-measurement data D16 DB1: Measurement signal database DB2: Normal database DB3: Simulated abnormality database DB4: Abnormal case database DB5: Actual normal database DB6: Actual abnormal database DB7: Abnormal cause database DB8: Pattern Collation database PT: Diagnostic pattern 208A: Assumed abnormal condition (dirt adhesion)
208B: Assumed abnormal condition (flow path abnormality)
208C: Assumed abnormal condition (leakage)
208D: Assumed abnormal conditions (dirt adhesion and flow path abnormality)
301: Physical model or mathematical model HEa: Heat exchanger HEb: Heat exchanger FA: High temperature fluid FB: Low temperature fluid

Claims (15)

An input unit for inputting measurement data in equipment to be monitored, a pattern collation unit for determining normality or abnormality of the measurement data using a diagnostic pattern, and a record of normality or abnormality determined by the pattern collation unit. It is composed of a pattern generation unit that generates the diagnostic pattern, including a first database that stores the measurement data and a second database that stores simulated abnormality data that simulates an abnormality in the equipment.
The pattern generator is a physical model or a mathematical model of the equipment to be monitored constructed by using the normal measurement data, and an abnormal condition of assumptions about an abnormality and an abnormal part in the equipment to be monitored. The second database is provided with a hypothetical anomalous condition setting unit for setting the above, and the second database for storing the simulated anomaly data generated by adding the assumed anomalous condition to the physical model or the mathematical model. An abnormality diagnosis system for equipment, which comprises obtaining the simulated diagnosis pattern using the stored simulated abnormality data. The equipment abnormality diagnosis system according to claim 1.
An abnormality diagnosis system for equipment, which comprises obtaining the actual diagnostic pattern by using the actual measurement data for which normality or abnormality is determined stored in the first database. The equipment abnormality diagnosis system according to claim 1 or 2.
An equipment abnormality diagnosis system, characterized in that the normal measurement data obtained at different times is given to the physical model or the mathematical model, and the parameters of the physical model or the mathematical model are adjusted. The equipment abnormality diagnosis system according to any one of claims 1 to 3.
An abnormality diagnosis system for equipment, characterized in that the measurement data of normal or abnormal results stored in the first database is added each time it is determined by the pattern collating unit. The equipment abnormality diagnosis system according to any one of claims 1 to 4.
The measurement data includes signal processing result data obtained by processing the measurement data, and the signal processing result data is processed in the same manner as the measurement data. The equipment abnormality diagnosis system according to any one of claims 1 to 5.
The abnormality diagnosis system of the equipment is provided with an abnormality evaluation unit, and the abnormality evaluation unit determines the result of pattern collation in the pattern collation unit, the abnormality of the measurement data, or the result of pattern collation in the pattern collation unit. An equipment abnormality diagnosis system, characterized in that, when the diagnosis pattern corresponding to the measurement data of normal or abnormality does not exist, the degree of influence of the generated abnormality on the equipment is obtained. The equipment abnormality diagnosis system according to claim 6.
The abnormality evaluation unit determines the abnormality of the measurement data as a result of pattern matching in the pattern matching unit, or the diagnostic pattern corresponding to the measurement data of normal or abnormal as a result of pattern matching in the pattern matching unit. An equipment abnormality diagnosis system characterized in that the location and cause of an abnormality in equipment are estimated when there is no such thing. The equipment abnormality diagnosis system according to any one of claims 1 to 7.
The abnormality diagnosis system of the equipment is provided with a monitor screen, and the data in the input unit, the data in the pattern matching unit, the data in the pattern generation unit, and the data in the abnormality evaluation unit in the abnormality diagnosis system are directly displayed on the monitor screen. Or an abnormality diagnosis system for equipment, which is characterized by being processed and displayed. The equipment abnormality diagnosis system according to any one of claims 1 to 8.
As a method of comparing the measurement data of the equipment acquired online or the signal processing result of the measurement data with the measurement data at the normal time and the abnormal time or the signal processing result of the measurement data, comparison of changes over time of the data, or An equipment anomaly diagnostic system characterized by using a decision tree, or a clustering process, or a neural network. The equipment abnormality diagnosis system according to any one of claims 1 to 9.
By importing the measurement data of equipment online and setting the values for the parameters corresponding to the measurement data among the parameters that make up the physical model or mathematical model, the values of the parameters corresponding to the non-measurement data are estimated and estimated. An abnormality diagnosis system for equipment, which is characterized by evaluating the cause of abnormality, the location of abnormality, or the degree of abnormality based on the obtained values. The equipment abnormality diagnosis system according to any one of claims 1 to 10.
The equipment abnormality diagnosis system is an equipment abnormality diagnosis system characterized in that equipment having a heat exchanger is targeted for abnormality diagnosis and a heat transfer model of the heat exchanger is used. The equipment abnormality diagnosis system according to claim 11.
An equipment abnormality diagnosis system characterized in that parameters related to heat transfer efficiency among the parameters in the heat transfer model of the heat exchanger are adjusted based on the measurement data values of the equipment. The equipment abnormality diagnosis system according to claim 11 or 12.
Abnormality diagnosis of equipment characterized by a decrease in heat transfer efficiency due to adhesion of dirt, clogging or leakage of piping through which fluid flows, or sensor failure as an abnormal event simulated by the heat transfer model of the heat exchanger. system. The equipment abnormality diagnosis system according to any one of claims 11 to 13.
As a non-measured parameter estimated by the heat transfer model of the heat exchanger, it is characterized by the heat transfer efficiency, the flow rate of the fluid, the amount of leakage from the pipe, or the original state amount at the time of sensor failure. Abnormality diagnosis system for equipment. The equipment abnormality diagnosis system according to any one of claims 11 to 14.
Equipment that is characterized by being accompanied by an information providing service that provides at least one of the type of abnormality identified in the abnormality diagnosis system and incidental parts, the location of the abnormality, the cause of the abnormality, the effect of the abnormality, and the secondary damage. Abnormality diagnosis system.

Images