JP2019153045A - Data processor and data processing method - Google Patents

Abstract

To upgrade the reliability of data, which is employed in monitoring the operation of a plant, by presenting a work instruction to an operator on the basis of a result of evaluation of each of the states of a measuring instrument and the plant.SOLUTION: A first diagnosis unit 32 included in a data processor 23 analyzes data, which is read from a first database on the basis of a condition read from a second database, according to a first statistical technique, and thus diagnoses presence or absence of an abnormality in a measuring instrument. A second diagnosis unit 33 analyzes data, which is read from the first database on the basis of a result of diagnosis by the first diagnosis unit and a condition read from the second database, according to a second statistical technique, and thus diagnoses presence or absence of an abnormality in performance of a facility. A state evaluation unit 34 selects information, which is used to give a work instruction to an operator of a plant, on the basis of a result of evaluation performed on each of the states of the measuring instrument and the plant according to a result of diagnosis of presence or absence of an abnormality in the measuring instrument and a result of diagnosis of presence or absence of an abnormality in the performance of the facility. A presentation unit presents the information selected by the state evaluation unit.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a data processing apparatus and a data processing method.

  A water treatment plant, a chemical plant, a power plant, etc. are comprised from various apparatuses and piping. Measuring devices for measuring water quality, temperature, flow rate, and the like are installed in the equipment and piping constituting these plants. And the measuring device is monitoring the state of an apparatus, a plant, and a product (for example, tap water if it is a water purification plant). In addition, the introduction of an abnormality detection / diagnosis system that uses these measurement values to detect an abnormality at a higher level or diagnose the content of the abnormality is also in progress.

  Various techniques have been proposed as an abnormality detection / diagnosis technique for analyzing measurement data output from a measuring instrument and grasping an operation state. For example, an abnormality diagnosis system based on a physical model and a general-purpose system include a diagnostic system using statistical methods such as principal component analysis, latent variable projection method, support vector machine, and the like. Although these methods are different, in general, not only measurement data but also a creation step (hereinafter referred to as “pre-processing”) for processing the measurement data into input data for analysis is required. The quality and creation time of input data for analysis affect the accuracy and speed of abnormality prediction, respectively.

  For example, water treatment plants that make up a water treatment system use river water or groundwater as raw water, and after supplying water and disinfecting water, finally supply water to customers via pipes. . Many of the water treatment facilities use surface water such as river water as the source of water, and the quality of raw water varies depending on the season and weather. Therefore, it is necessary for the maintenance manager of the water purification treatment to adjust the operation of the water purification treatment facility according to the quality of the raw water. Water purification facilities are operated based on plant operation / monitoring data. However, in order to realize continuous water supply for 24 hours with a small number of people, the importance of abnormality detection and diagnosis of equipment is increasing.

  However, the water used in water purification treatment not only contains substances derived from rivers and substances used in water purification treatment, but also contains microorganisms. These substances deteriorate the measurement accuracy of water quality and flow rate. Cause it. Moreover, the measurement data includes abnormal values and drifts due to noise of devices and signal transmission lines, events such as maintenance, malfunctions of measuring instruments, and the like. For this reason, when detecting abnormality or creating diagnostic data, it is necessary to remove abnormal values, drifts, and the like from the measurement data by setting a range based on physical characteristics or correcting with an evaluation formula.

  Here, Patent Document 1 discloses a technique for determining drifts of a plurality of measuring instruments that are provided in a plant facility and are correlated with measured process data.

JP 2009-175870 A

  By the way, in Patent Document 1, four water level meters that measure the reactor water level are listed as examples as a plurality of measuring instruments with which correlation is recognized. These water level gauges are used as measuring instruments that measure the same items in the same way in a nuclear reactor. In addition, during the commercial operation in which the operation state in which power generation is constantly performed is maintained, the physical properties and impurity concentration of water that is a measurement target of these water level meters do not change greatly.

  However, in a general water treatment system, a plurality of measuring instruments often measure water that is a measurement target in different situations. For this reason, as in a water treatment system, the situation in which the impurity concentration of the water to be measured and the amount of water to be handled change greatly, or the combination of measuring instruments with different measurement items and measurement ranges that are correlated with the measurement values can be selected. Then, data preprocessing was difficult. For this reason, the quality of the input data for analysis is low for data that has not been appropriately preprocessed. When such input data for analysis is used, the accuracy and speed of abnormality prediction is low.

  The present invention has been made in view of such a situation, and an object thereof is to improve the reliability of data used for plant operation monitoring.

  A data processing apparatus according to the present invention is a first database that stores, in time series, data output from equipment constituting a plant as the plant is operated and data output from a measuring instrument that measures a processing target of the plant. A second database in which conditions for analysis or diagnosis are defined, and analyzing data read from the first database based on the conditions read from the second database by a first statistical method. A first diagnostic unit for diagnosing the presence / absence of abnormality, a diagnosis result of the presence / absence of abnormality of the measuring instrument, and data read from the first database based on conditions read from the second database are analyzed by a second statistical method. The second diagnostic unit for diagnosing the presence or absence of equipment performance abnormality, the diagnostic result of the presence or absence of abnormality of the measuring instrument, and the presence or absence of equipment performance abnormality A state evaluation unit that selects information for giving a work instruction to a plant operator based on an evaluation result obtained by evaluating the state of the measuring instrument and the plant based on the diagnosis result, and a presentation that presents the information selected by the state evaluation unit A section.

According to the present invention, the operator can perform appropriate work based on the work instruction presented to the operator based on the evaluation result obtained by evaluating the state of the measuring instrument and the plant. Can improve the reliability of the data.
Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

It is a block diagram which shows the structural example of the data pre-processing and analysis system for plant diagnosis which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the structural example of the analyzer which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the hardware structural example of the computer which concerns on the 1st Embodiment of this invention. It is explanatory drawing which shows the example of the category classification | category of the driving | running data by the category analysis part which concerns on the 1st Embodiment of this invention. It is explanatory drawing which shows the example of the correlation of the category after the classification | category by the category analysis part which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the data processing method by the analyzer which concerns on the 1st Embodiment of this invention. It is a table | surface which shows an example of the data item which shows the analysis / diagnostic condition which the category analysis part and measuring device diagnostic part which concern on the 1st Embodiment of this invention can acquire from analysis / diagnostic condition DB. It is a table | surface which shows an example of the classification item for classifying the content displayed on the display part which concerns on the 1st Embodiment of this invention. It is a table | surface which shows an example of the content displayed on a display part according to the classification | category which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the data processing method by the analyzer which concerns on the 2nd Embodiment of this invention.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the accompanying drawings. In the present specification and drawings, components having substantially the same function or configuration are denoted by the same reference numerals, and redundant description is omitted.

[First Embodiment]
First, the plant diagnosis data preprocessing / analysis system 100 according to the first embodiment of the present invention will be described.
FIG. 1 is a configuration diagram of a plant diagnosis data pre-processing / analysis system 100.
The plant diagnosis data preprocessing / analysis system 100 is an example of a data processing system that processes data used for abnormality detection / diagnosis of a water supply facility 1 configured as an example of a plant. The plant according to the present embodiment is a treatment plant for performing water purification treatment, sewage treatment, water intake, water collection, or water supply / distribution, and particularly the water supply facility 1 will be described. The plant diagnosis data preprocessing / analysis system 100 includes equipment and measuring instruments installed in the water supply facility 1 and an analysis device 23. The object managed by the plant diagnosis data preprocessing / analysis system 100 is the water supply facility 1. The water supply facility 1 includes a water intake facility 2, a water purification treatment facility 3, a water transmission / distribution facility 4, and a water supply facility 5.

  The water treatment facility 3 includes, for example, a landing well 10 that receives raw water, a mixing basin 11 in which a flocculant is added and rapidly stirred, a floc formation pond 12 in which slow stirring is performed to grow flocs, and a precipitate that separates the grown flocs by sedimentation. A pond 13, a filter pond 14 for filtering water, and a water purification pond 15 for storing purified water are provided. In this Embodiment, the measuring device which measures a process flow rate, water quality, etc. is the water quality measuring device which measures the water quality of the water as a process target of the water supply facility 1, for example, and the turbidimeters 16a-16c and the flow meter 17 are measured. There are a water temperature meter 18 and a pH meter 19.

  A turbidity meter 16a, a flow meter 17, a water temperature meter 18, and a pH meter 19 installed in the landing well 10 measure the quality of raw water and the amount of water (flow rate) for water purification treatment, respectively. The turbidimeter 16b is installed at the outlet of the settling basin 13 and is used to determine the performance of the precipitation process. Moreover, the turbidimeter 16c is installed in the exit of the filtration basin 14, and measures the water quality after filtration. The water temperature gauge 18 measures the temperature of water stored in the landing well 10. The pH meter 19 measures the pH of the water stored in the landing well 10.

  Water treatment by chemical injection is performed at the water purification treatment facility 3. For this reason, the water purification treatment facility 3 has an alkaline agent injection facility 20 for injecting an alkaline agent such as a sodium hydroxide solution into the landing well 10, and a coagulant injection for injecting a coagulant such as PAC (polyaluminum chloride) into the mixing basin 11. Equipment 21 is provided. Measurement information including measurement values of measuring instruments such as the turbidity meter 16 a, flow meter 17, water temperature gauge 18, pH meter 19, and operation information of the measuring instrument are transmitted to the analyzer 23 via the control LAN 22.

FIG. 2 is a block diagram illustrating a configuration example of the analysis device 23.
The analysis device 23 includes a communication I / F (Interface) 31, a measuring instrument diagnosis unit 32, a category analysis unit 33, a state evaluation unit 34, a process / water quality DB (Data Base) 35, an analysis / diagnosis condition DB 36, a display unit 37, And an internal bus 38.

  The communication I / F 31 is connected to a control LAN (Local Area Network) 22, and acquires data indicating measured values such as operating conditions and water quality from the equipment configuring the water supply facility 1. In the following description, the data indicating the operating condition of the water supply facility 1 and the data indicating the measurement value measured by the measuring instrument are also referred to as “operation data”. The operation data is an example of data for plant diagnosis used for managing the operation state of the water facility 1 and performing diagnosis of the water facility 1. The operation data is stored in the process / water quality DB 35 at a set frequency.

  The measuring instrument diagnosis unit 32 (an example of a first diagnosis unit) analyzes the data read from the process / water quality DB 35 by using the first statistical method, thereby diagnosing the presence or absence of an abnormality of the measuring instrument. For example, the instrument diagnosis unit 32 uses a statistical method such as cluster analysis, a physical model, or a correlation equation obtained from another measurement value, so that the value indicated by the instrument is correct within a predetermined error range. Diagnose whether it shows a value. For this reason, the measuring instrument diagnosis unit 32 uses the diagnosis conditions stored in the analysis / diagnosis condition DB 36, that is, the data items and numerical values to be input, and the data set stored in the process / water quality DB 35 as a target. Diagnose the measuring instrument. As a first statistical method used in the instrument diagnosis unit 32, an adaptive resonance theory (hereinafter referred to as “ART”) shown as an example of a category analysis method used in the category analysis unit 33 is used. Alternatively, a simple linear model different from ART or a test method may be used.

  The category analysis unit 33 (an example of the second diagnosis unit) analyzes the data read from the process / water quality DB 35 based on the diagnosis result of the measuring device diagnosis unit 32 by using the second statistical method, thereby causing abnormal performance of the equipment. Diagnose the presence or absence of. For this reason, the category analysis unit 33 uses, for example, an engine that executes a technique using ART that classifies data read from the process / water quality DB 35 into a category or a cluster analysis technique similar to ART as the second statistical technique. Prepare. The category analysis unit 33 uses the analysis conditions stored in the analysis / diagnosis condition DB 36, that is, the data items and numerical values to be input, and a data set that is a cluster of various data stored in the process / water quality DB 35, The category into which this data set is classified is output.

  The category analysis method used by the category analysis unit 33 is not limited, but in the present embodiment, for example, the following description is continued assuming that ART is used as the category analysis method. The category analysis method using ART performed by the category analysis unit 33 is a method of classifying a data set stored in the process / water quality DB 35 into a category according to a combination of a plurality of operation data output from a plurality of measuring instruments. is there. Then, in the category analysis method using ART, the combination of operation data output when the water supply facility 1 is operating normally is stored as a normal state, and if it has never appeared in the past, a new category is stored. Generate automatically. The category analysis method using ART has an advantage that it does not require condition setting by the user. A category analysis method using ART will be described later with reference to FIGS.

  The state evaluation unit 34 operates the water supply facility 1 based on the evaluation result obtained by evaluating the state of the measurement device and the water facility 1 based on the result of the measurement by the measurement device diagnosis unit 32 and the analysis result by the category analysis unit 33. Select information to give work instructions to employees. For example, the state evaluation unit 34 detects that a new category has been automatically generated by the category analysis unit 33 as a state change of the measuring device and the water facility 1, so that the state of the measuring device and the water facility 1 is normal. Or whether it is abnormal. Then, the state evaluation unit 34 outputs information for giving a work instruction to the operator on the display unit 37.

  The process / water quality DB 35 (an example of a first database) measures the process data output from the equipment configuring the water supply facility 1 as the water supply facility 1 is operated, and the water quality of the water to be processed by the water supply facility 1. Stores the measurement data output from the measuring instrument in time series. The time series data stored in the process / water quality DB 35 is used by the analysis device 23. This time series data is operation data acquired online from the water supply facility 1 in a time series, and includes data output from various measuring instruments. For example, as time-series data items, raw water turbidity, pH, alkalinity, water temperature, chlorine consumption, chemical injection rate in the landing well 10 and the mixing pond 11, water intake, precipitation treated water turbidity, filtration Examples include speed, head loss of filter basin, turbidity of filter basin, residual chlorine concentration, cleaning performance of filter basin, sludge concentration, calibration records of measuring instruments, and maintenance inspection records of various devices. The frequency with which the process / water quality DB 35 acquires and stores time-series data is preferably, for example, 1 second intervals, but the characteristics of the water supply facility 1 (such as the rate of change of water quality and the amount of water and the rate of change of the amount of chemical injection) are diagnosed. Depending on the required time interval, data may be acquired and stored at intervals longer than 1 second.

  The analysis / diagnosis condition DB 36 (an example of a second database) stores conditions for analysis or diagnosis such as data items of data used in the measuring instrument diagnosis unit 32 and the category analysis unit 33 and the frequency of analysis / diagnosis. For example, the target of category analysis by the category analysis unit 33 is “flocculation treatment water turbidity”, and input items for category analysis are “flocculating agent injection rate”, “water intake flow rate”, “raw water turbidity”, “ "Water temperature" can be taken. Further, the category analysis target may be “water pump”, and “pump flow rate”, “current”, and “pressure” may be taken as input items for category analysis.

  In addition, the object of the instrument diagnosis by the instrument diagnosis unit 32 is “precipitation water turbidity meter”, and “raw water turbidity”, “filtered water turbidity”, “coagulant injection rate” are input items used for the instrument diagnosis. Can take. In addition, the instrument diagnosis target is “flocculating agent flow meter”, and the input items used for instrument diagnosis are “raw water turbidity”, “water intake”, “precipitation water turbidity”, “stirring speed”, etc. You can also take.

  The display unit 37 (an example of a presentation unit) displays the evaluation results obtained from the state evaluation unit 34 and support information including work instructions to the operators of the water facility 1 (not shown) connected to the analyzer 23. It is displayed on a PC (Personal Computer) display or portable terminal, or printed on a paper medium such as a form.

Next, the hardware configuration of the computer C constituting the analysis device 23 will be described.
FIG. 3 is a block diagram illustrating a hardware configuration example of the computer C.

  The computer C is hardware used as a so-called computer. The computer C includes a CPU (Central Processing Unit) C1, a ROM (Read Only Memory) C2, and a RAM (Random Access Memory) C3 connected to the bus C4. Further, the computer C includes a display device C5, an input device C6, a nonvolatile storage C7, and a network interface C8.

  The CPU C1 reads out the program code of software that implements each function according to the present embodiment from the ROM C2, and executes it. In the RAM C3, variables, parameters and the like generated during the arithmetic processing are temporarily written. The display device C5 is a liquid crystal display monitor, for example, and displays the results of processing performed by the computer C to the operator. For example, a keyboard, a mouse, or the like is used as the input device C6, and an operator can perform predetermined operation inputs and instructions.

  As the nonvolatile storage C7, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), a flexible disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a magnetic tape, a nonvolatile memory, or the like is used. It is done. In addition to the OS (Operating System) and various parameters, a program for causing the computer C to function is recorded in the nonvolatile storage C7. The ROM C2 and the non-volatile storage C7 permanently record programs and data necessary for the operation of the CPU C1, and are computer-readable non-transitory that stores programs executed by the computer C. Used as an example of a recording medium.

  As the network interface C8, for example, a NIC (Network Interface Card) or the like is used, and various types of data can be transmitted and received between devices via a LAN (Local Area Network) connected to a terminal, a dedicated line, or the like. is there.

Next, ART, which is one of the category analysis methods used in the category analysis unit 33, will be described with reference to FIGS.
FIG. 4 is an explanatory diagram showing an example of category classification of operation data by the category analysis unit 33.
FIG. 5 is an explanatory diagram illustrating an example of the correlation of categories after classification by the category analysis unit 33.

  In the upper part of FIG. 4, the operation data A output from the two types of measuring instruments in both the normal operation period in which the water supply facility 1 is operating normally and the diagnosis period in which the operation state of the water supply facility 1 is diagnosed and An example of the time change of the operation data B is shown. Here, the operation data A and the operation data B are two kinds of data selected from data collected from each measuring instrument such as the turbidimeter 16a, the flow meter 17, the water temperature meter 18, and the pH meter 19 shown in FIG. Represents data. The operation data A and the operation data B are within the upper and lower limits of the alarm output in both the normal operation period and the diagnosis period.

  First, the category analysis unit 33 learns in advance the correlation between the operation data A and the operation data B during the normal operation period. At this time, as shown in the upper part of FIG. 4, as the correlation between the operation data A and the operation data B, (1) the operation data A is large, the operation data B is small, and (2) both the operation data A and the operation data B are Data groups showing three different correlations are extracted: small, (3) operation data B is large, and operation data A is small.

  In the lower part of FIG. 4, the category analysis unit 33 classifies the data groups indicating the correlation between the operation data A and the operation data B into the categories “1” to “3”, and these data groups are shown along with time changes. It is. In this example, in order to simplify the description, the case where the correlation between the operation data A and the operation data B in the category analysis is identified by the magnitude relationship of each operation data is shown. However, the present invention is not limited to this example, and the category analysis unit 33 may compare the difference between the operation data A and the operation data B with a predetermined threshold and perform category analysis on the difference.

  Next, an operation example of the category analysis unit 33 in the diagnosis period after the category analysis unit 33 learns the correlation between the operation data A and the operation data B shown in the categories “1” to “3” will be described. When the operation data A and the operation data B are input to the category analysis unit 33, the operation data A and the operation data B in the first first diagnosis period in the diagnosis period shown in the upper part of FIG. Similar to the characteristic with the number “2”. Therefore, the category analysis unit 33 classifies the data group indicating the correlation between the operation data A and the operation data B as the category number “2”.

  Subsequently, the correlation between the operation data A and the operation data B input in the second diagnosis period is large for each data, and is not similar to the characteristics of any of the learned categories “1” to “3”. For this reason, the category analysis unit 33 classifies the data group indicating the correlation between the operation data A and the operation data B in the second diagnosis period as a new category. As a result, as shown in FIG. 5, the category analysis unit 33 classifies the category “4” as a new category in addition to the learned categories “1” to “3”.

  Based on the analysis result by the category analysis unit 33, the state evaluation unit 34 has a normal state of the water supply facility 1 if the category classified by the category analysis unit 33 has the same characteristics as the operation data used for learning. And evaluate. On the other hand, when the characteristics of the operation data A and B are different from the operation data used for learning, that is, when the operation data A and B are classified into a new category by the category analysis unit 33, the state evaluation unit 34 The state of 1 can be evaluated as abnormal. 4 and 5 show an example in which the category analysis is performed on the two operation data A and B. In reality, the category analysis is performed on more operation data.

Next, processing by the analysis device 23 will be described.
FIG. 6 is a flowchart illustrating an example of a data processing method by the analysis device 23.

  First, the measuring instrument diagnosis unit 32 acquires analysis / diagnosis conditions from the analysis / diagnosis condition DB 36 (S1). The analysis / diagnosis conditions that can be acquired from the analysis / diagnosis condition DB 36 by the measuring instrument diagnosis unit 32 are shown in FIG.

  Next, the measuring instrument diagnosis unit 32 acquires analysis / diagnosis data from the process / water quality DB 35 (S2). The analysis / diagnosis data is operation data stored in the process / water quality DB 35, and is operation data acquired from the process / water quality DB 35 in accordance with a measuring instrument to be diagnosed by the measuring instrument diagnosis unit 32. When the measuring instrument diagnosis unit 32 acquires data for analysis / diagnosis, a data set that takes into account the time difference of water treatment may be assembled.

  For example, it takes several hours for the water received by the receiving well 10 from the intake facility 2 to reach the outlet of the settling basin 13. For this reason, even if pollution occurs in the water that has reached the outlet of the sedimentation basin 13, the water received by the landing well 10 at the same time is not necessarily contaminated. For this reason, the measuring instrument diagnosis unit 32 diagnoses the measuring instrument in the water supply facility 1 by assembling operation data representing the raw water quality that has gone back by the time difference from the time of measuring the quality of the precipitated treated water as a data set. In addition, the water quality of the water processed at the water supply facility 1 varies from day to day. For this reason, the period for which the measuring instrument diagnosis unit 32 goes back to acquire operation data may be in units of days.

  Moreover, in the water supply facility 1, unsteady operations such as construction and periodic maintenance and inspection of measuring instruments are performed. Data measured by a measuring instrument during unsteady work and stored in the process / water quality DB 35 is not suitable for diagnosis or category analysis. For this reason, during the period of unsteady work, an operator operates a data logger installed at a place where construction or inspection is performed, thereby stopping the collection of data output from the measuring instrument by the data logger. Also good. Thereby, the storage of the data in the process / water quality DB 35 in the implementation period of the unsteady work can be stopped.

  In addition to the operation of stopping data collection by the data logger, the worker operates the data logger during the operation period of the non-stationary work, thereby transmitting a working signal to the analyzer 23. Also good. In the analysis device 23, the process / water quality DB 35 is set so that the data stored in the process / water quality DB 35 is not analyzed during the operation signal being transmitted and during the stable waiting time after the work signal is stopped. Flag the corresponding data stored in. And it is desirable for the analyzer 23 to exclude the flagged data when performing diagnosis or category analysis. In this way, for example, abnormal values of data that may be output by the measuring instrument due to construction or inspection performed during a predetermined period can be reliably removed from the analysis target.

  After step S2, the measuring instrument diagnosis unit 32 diagnoses the measuring instrument (S3). Here, the description will be continued assuming that the measuring instrument diagnosis unit 32 diagnoses one measuring instrument (a) among a plurality of measuring instruments installed in the water supply facility 1.

Here, analysis / diagnosis conditions will be described.
FIG. 7 is a table showing an example of data items indicating the analysis / diagnosis conditions that can be acquired from the analysis / diagnosis condition DB 36 by the measuring instrument diagnosis unit 32 and the category analysis unit 33. The table shown in FIG. 7 includes fields of classification, target, measurement / inspection item, and defect example.

In the classification field, items acquired by either category analysis by the category analysis unit 33 or measurement instrument diagnosis by the measurement instrument diagnosis unit 32 are shown.
The target field is an item indicating a category analysis target and an instrument diagnosis target according to the classification item in the classification field. For example, it is indicated that the target of category analysis is “abnormal coagulation sedimentation process”, and the target of measuring instrument diagnosis is “precipitation water turbidity indicated by the measured value of turbidimeter 16b”.

  The measurement / inspection item field is an item indicating a measurement item or an inspection item used for diagnosis or category analysis for each target item in the target field. For example, the measurement / inspection items used for category analysis are “flocculating agent injection rate”, “alkaline agent injection rate”, etc., and the measurement / inspection items used for measuring instrument diagnosis are “raw water turbidity”, “filtration” It is shown to be “water turbidity” or the like.

  The defect example field is an item indicating a defect example that may occur for each measurement / inspection item in the measurement / inspection item field. As an example of a defect that can be detected by category analysis, for example, if the flocculant injection rate is abnormal, that is, if the amount of flocculant injected is abnormal, the occurrence of clogging of the pipe is indicated. Note that there are no examples of defects that can be detected by instrument diagnosis.

Next, the “aggregation-precipitation process abnormality” that is an analysis target when the classification item is category analysis will be described.
The coagulation sedimentation process abnormality is an abnormality that can be determined by directly measuring the turbidity of the precipitation-treated water using the turbidimeter 16b. However, the category analysis unit 33 performs category analysis using the items shown in the table shown in FIG. As already described with reference to FIG. 4 and FIG. 5, when the category analysis unit 33 performs the category analysis, the category data is classified in advance by using the performance data indicating the normal processing state in the normal operation period as the teacher data. deep. Then, as a result of the category analysis unit 33 classifying the data to be diagnosed into categories during the diagnosis period, if a new category appears, it indicates that there is an abnormal cause in water treatment or raw water quality that did not exist in the past.

  On the other hand, it is necessary to determine whether each item used for category analysis itself is a normal value. For example, although the turbidimeter 16b is not defined, there is a possibility that the measuring instrument does not show a correct value due to the cause shown in the failure example field of FIG. 7 in another measuring instrument. As a condition for detecting such a defect, the category analysis unit 33 acquires a setting value for each item from the analysis / diagnosis condition DB 36. For example, in the case of precipitation-treated water turbidity, the category analysis unit 33 sets the turbidity measured at another measurement point or the injection rate of the flocculant directly related to the removal of the turbidity as a set value. Thus, the category analysis unit 33 can determine whether each item used for category analysis is a normal value.

Returning to the description of the processing in FIG.
After step S3, the measuring instrument diagnosis unit 32 determines whether or not the measuring instrument (a) is abnormal (S4). When it is determined that there is no abnormality in the measuring instrument (a) (NO in S4), the process proceeds to the subsequent step S6.

  On the other hand, when it is determined that the measuring instrument (a) is abnormal (YES in S4), the category analysis unit 33 removes the data output from the measuring instrument (a) (hereinafter referred to as “measuring instrument (a Category analysis is also performed (also referred to as “data set excluding“) ”(S5). Accordingly, the category analysis unit 33 uses the data set obtained by removing the data output from the measuring instrument (a) diagnosed as abnormal by the measuring instrument diagnosis unit 32 from the data read from the process / water quality DB 35. Analysis can be performed.

  Therefore, in the process of step S5, the category analysis unit 33 removes the measuring instrument (a) from the item group initially registered in the analysis / diagnosis condition DB 36 as a data set used for category analysis. Is used. At this time, tuning with the data set excluding the measuring instrument (a) is required. For this reason, the category analysis unit 33 performs tuning using only the corresponding items excluding the measuring instrument (a) among the teacher data used when tuning all the data items.

  After step S4 NO or step S5, the category analysis unit 33 uses a data set including the measuring instrument (a), that is, a data set that is a set of data output from the measuring instrument of the initially registered item group. Then, category analysis is performed (S6). As described above, through the processing of steps S5 and S6, the category analysis unit 33 outputs a normal or abnormal category analyzed according to each data set. For example, the number of categories classified when the measuring instrument (a) is normal is different from the number of categories classified when the measuring instrument (a) is abnormal. For this reason, the state evaluation part 34 can judge the normality or abnormality of a measuring device (a) by comparing the number of categories.

  Next, the state evaluation unit 34 selects the contents to be displayed on the display unit 37 in order to give a work instruction to the operator according to the result of the instrument diagnosis and category analysis (particularly, whether or not a new category is generated). (S7). For example, if the result of the category analysis using the data set including the measuring instrument (a) indicates an abnormality, and the result of the category analysis using the data set excluding the measuring instrument (a) indicates normal. In this case, it is understood that an abnormality has occurred in the measuring instrument (a). Then, regarding the measuring instrument (a), the display contents for instructing the operator that a new category has occurred and confirming the state of the measuring instrument (a) are selected by the state evaluation unit 34.

  The display unit 37 outputs display contents indicating a work instruction to the operator selected by the state evaluation unit 34 (S8), and ends this process. The operator can confirm the state of the measuring instrument (a), for example, according to the work instruction indicated in the display content.

FIG. 8 is a table showing an example of classification items for classifying the contents displayed by the state evaluation unit 34 on the display unit 37.
The table shown in FIG. 8 includes items of category analysis, instrument diagnosis: no abnormality, and instrument diagnosis: abnormality.

  In the category analysis item, an item “with new category or with abnormality” and an item “with new category or without abnormality” are stored. In addition, the items of “no abnormality of measuring instrument” and the item of “abnormality of measuring instrument” are stored in the measuring instrument diagnosis item. Furthermore, an item “data usage of measuring instrument (a)” and an item “data not using measuring instrument (a)” are stored.

  The table shown in FIG. 8 stores, for example, five types of classifications AA, A, B, C, and D as classification items corresponding to combinations of category analysis items and measuring instrument diagnostic items. Each classification corresponds to the display content output to the display unit 37 shown in FIG. For this reason, the presence / absence of abnormality in the data indicated by the result of the instrument diagnosis by the instrument diagnosis unit 32, the presence / absence of a new category indicated by the result of the category analysis by the category analysis unit 33, or whether the instrument is abnormally categorized By such combination, the classification item indicating the display content of the display unit 37 is selected.

  FIG. 9 is a table showing an example of contents displayed on the display unit 37 according to the classification items. FIG. 9 shows that the content displayed on the display unit 37 is defined for each of the classifications AA, A, B, C, and D.

The classification AA represents the display content that prompts a direct response because it is most likely that an abnormality has occurred in the category analysis target.
Class A may indicate an abnormality in the measuring instrument, but it is abnormal in the category analysis performed in combination with other data. Therefore, priority is given to the confirmation of the measuring instrument, and the display content indicates a response thereafter.

The classification B depends on the contribution of the measuring instrument (a) in the category analysis, and thus the reliability of the determination of abnormality is low.
The classification C represents the display content instructing to perform category analysis again after checking the measuring instrument because there is no abnormality in the new category.
The classification D represents the display contents that output that there is no particular problem at present.

  As described above, the state evaluation unit 34 selects information that gives an instruction to give priority to a countermeasure for the measuring device (a) diagnosed as abnormal by the measuring device diagnosis unit 32. Thereby, the operator who operates the analyzer 23 can preferentially implement the countermeasure for the measuring instrument (a) diagnosed as having an abnormality based on the instruction displayed on the display unit 37. For this reason, the accuracy of the operation data to be analyzed by the analysis device 23 is increased, and the operation management of the water facility 1 can be appropriately performed.

  In the plant diagnosis data preprocessing / analysis system 100 according to the first embodiment described above, the occurrence of an abnormality in the measuring instrument or the water supply facility 1 by the two-stage configuration of the diagnosis of the measuring instrument and the category analysis of the operation data. The presence or absence can be detected. Then, based on the diagnosis result, the procedure to be taken by the operator with respect to the measuring instrument, equipment, or process (operation) is shown, so that the operator can take appropriate measures, and as a result, stable operation of the water supply facility 1 is realized. There is an effect that can be done. Moreover, since appropriate measures are taken by the operator based on the display content output on the display unit 37, the accuracy of data necessary for diagnosing the water facility 1 is increased, and the measuring instrument or the water facility 1 The accuracy of detecting defects can also be improved.

  Furthermore, conventionally, in the water treatment process, it is necessary to consider the correlation of data output from a plurality of measuring instruments in which the number of the same sensors is small and it is difficult to create a correlation formula or an evaluation formula. However, by using a statistical method as in the present embodiment, it is possible to detect the diagnosis of the measuring instrument (particularly, the presence or absence of drift of the measuring instrument), and promptly show the diagnostic result of the measuring instrument to the operator. be able to. Further, since the statistical method is used, the plant diagnosis data preprocessing / analysis system 100 according to the present embodiment can be introduced quickly and easily.

[Modification of First Embodiment]
In the first embodiment described above, only the measuring instrument (a) is a target for abnormality diagnosis of the measuring instrument. However, abnormality diagnosis is performed on a plurality of measuring instruments or all measuring instruments used for category analysis. Also good. For example, after the turbidimeter 16a is first selected as the measuring instrument (a) and the process shown in FIG. 6 is executed, the turbidimeter 16b is selected as the measuring instrument (a) and the process shown in FIG. Execute. Hereinafter, similarly, all the measuring instruments constituting the water supply facility 1 are selected, and the processes shown in FIG. 6 are executed in order.

  As a result of selecting the measuring instrument (a) and executing the process as described above, if abnormality is observed in a plurality of measuring instruments, the state evaluation unit 34 performs the corresponding classification, that is, the process of step S7 to increase the urgency. Selects the display contents. At this time, the state evaluation unit 34 selects display contents that allow the operator to check the state and accuracy of the measuring instrument. However, the state evaluation unit 34 may use the timing of calibration of the measuring instrument and the difference between measured values before and after calibration (degree of drift) as a determination index in order to prioritize the confirmation work to be performed by the operator. And the measured value output immediately after calibrating a measuring device may be used as teacher data. Moreover, the analyzer 23 may raise the calibration frequency with respect to the measuring device in which the occurrence of drift is predicted in advance.

  In the first embodiment, the instrument diagnosis unit 32 and the category analysis unit 33 perform analysis by different statistical methods, but they are unified into one statistical method (for example, ART). Also good. For example, in the first embodiment, the diagnosis of the measuring instrument by the measuring instrument diagnosing unit 32 determines whether or not there is an abnormality depending on whether or not the data output by the measuring instrument exceeds a certain set value. The presence / absence of abnormality may be determined by performing category analysis on the combination of the above data by ART. As a result, one type of engine is used for the statistical method, and the format for preprocessing and organizing the input data for analysis is unified, so that the system configuration can be simplified.

  In addition, when an abnormality is found in the measuring instrument, the category analyzing unit 33 performs category analysis by excluding data (measurement items) output by the measuring instrument. As a result, the state evaluation unit 34 may cause the display unit 37 to display the result of comparing the analysis accuracy when the measurement items are reduced, for example, the detection probability of abnormality by the teacher data with the case where the measurement items are not reduced. .

[Second Embodiment]
Next, a plant diagnosis data pre-processing / analysis system according to a second embodiment of the present invention will be described. The plant diagnosis data preprocessing / analysis system according to the present embodiment can perform category analysis by excluding the influence of drift of the measuring instrument from the data output from the measuring instrument, for example.

  FIG. 10 is a flowchart illustrating an example of a data processing method by the analysis apparatus 23 according to the second embodiment. Note that the processing from step S11 to step S15 in FIG. 10 is the same as the processing from step S1 to step S5 in FIG. 6 shown in the first embodiment, and thus detailed description thereof is omitted.

  If it is determined in step S14 that there is an abnormality in the measured value indicated in the data output from the measuring instrument (a) (YES in S14), the category analysis by the data set excluding the measuring instrument (a) ( In parallel with S15), the processes shown in steps S16 and S17 are performed. In this process, the category analysis unit 33 calculates a correction value for correcting the measurement value of the measuring instrument (a) (S16), and performs category analysis using the correction value of the measuring instrument (a). At this time, the category analysis unit 33 corrects the measurement value indicated in the data output from the measurement device (a) diagnosed as abnormal by the measurement device diagnosis unit 32 from the characteristics of the measurement device (a). Diagnosis by category analysis is performed using a data set in place of values.

  Here, the correction value of the measuring instrument (a) calculated by the category analysis unit 33 goes back to the time when the measuring instrument (a) does not show any abnormality, and uses the value (x0) at that time as it is. Further, as a correction value of the measuring instrument (a), for example, using a value (x) when the measuring instrument (a) indicates abnormality, the result of obtaining x0 ± | x−x0 | A value within the range of the swing width may be used as the correction value on the assumption of the swing width. In this way, if there is an abnormality in the measurement value of the measuring instrument (a), the category analysis unit 33 calculates the correction value, so that the category analysis unit 33 can set the initial input items set in the analysis / diagnosis condition DB 36. It is possible to perform category analysis by arranging all of the above.

  Thereafter, the category analysis unit 33 performs category analysis using a data set including the measuring instrument (a) (S18). Then, the state evaluation unit 34 selects contents to be displayed on the display unit 37 in order to give a work instruction to the operator according to the result of the instrument diagnosis and the category analysis (particularly, whether or not a new category is generated). (S19). Here, the classifications selected by the state evaluation unit 34 are, for example, classifications A and C in the table of FIG. Then, the display unit 37 outputs a work instruction to the operator selected by the state evaluation unit 34 (S20), and ends this process.

  In the plant diagnosis data preprocessing / analysis system 100 according to the second embodiment described above, if there is a possibility that the measurement value is abnormal, an abnormality has occurred in the measuring instrument by using the correction value of the data. Even if it is a case, it is possible to perform category analysis with appropriate data to diagnose the water supply facility 1. For this reason, the accuracy of data necessary for diagnosing the water supply facility 1 is increased, and the accuracy of device / process diagnosis can be improved. Further, the same effects as those of the plant diagnosis data preprocessing / analysis system 100 according to the first embodiment described above can be obtained.

  In addition, although the 1st and 2nd embodiment mentioned above was taken as the example applied to the water supply facility 1, other water treatment facilities, such as a sewerage facility and a seawater desalination facility, a chemical plant, a power plant, a thermal power plant It may be applied to various plants such as nuclear power plants. Further, the processing target of the analyzer 23 is not limited to the plant. The analysis device 23 can be used for data preprocessing in a system that is required to increase the accuracy of data in order to analyze a state using a measurement value.

Further, the present invention is not limited to the above-described embodiments, and it is needless to say that other various application examples and modifications can be taken without departing from the gist of the present invention described in the claims.
For example, the above-described embodiment is a detailed and specific description of the configuration of the apparatus and the system in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to one having all the described configurations. In addition, a part of the configuration of the embodiment described here can be replaced with the configuration of the other embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. Is possible. Moreover, it is also possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.
Further, the control lines and information lines indicate what is considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. Actually, it may be considered that almost all the components are connected to each other.

  DESCRIPTION OF SYMBOLS 1 ... Water supply facility, 2 ... Water intake facility, 3 ... Water purification treatment facility, 22 ... Control LAN, 23 ... Analyzing device, 31 ... Communication I / F, 32 ... Measuring instrument diagnostic part, 33 ... Category analysis part, 34 ... State evaluation 35, process / water quality DB, 36 ... analysis / diagnosis condition DB, 37 ... display unit, 38 ... internal bus, 100 ... data pre-processing / analysis system for plant diagnosis

Claims (7)

A first database that stores, in time series, data output from equipment constituting the plant as the plant is operated, and data output from a measuring instrument that measures a processing target of the plant;
A second database in which conditions for analysis or diagnosis are defined;
Analyzing the data read from the first database based on the condition read from the second database by a first statistical method, thereby diagnosing the presence or absence of abnormality of the measuring instrument;
Analyzing the diagnostic result of the presence or absence of abnormality of the measuring instrument and the data read from the first database based on the condition read from the second database, the performance of the facility A second diagnostic unit for diagnosing the presence or absence of abnormality;
A work instruction is given to an operator of the plant based on an evaluation result obtained by evaluating a state of the measuring instrument and the plant based on a diagnosis result of whether or not the measuring instrument is abnormal and a diagnostic result of whether or not the equipment is abnormal in performance. A state evaluation unit for selecting information for
A data processing apparatus comprising: a presentation unit that presents the information selected by the state evaluation unit. The second diagnostic unit uses the data set obtained by removing data output from the measuring instrument diagnosed as abnormal by the first diagnostic unit from the data read from the first database. The data processing apparatus according to claim 1, wherein diagnosis is performed by a general technique. The second diagnostic unit is a data set in which the measurement value indicated in the data output from the measuring instrument diagnosed as abnormal by the first diagnostic unit is replaced with a correction value calculated from characteristics of the measuring instrument. The data processing apparatus according to claim 2, wherein diagnosis is performed using the second statistical technique. The data processing apparatus according to claim 3, wherein the state evaluation unit selects information that gives an instruction to give priority to a countermeasure for the measuring instrument diagnosed as abnormal by the first diagnosis unit. The first statistical method or the second statistical method is a method using an adaptive resonance theory for classifying the data read from the first database into a category, or a cluster analysis method similar to the adaptive resonance theory. The data processing apparatus according to claim 4. The plant is a treatment plant for performing water purification, sewage treatment, water intake, water collection, or water supply and distribution,
The data processing device according to claim 5, wherein the measuring device is a water quality measuring device that measures the quality of water as the processing target. Based on the conditions read from the second database in which conditions for analysis or diagnosis are defined, the data output from the equipment constituting the plant as the plant operates and the processing target of the plant are measured. Diagnosing the presence or absence of abnormality of the measuring instrument by analyzing the data read from the first database storing the data output from the measuring instrument in time series by a first statistical method;
Analyzing the diagnostic result of the presence or absence of abnormality of the measuring instrument and the data read from the first database based on the condition read from the second database, the performance of the facility A step of diagnosing whether there is an abnormality,
A work instruction is given to an operator of the plant based on an evaluation result obtained by evaluating a state of the measuring instrument and the plant based on a diagnosis result of whether or not the measuring instrument is abnormal and a diagnostic result of whether or not the equipment is abnormal in performance. Selecting information for, and
Presenting the selected information. A data processing method.

Images