JP3176692B2 - Process operation support system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for supporting the operation of a process such as a plant, and more particularly to a system which can provide information representing the state of a process in an expression which can be easily understood by an operator. It relates to a process operation support system.

[0002]

2. Description of the Related Art In plants such as chemical plants,
Because process fluctuations in the plant have a significant effect on the final product, monitoring of process trends and changes, monitoring of the abnormalities of the instruments and sensors that make up the control loop, and whether the process is operating in accordance with the operation pattern Monitoring is required. Normally, a system for monitoring the operation of a plant is configured to output an alarm when there is an abnormality in process data as compared with a predetermined reference. As an element for outputting an alarm, for example, upper and lower limits, a change rate, a deviation, and the like regarding data representing a process state are used. The plant operator performs an operation on the plant by monitoring them or by generating an instantaneous alarm in control cycle units.

[0003] However, in many cases, these alarms are already slow after they are generated. for that reason,
It is necessary to catch signs of plant abnormalities and take measures to avoid abnormal situations and accidents in advance. For this monitoring, for example, it is necessary to accurately grasp the characteristics of the time-series data representing the state of the process. Therefore, conventionally, in operation of the plant, in addition to the alarm processing, a trend graph of various process data is displayed on a display device in parallel with the alarm processing, and the operator monitors the trend graph. That is, the operator operates the plant by constantly monitoring these trend graphs to detect trends and changes in the process. This trend graph monitoring is especially useful when starting up a plant.
This is a very important task for the operator when the state of the plant changes greatly, such as at the time of a grade change or shutdown.

[0004] On the other hand, plant operation is performed based on a plant operation plan that has been previously considered as an operator's personal experience and design knowledge. Also, if the current operating condition of the plant becomes abnormal, the operator recalls similar cases or searches for records in the past cases that he or she has experienced, and employs the situation at that time. Driving with reference to the operation performed.

[0005]

By the way, in order to catch a precursor of a plant abnormality, it is necessary to read from the trend graph what kind of process change has occurred in the past. In this case, whether or not a change in a process on the trend is determined to be an abnormal factor depends on individual differences among operators. However, with the conventional technology,
There was no consideration on this point, and the reading of a change that was a precursor of an abnormality from the trend graph was judged by the subjective personality of the operator. For this reason, the judgment is influenced by the skill level, attention, physical condition, and the like of the operator, and there has been a problem that the precursor phenomenon may be overlooked or mistaken.

In this type of process, not only is sensor information indicating the state of the process received, but also, based on the sensor information, whether or not the process is abnormal, and where an abnormal location is determined in the event of an abnormality, are determined. It is preferable to make inferences. To do this, it is necessary to gather the knowledge of experienced operators, make rules, and build inference knowledge.

[0007] However, when it is attempted to form a rule using various operator's knowledge from the process, the combination of the operator's knowledge with the actual measurement data is not uniform.
Unlike measured values, they are vaguely expressed. This complicates the rules in a very wide variety. For this reason, there is a problem that construction of rules is not easy.

A first object of the present invention is to extract characteristics of information indicating a state of a process without being influenced by personal conditions of a monitoring operator, and to accurately grasp the state of a process. It is to provide a process operation support system.

A second object of the present invention is to express the characteristics of the process so that the operator can easily understand it, and to combine this expression with the operator's knowledge as it is, making it easy to construct rules. It is to provide a process operation support system that can be used for

[0010]

According to a first aspect of the present invention, to attain the first object, data representing a state of a process is fetched and converted into data which can be processed by a subsequent processing means. Process data capture processing means for processing, symbol processing means for generating symbol data representing the process state symbolically based on data representing the process state, and symbol data for accumulating the symbol data Storage means, display processing means for extracting data in a specified range from the encoded data storage means and outputting the display data as display data, and input / output means for displaying the display data on a screen and receiving an external instruction. Wherein the symbol processing means obtains a degree of conformity between data representing a process state and symbol data representing the process state symbolically. Has a symbol data storage means, the process operation support system, characterized in that for storing the fitness with the symbol data is provided.

[0011] The display processing means may display the symbolic representation of the encoded data as well as the degree of matching.

The symbolization processing means includes a symbolization processing dictionary having standard vector data and symbol expression data for symbolically expressing the characteristics of the standard vector data, and a vector series obtained by linearly approximating the input data. Produces
For this vector sequence, a symbolization processing unit that searches for similar standard vector sequences using a dictionary for symbolization processing, extracts symbol expression data corresponding to the most similar standard vector, and performs symbolization processing, It has a similarity calculation processing unit that calculates a similarity between a broken line approximation vector series of data and a standard vector series, and can output the similarity as a degree of conformity to the input data of the symbolized data. .

According to the present invention, in order to achieve the second object, in addition to the above-mentioned structure, a knowledge acquisition input means for generating an inference rule described using the encoded data and its fitness is provided. And

[0014] The knowledge acquisition input means associates the representation of the symbolized data representing the state of the process with a membership function of fuzzy inference, and associates the degree of adaptation with the degree of adaptation to the membership function. And a function of generating a description of an IF THEN rule.

Further, the knowledge acquisition input means has in advance a graph format represented by a box defining an event occurring in the process, and a connection symbol defining a connection direction between the box and the propagation direction. For
The designation of belonging to the IF part and the designation of belonging to the THEN part are received. For the box belonging to the IF part, a name indicating the location where the event occurred, symbolized data indicating the content of the event, its fitness, In addition, at least an input of an application condition of the degree of conformity is received, and an input of a result of inference is received in a box belonging to the THEN unit, and a connection relationship between the boxes in which an event is defined, And a function of receiving at least an input of a connection symbol designating the propagation direction of those events.

Here, connection symbols for designating the transmission direction of events are OR connections for boxes connected in parallel and AND connections for boxes connected in tandem.
Each connection can be defined.

The knowledge acquisition input means generates an IF condition from a box belonging to the IF section among the above boxes based on a logical relationship of the connection symbols thereof, and a THEN section logically connected to these boxes. The function of generating the conclusion of the THEN part based on the box belonging to and generating the description of the inference rule can be provided.

The present invention may further include a knowledge base for storing the inference rules as knowledge, and an inference engine for inferring the state of the process using the knowledge-based inference rules. In addition, a configuration may be provided that includes an abnormal treatment guide output unit that outputs a guide for the abnormal treatment via the input / output unit based on the inference result.

Further, the present invention provides a knowledge base for storing inference rules as knowledge, an inference engine for inferring the state of a process using the inference rules of the knowledge base, and information representing the state of the process. A configuration may be provided which further includes a display for inquiring about undetectable information and prompting the input, and a process status inquiry means for receiving an input of the corresponding information.

Here, the knowledge base is an IF described by coded data representing the state of the process and its fitness.
A process status inquiry rule, which has a section and a THEN section corresponding to the information indicating the state of the process, which is necessary for determining the state and which describes information that cannot be detected by the sensor, is stored. Configuration.

The inference engine can have a function of executing inference based on knowledge-based inquiry rules and outputting necessary inquiry items.

The process status inquiry means may have a function of displaying the inquiry items output from the inference engine via the input / output means.

Further, the present invention can be configured to further include a symbol integration unit for naming new symbol data for a plurality of symbol data arranged in time series.

[0024]

According to the present invention, data representing the state of a process is fetched by the process data fetching means.
The data is converted into data that can be processed by the symbol processing means in the subsequent stage. The symbolization processing means generates symbolized data that symbolically represents the characteristics of the data representing the state of the process. In this symbolization processing means, for example,
Generate a vector series by linear approximation using a filter,
Using a dictionary, this vector sequence is integrated into a vector sequence composed of fewer vectors, and the sequentially obtained vector sequences are compared with standard vector sequences registered in the dictionary, respectively, and a name of a similar standard vector sequence is obtained. Extract as a symbol.

The encoded data is accumulated as an encoded data string in an encoded data storage means. For example, since many plant data changes in a time series, this encoded data string is accumulated in a time series, for example. Of course, spatially distributed data
For example, they can be stored in a spatial sequence such as a tree structure.

From the stored data, the designated range is extracted from the symbolized data storage means by the display processing means and output as display data. The range to be taken out can be set arbitrarily. For example, data of one sampling cycle including the present, data of one or several sampling cycles including a certain date and time in the past, and the like can be designated. Further, the latest data may be automatically updated and output.

The input / output means issues an instruction to output the encoded data, an instruction to search the encoded data, an instruction to operate the process, and the like, and displays the display data on a screen.
Since the data displayed here is, for example, symbolized data such as a character string, the operator can easily and easily change the state of the process based on the meaning of the symbol.
It can be grasped accurately without causing individual differences.

Further, by using the symbolized operation instructions, the ambiguous instructions of the operator can be applied to the process operation without individual differences.

Further, according to the present invention, by providing the knowledge acquisition input means, it is possible to generate an inference rule described using the coded data and its fitness. That is, for example, the knowledge acquisition input means associates the representation of the symbolized data representing the state of the process with a membership function of fuzzy inference, and associates the degree of adaptation with the degree of adaptation to the membership function. And generates an IF THEN rule description.

The knowledge acquisition input means prepares in advance a graph format represented by a box defining an event occurring in the process, and a connection symbol defining a connection direction between the box and a propagation direction thereof. By filling in the necessary information in this, FTA (Foult Trerance Anari
sys) can be formed. The designation of belonging to the IF section and the designation of belonging to the THEN section are accepted for the box. For the box belonging to the IF section, at least the input of the name indicating the location where the event occurred, the encoded data indicating the content of the event, its fitness, and the application condition of the fitness is accepted. , Accept the input of the result of the inference. In addition, with respect to boxes in which events are defined, at least input of a connection symbol that specifies a connection relationship between them and a propagation direction of those events is received. In this case, connection symbols specifying the propagation direction of an event are defined as an OR connection for connecting boxes in parallel and an AND connection for connecting boxes in tandem.

Thus, the knowledge acquisition input means generates IF conditions from the boxes belonging to the IF section among the above boxes based on the logical relationship of their connection symbols, and is logically connected to these boxes. , Based on the boxes belonging to the THEN part, the conclusion of the THEN part can be generated to generate a description of the inference rule.

[0032]

Embodiments of the present invention will be described below with reference to the drawings.

One of the features of the present invention is that the features of process data from a plant or the like can be shown to an operator using symbolic expressions. Therefore, prior to the description of the embodiment relating to the process operation support system, a symbol processing system used as a symbol processing means in the support system will be described. Here, the encoding processing system will be described as an independent encoding device, but it goes without saying that it can be configured as one component of a process operation support system described later.

The details of the first embodiment of the encoding apparatus according to the present invention will be described with reference to FIGS.

FIG. 1 is a block diagram showing the configuration of the encoding apparatus. In the figure, reference numeral 110 denotes a communication device, which receives an identification number of a pattern data to be processed, a sampling cycle, a processing target section, a conversion coefficient, and the like through a communication network. Reference numeral 120 denotes a data storage device which stores information indicating a state of a process. As information representing the state of the process, for example, in the field of process control such as a power plant, a chemical plant, and a water treatment plant, time-series data such as temperature, pressure, and flow rate of each part of the process or data of a spatial distribution pattern. However, in the process of processing data in fields such as finance, securities, and distribution,
Time-series data or pattern data such as economic indicators, sales information, and management information are stored.

An input processing unit 130 is activated by a signal from the communication device 110. Input processing unit 13
0, based on the information of the identification number, the sampling period, and the processing section sent from the communication device 110, the corresponding data is extracted from the processing target section at a given sampling period from the data storage device 120, and Send to 140.

The polygonal line approximation processing unit 150 includes a memory 140
Is converted into a vector series of line segments constituting the polygonal line, and the result is sent to the integration processing unit 160. 155 is a filter storage unit,
A feature extraction filter used in the polygonal line approximation processing unit 150 is stored.

Reference numeral 160 denotes an integration processing unit which integrates a series of broken line vectors obtained by the broken line approximation processing unit 150 and sends the result to the dictionary search processing unit 170. Reference numeral 165 stores integration parameters used in the integration processing unit 160.

The dictionary search processing unit 170 includes the integrated processing unit 16
0 and the standard pattern storage unit 1
75 with the vector sequence of events stored in
Search for a vector series of correspondence candidates. Reference numeral 175 denotes a standard pattern storage unit which stores a standard pattern corresponding to an event name for identifying an input pattern as a vector sequence approximated by a broken line and a name for each vector sequence. The standard pattern and its corresponding name are prepared in advance by the system. The user may register an arbitrary name. Also, an arbitrary pattern may be registered together with its name as a standard pattern.

Reference numeral 180 denotes a similarity calculation processing unit which calculates a similarity and a scale between the vector series of the correspondence candidate obtained by the dictionary search processing unit 170 and the vector series of events stored in the standard pattern storage unit 175. The calculation is performed, and the event name, the similarity, the scale, the section of the corresponding data, and the like of the standard pattern whose scale is within a certain range and the similarity is a certain value or more are sent to the memory 190.

The processing result stored in the memory 190 is transmitted to another information processing system or the like via the communication device 110.

The encoding device according to the present embodiment includes a central processing unit,
It is configured by a computer system having a main storage device, an auxiliary storage device, various interface devices, and the like. Various functions of the encoding device of the present embodiment are realized by the central processing unit executing a program stored in the main storage device, and the following operations are executed. As a hardware system for realizing a symbol or a device, for example, a system as shown in FIG. 19 can be used.

Next, the operation of each section of the encoding apparatus based on the above configuration will be described in detail.

FIG. 2 is a flowchart showing the operation of the broken line approximation processing section 150. When roughly classified, a pattern division process 210 for calculating the product unevenness of the pattern f (t) of the input data and the feature extraction filter W2 (x) to calculate the unevenness of the pattern and obtaining a point τp at which the pattern is divided into polygonal lines.
Using the feature extraction filter W0 (x), an average value of data in the vicinity of the division point of the pattern, and an estimation process 220 of obtaining an estimated value fe (τp) of the data at the division point.
, End-end estimation processing 230 for obtaining estimated values at both ends from regression lines at both ends of a data section, and vector series calculation processing 240 for calculating a vector series when a pattern is approximated by a broken line using the dividing points and the estimated values at both ends. It is composed of

In the pattern division processing 210, data is normalized by multiplying the input data by the conversion coefficient transmitted through the communication network, and the processed data f (t) (t
= 0, 1,..., T). Next, for the input pattern data f (t) (t = 0, 1,..., T), the feature extraction filter W2 stored in the filter storage unit 155 is used.
Sum of products calculation using (x)

[0046]

(Equation 1)

Thus, the characteristic amount g2 (t) indicating the degree of unevenness
Is calculated. Here, a is a constant indicating the spread of the filter. With respect to the example of the input pattern shown in FIG.
FIG. 3B shows the result of the product-sum calculation performed by the feature extraction filter W2 (x) shown in FIG. Calculating the division point τp (p = 1, 2,..., K) of the input pattern by extracting points having positive and negative extreme values whose absolute values are larger than appropriate constants from the calculation results shown in Equation 1. Can be. In the example of FIG. 3B, τp (p = 1, 2,..., 9) is obtained as a division point.

In the estimating process 220, at the division point τp obtained by the pattern dividing process 210, the feature extraction filter W 0 stored in the filter storage unit 155 is used.
Sum of products calculation using (x)

[0049]

(Equation 2)

As a result, the characteristic value g0 (τp) representing the average value
Is calculated. FIG. 5B shows an example of the feature extraction filter W0 (x). Estimated value fe of pattern at division point τp
(τp) is calculated by using a feature amount g2 (τp) indicating unevenness and a feature amount g0 (τp) indicating an average value.

[0051]

[Mathematical formula-see original document] fe ([tau] p) = W0 (0) * g0 ([tau] p) + W2 (0) * g2 ([tau] p) (Expression 3)

In the both ends estimation process 230, the estimated values of both ends of the input pattern are obtained. The estimated values (τ1, fe (τ1)) of the division points obtained in the estimation process 220 and (τ
k, fe (τk))

[0053]

(Equation 4)

[0054]

(Equation 5)

Is calculated by

Equations 4 and 5 show that t = 0 and t = T respectively.
, The estimated values f at both ends of the input pattern
e (0) (= h1 (0)) and fe (T) (= h2 (T)) are obtained.

In the vector series calculation process 240,
The point sequence (0, fe (0)) and (τp, fe (τp)) (p = 1,
2,..., K) and (T, fe (T)), a vector sequence Ai = (Pi, Qi) (i = 1, 2,.
k + 1),

[0058]

(Equation 6)

Is obtained by An example of the obtained vector sequence is shown in FIG. Each vector is given the name of the closest standard vector (slope gradient) among the standard vectors shown in FIG. That is, the inner product of each vector and the standard vector is determined, the standard vector having the largest inner product is selected as the closest standard vector, and the name is given. Each of the named vectors is shown in FIG.
The data is transmitted to the integration processing unit 160 as data as shown in (d).

Next, details of the processing in the integration processing section 160 will be described. The integration processing is performed when the processing result of the polygonal line approximation processing unit 150 includes two or more vectors. The process includes three steps as shown in FIG. 710 is a step of integrating adjacent vectors having the same name, 720 is a step of integrating short “balanced” vectors, and 730 is a step of integrating short vectors other than “balanced”. Hereinafter, the processing in each step will be described in detail.

(1) Step 710: Adjacent vectors having the same name are integrated. For example, in the case of FIG.
Two consecutive "balanced" vectors are combined into one "balanced" vector.

(2) Step 720: A vector given the name "balance" and having a length equal to or less than a predetermined value is integrated with other vectors.

(2-1) When a vector having a variation width equal to or smaller than a certain value stored in the integrated parameter storage unit 165 is located before and after, the vector is integrated with the vector. For example, FIG.
In case (b), the short "equilibrium" vector is preceded by the short "descent" vector, so that both are integrated to generate a new "descent" vector.

(2-2) If the angle between the preceding and succeeding vectors is equal to or smaller than a certain value stored in the integrated parameter storage unit 165, the vectors are integrated. For example, in the case of FIG. 8C, since cos (θ1)> cos (θ2), the “equilibrium” vector is integrated with the previous “descent” vector to form a new “descent” vector.

(3) Step 730: A vector having a name other than “equilibrium” and a small change width is integrated with other vectors.

(3-1) When a short vector whose change width is equal to or smaller than a certain value stored in the integrated parameter storage unit 165 is before and after, the vector is integrated with the vector. For example,
In the case of FIG. 9A, a short “fall” vector and a short “rise” vector are integrated to generate a new “balanced” vector.

(3-2) If the angle between the preceding and succeeding vectors is equal to or smaller than a predetermined value stored in the integrated parameter storage unit 165, the vectors are integrated. For example,
In the case of FIG. 9B, since cos (θ1)> cos (θ2), the short “rise” vector is integrated with the previous “fall” vector to generate a new “balanced” vector.

In the above cases (2) and (3), the name of the newly generated vector is changed to the name of the standard vector (FIG. 6) having the smallest angle with the vector. FIG. 4A illustrates the vector sequence after integration. This vector sequence is sent to the dictionary search processing unit 170 as data as shown in FIG.

Next, the operation of the dictionary search processing section 170 will be described. FIG. 10 is a flowchart showing the operation of the dictionary search processing unit 170. As shown in FIG. 11, the standard pattern storage unit 175 is provided with a number for each event name to be identified, and is stored in numerical order. Further, each vector of the vector series constituting the event is constituted by a vector having the same inclination as the standard vector in FIG. 6, and the name of the corresponding standard vector is given. FIG. 11 shows the standard pattern storage unit 175.
Shows an example of the event name stored in the event name. In FIG. 11, the event name “stepwise rise” is stored as number 1,
The event name “fall from rise” is stored as number 2.

The processing of the dictionary search processing section 170 shown in FIG. 10 will be described in detail. First, in step 1010, a vector sequence Ai (i =
1, 2,..., M) are input from the integration processing unit 160. Next, in step 1020,
The number N of the event name to be searched is set to 1, and the search is started. In the search process 1030, if a vector sequence similar to the event in the standard pattern dictionary is selected from the vector sequence of the input pattern as a correspondence candidate, the process proceeds to step 10
By 40, the flow of the process proceeds to the transmission process 1050, where the number N of the event name and the vector sequence of the correspondence candidate are sent to the similarity calculation processing unit 180. If there is no corresponding candidate, it is determined in step 1060 whether or not the search for all event names has been completed (the number of standard patterns is Nmax). If the search for all the standard patterns has not been completed, in step 1070, the event name number is updated “N = N + 1”, and the search for the next standard pattern is continued by the search processing 1030.

FIG. 12 is a flowchart for explaining the details of the processing in the search processing 1030. Where Ai
(I = 1, 2,..., M) is a vector sequence that forms the input pattern and is sent from the integration processing unit 160, Vj (j = j
1, 2,..., N) are vector series of standard patterns of events stored in the standard pattern storage unit 175. First, “is = 1” is set in step 1210 to search for a vector of the standard pattern corresponding to the first vector of the input pattern. Next, step 12
20, "j = 1, i = is" is set, and the search is started. The vector number of the standard pattern being searched is m
If it is +1, the search ends at step 1230. The name of the vector Ai of the input pattern and the name of the vector Vj of the standard pattern are compared in step 1240.
A comparison with the next input pattern vector is made. If they match, the pair of Ai and Vj is stored in the temporary storage 1255 in step 1250. In step 1260, it is determined whether the search for all the vectors of the standard pattern has been completed. If j ≠ n, step 1270
, "I = i + 1, j = j + 1", and the search for the next vector is continued. If j = n, the process “is = ik + 1” in step 1280 is performed, and
Search processing for a standard pattern is performed from the s-th input vector. Here, ik is the number of the vector of the latest input pattern in which the name of the vector Vj of the standard pattern matches, and among the correspondence candidate strings stored in the temporary storage, the vector of the latest input pattern corresponding to Vj Number.

In the similarity calculation processing section 180, FIG.
2 according to the contents of the temporary storage 1255, the similarity S and scale K between the input pattern and the standard pattern are

[0073]

(Equation 7)

[0074]

(Equation 8)

Is calculated. Here, Vi (i =
1, 2,..., M) are vector series of standard patterns, Ai
j is a correspondence candidate of the vector series of the input pattern.
i and Aij (i = 1, 2,..., ni) are correspondence candidates. S in Equation 7 is for calculating a correlation coefficient between vector sequences, and is 1.0 if the patterns are completely similar. Equation 8 is a scale (scale) when the vector series of the input pattern most closely matches the vector series of the standard pattern. A certain range in which the similarity and the scale are determined for the standard pattern obtained by the similarity calculation processing unit 180

[0076]

The event names of the standard patterns satisfying Smax ≦ S, Kmin ≦ K ≦ Kmax (Expression 9) are arranged in ascending order of similarity and stored in the memory 190. FIG.
(C) shows an example of the encoding result.

The symbolization result (event name, similarity, scale, section of matched data) stored in the memory 190 is transmitted to another information processing system or the like via the communication device 110.

In the encoding apparatus according to the present embodiment, a search for a candidate portion of an input pattern that matches a standard pattern can be performed by using the name of a vector. There is an effect that a small number of candidates can be searched. Furthermore, since the correlation coefficient is calculated between the vector sequences and the degree of similarity is calculated, similar patterns can be searched, and the number of event patterns stored in the standard pattern storage unit 175 can be reduced. ing.

A second embodiment of the encoding apparatus according to the present invention will be described with reference to FIGS. The first embodiment is an embodiment in which pattern data stored in a database or the like is collectively converted into a symbol, whereas the second embodiment is a process in which data obtained from a process is momentarily converted into a symbol. This is an embodiment in the case of conversion. In this case, since the data up to one data before the current one in the sampling period has already been converted into the vector sequence, the processing based on the addition of one data may be performed for each sampling period.

FIG. 13 shows the configuration of the encoding apparatus of this embodiment. In the figure, reference numeral 1310 denotes a communication device similar to 110, which receives an identification number of pattern data to be processed through a communication network. Reference numeral 1320 denotes a data storage device, which stores together with process data a vector sequence, an event name, and the like, which are processing results up to the previous sampling cycle. The input processing unit 1330 is activated by a signal from the communication device 1310,
Data information corresponding to the identification number (processing results up to the previous sampling data and the latest sampling data) is read from 320 and sent to the memory 1340.

Reference numeral 1350 denotes a broken line approximation processing unit. 13
Reference numeral 55 denotes a filter storage unit, and a polygonal line approximation processing unit 135
A feature extraction filter used at 0 is stored. In this embodiment, a feature filter as shown in FIG. 15 is used as the real-time processing in order to perform processing from the latest data to the past data. The polygonal line approximation processing unit 1350 performs feature extraction from a portion including the latest data of the input pattern stored in the memory 1340, approximates the pattern data with a polygonal line, and converts the data into a vector sequence of line segments constituting the polygonal line.

Reference numeral 1360 denotes an integration processing unit, which integrates the series of broken line vectors obtained by the broken line approximation processing unit 1350 and sends it to the dictionary search processing unit 1370. 1365
Stores integration parameters used in the integration processing unit 1360.

The dictionary search processing unit 1370 compares the vector sequence of the input pattern sent from the polygonal line approximation processing unit 1350 of FIG. 13 with the vector sequence of the standard pattern stored in the standard pattern dictionary 1375, and Search for a vector series. Standard pattern storage unit 1375
As in the first embodiment, an event name, a vector sequence thereof, and a name of each vector are stored as shown in FIG. 1380 is a similarity calculation processing unit,
The similarity and scale between the vector series of the correspondence candidate obtained by the dictionary search processing unit 1370 and the vector series of the standard pattern are calculated. If the obtained event name is different from the previous sampling time and the scale is within a certain range and the similarity is a certain value or more, the standard pattern number, the corresponding input pattern number and the similarity , And the measure are stored in the memory 1390. The result stored in the memory 1390 is transmitted to another information processing system or the like via the communication device 1310.

Next, the operation of the encoding apparatus based on the above configuration will be described in detail.

FIG. 14 shows the broken line approximation processing unit 1350
6 is a flowchart showing the operation of the first embodiment. When roughly classified, a pattern division process 1410 for calculating the unevenness of the pattern by the product sum calculation of the input pattern f (t) and the feature extraction filter W2 (x) and obtaining the point τp at which the pattern is divided into polygonal lines 1410
And feature extraction filters W0 (x), W1 (x), and W2
By using (x), the average value of the data near the division point of the pattern is obtained, and the estimated value fe (τ
p), an end-point estimation process 1450 for obtaining an end-point estimated value from a regression line at the end of the data section, and a vector series obtained when the pattern is approximated by a polygonal line using the division points and the end-point estimated values. And a vector series calculation process 1460 for calculation.

First, in the pattern division processing 1410, the input pattern data f (t) (t = −T, −T +
1,..., 0) for the product-sum calculation using the feature extraction filter W2 (x) stored in the filter storage unit 1355

[0087]

(Equation 10)

As a result, a feature amount g2 (0) indicating the degree of unevenness near the current time is calculated. Here, a is a constant indicating the spread of the filter. As a result of performing a product-sum calculation on the example of the input pattern shown in FIG. 16A using the feature extraction filter W2 (x) shown in FIG.

[0089]

[Equation 11]

When the above condition is satisfied, t = -b (b is a constant value determined by the filter) is a new division point. If a new dividing point is not generated by the processing step 1420, the process proceeds to a processing step 1450, and if a new dividing point is generated, the process proceeds to a processing step 1430.

In the estimating process 1430, the feature extraction filter Wk (x) (k = 0, 1, 2) stored in the filter storage unit 1355 is used at the dividing point τp obtained by the pattern dividing process 1410. Sum of products calculation

[0092]

(Equation 12)

As a result, the characteristic amount g0 (τ
p), a feature amount g1 (τp) representing the inclination, and a feature amount g2 (τp) representing the unevenness. FIG. 15 shows an example of the feature extraction filters W0 (x), W1 (x), and W2 (x). The estimated value fe (τp) of the pattern at the division point τp is

[0094]

(Equation 13)

Is calculated.

Next, in the termination estimation processing 1440,
Find the estimated value at the end of the input pattern. Estimation process 143
0 (τp, fe (τ
p)) through a regression line

[0097]

[Equation 14]

Is calculated.

By substituting t = 0 into Equation 14, the estimated value fe (0) (= h) at the end (t = 0) of the input pattern is obtained.
(0)) is required.

In the vector series calculation processing 1440, the sequence of points (τp, fe (τp)), (0, fe (0)) obtained in the estimation processing 1430 and the termination estimation processing 1430
From the vector sequence Aj = (P
j, Qj),

[0101]

(Equation 15)

Is determined by FIG. 16B shows a vector sequence obtained from an example of a point sequence when a new division point τp is generated. Each of these vectors is given the name of the closest (similar to the gradient) vector among the standard vectors shown in FIG. 6 and sent to the integration processing unit 1360.

Next, details of the processing in the integration processing unit 1360 will be described. The integration process is performed by the polygonal line approximation processing unit 135.
As a result of 0, when the number of vectors is three or more, the processing is performed on the vector two before the latest vector. The content of the processing is the same as the processing shown in FIG. FIG. 16C shows an example of the processing result.

The operation of dictionary search processing section 1370 is the same as dictionary search processing section 170 in FIG.
The operation of 380 is the same as the similarity 180 of FIG. FIG. 17 shows an example of the obtained symbols.

The symbolization result (event name, similarity, scale, section of matched data) stored in memory 1390 is
The information is transmitted to another information processing system or the like via the communication device 1310.

In the encoding apparatus according to the present embodiment, not only the same effects as in the first embodiment, but also the processing is simple,
Since a symbolic representation of data can be obtained at high speed, a large amount of process data can be processed by an inexpensive processing device.

To better understand the implementation of the symbolizer and to evaluate its inherent potential and advantages, the following provides a brief mathematical background and basis for the techniques used to implement the implementation. Present. The expressions introduced also provide insight into the functionality of the embodiment.

The feature extraction filters W0 (x), W1 (x), and W2 (x) stored in the filter storage unit 155 or 1355 and used in the polygonal line approximation processing unit 150 or 1350 are as follows. A filter for obtaining an average value, W1 (x) is a filter for obtaining an average inclination, and W2 (x) is a filter for obtaining an average unevenness of a pattern. Below, W0 (x), W1 (x), and W2
The method of obtaining (x) is organized and shown.

As a method for developing pattern data,

[0110]

(Equation 16)

A polynomial Hm (x) defined by Where E (x) is a suitable differentiable function, a and b
Is a constant indicating a defined section. At this time,

[0112]

[Equation 17]

In other words,

[0114]

(Equation 18)

Constitutes an orthonormal function system.

When the time series data f (t) is expanded near time t using the orthonormal function system ψm (x) as it is,
In order to obtain a good approximation of the original data, it is necessary to obtain higher-order terms. In order to avoid this, the time series data is

[0117]

[Equation 19]

Will be handled after conversion.

By expanding the equation (19) using the orthogonal function system of the equation (18),

[0120]

(Equation 20)

Is obtained. Here, the expansion coefficient am (t) (m =
0,1, ..., ∞)

[0122]

(Equation 21)

Is determined by From Equations 19 and 20, the time-series data f (t) is near the time t,

[0124]

(Equation 22)

This is expanded.

The differentiable function E (x) is

[0127]

(Equation 23)

Assuming that b = −a, an exponential function represented by the following formula, Hm (x) is an m-order Hermite polynomial from equation (16). In addition, according to Equation 22, the time series data f (t) is expanded into a polynomial in the vicinity of the time t. At this time, the expansion filter Wm (x) is obtained by adding Expression 21 and Expression 18 to Expression 22.
By substituting

[0129]

(Equation 24)

From the above,

[0131]

(Equation 25)

Are represented as follows.

The time series data estimation value fe (t) at time t is obtained by using the result of the polynomial expansion by the feature extraction filter expressed by Expression 25. By substituting x = 0 into Equation 22, fe (t) becomes

[0134]

(Equation 26)

Is obtained.

A quadratic term a2 (t) of the expansion coefficient according to Equation 26
Indicates a state of unevenness of the time-series data near time t. Therefore, for the time series data f (t),
The time at which the gradient of the time-series data changes can be obtained by obtaining the expansion coefficient a2 (t) of the quadratic polynomial according to 4 and extracting the time having the extreme value.

Hereinafter, a method of obtaining the similarity and the scale will be described. A vector sequence of one symbol stored in the standard pattern storage unit 175 or 1375 is represented by Vi (i =
1, 2,..., M), and the vector series of the corresponding candidates included in the time-series data is Ai (i = 1, 2,..., M). When the vector sequence stored in the standard pattern storage unit 175 or 1375 is multiplied by K, the difference between the two vector sequences is 2
The squared error E is

[0138]

[Equation 27]

Are represented as follows. First, a scale K that minimizes the square error E is obtained. Expanding Equation 27 gives

[0140]

[Equation 28]

Is obtained. Here, (,) indicates the inner product of the vectors. Since Equation 28 is a quadratic equation with a quadratic coefficient relating to K being positive, the scale K that gives the minimum value of the square error E is

[0142]

(Equation 29)

From the above,

[0144]

[Equation 30]

Is given. The square error E at that time is obtained by substituting Equation 30 into Equation 28,

[0146]

(Equation 31)

Is given by The square error E in Equation 31
Is normalized with respect to the vector series of the time series pattern,

[0148]

(Equation 32)

Is as follows. Therefore, an evaluation index of the degree of correspondence based on the square error E normalized by the scale K is defined by the similarity S using the square root of the second term of Expression 32.

[0150]

[Equation 33]

In Equation 32, the square error E is the same whether the scale K in Equation 30 is positive or negative. But,
When a symbolic name is given to the time-series data, the similarity must be reduced if they are inverted and match. Number 33
Is given by considering the square root of the second term on the right side of Equation 32.
An appropriate similarity can be obtained. The similarity S represented by Expression 32 satisfies −1.0 ≦ S ≦ 1.0, and has a form similar to that of a normal correlation coefficient, and is considered as a correlation coefficient between vector sequences. be able to.

As described above, the equations of the similarity S and the scale K were obtained on the assumption that the vector series of the time series data and the vector series of the standard pattern corresponded one to one. In general, one vector may correspond to a plurality of vectors between the vector sequence of the standard pattern and the vector sequence of the time-series data. In this case, the similarity S and the scale K shown in Equations 33 and 30 are calculated.

[0153]

(Equation 34)

[0154]

(Equation 35)

It is used after being modified. Here, the vector series of pattern data corresponding to the vector Vj of the standard pattern is Aij (j = 1, 2,..., Ni).

The symbolizing method according to the present embodiment is characterized in that by converting time-series data into a vector series and handling it, it is possible to perform collation in consideration of pattern displacement and changes in shape.

Hereinafter, embodiments of the process operation support system of the present invention will be described in detail with reference to the drawings.

FIG. 18 shows the configuration of an embodiment of the process operation system of the present invention. This embodiment is an example applied to the process operation of a plant.

The process operation system shown in FIG. 18 includes a process data capture processing unit 1 which captures process data from various sensors Sn and the like of a monitoring target plant as numerical data in a time series, and symbolizes the characteristics of the captured data. A symbolization processing unit 2 for generating symbolized data to be represented, a symbolization processing dictionary 3 used for the symbolization processing, and the symbolized data is output by corresponding process data (numerical data) and symbolization processing. Corresponds to the encoded historical file 5 that stores the similarity in time series together as membership information and the encoded data storage processing unit 4 that performs storage processing therefor, and the encoded data stored in the encoded data historical file 5 Fuz to be displayed together with the membership data
a zy output processing unit 31, an inference engine 32 for executing inference by knowledge processing, a symbolization processing unit 2 corresponding to process data and symbolized data used for starting and inferring inference in knowledge processing, and a symbolized data historical file 5 has a time-series data support processing unit 15 to be passed to the inference engine 32, a knowledge base 33 for storing the knowledge, and a knowledge acquisition input processing unit 34 for inputting data to the knowledge base 33. .

Further, an abnormality treatment guide output unit 35 for outputting guidance indicating a treatment method for the abnormality obtained in the inference execution, and a plant status inquiry processing unit 36 for directly inquiring the operator in the inference execution.
And a little more signal reception processing unit 3 for converting ambiguous encoded data into numerical data and outputting it to control data output processing unit 12 when outputting control data in inference execution.
7 and a symbol inversion processing unit 11, a symbol integration information input unit 38 for integrating a plurality of signal data to create new symbol data, and a result data captured by the process data capture processor 1. A result data file 41 to be stored, a symbol integration information conversion unit 39 that generates symbol integration information using input information and the result data, a symbol integration information file 40 that stores symbol integration information,
A man-machine device 10 for performing the input operation is provided.

In this embodiment, the operation amount for controlling the process is controlled by the control arithmetic processing unit 1 based on the process data.
7 and a control data output processing unit that sends control data to the operation target based on the calculation result and sends control data based on the output of the inference engine 32 or the output of the signal reception unit 37 to the corresponding operation target. 12 is provided. In a plant operation system having a controller, the control operation processing unit 17 and the control data output processing unit 12 may be separately provided in the controller. Further, the process operation system of the present embodiment may be incorporated in a control system such as a plant controller.

The symbolization processing unit 2 and the symbolization dictionary 3
This constitutes a symbol processing means, which is configured in the same manner as the system shown in FIG. Here, it is built not in an independent system but in an operation system.

The man-machine device 10 includes a CRT display 10a, a touch panel 10b, and a keyboard 10c.
Having. The man-machine device 10 constitutes input means.

FIG. 19 shows an outline of a system configuration of hardware constituting the process operation system shown in FIG. As shown in FIG. 19, in the present embodiment, a man-machine device 10, a CPU (central processing unit) 29 that executes various processes and realizes various means of the present invention, a symbolization dictionary 3,
Signaled data historical file 5, knowledge base 3
3, a file device 21 for storing data such as a symbolized integrated information file 40 and a performance data file 41, and a CP
A process input / output device 22 for inputting / outputting data between the U29 and the plant is provided as hardware. The file device 21 includes a main storage device and an auxiliary storage device.

Next, the operation of this embodiment will be described for each of the following functions.

(1) Symbolization Processing The data from the various sensors Sn of the plant is captured by the process data capture processor 1. The process data fetching unit 1 collects data at a fixed sampling period and stores data for a fixed period. Here, the process data becomes time-series data and is input to the symbol processing unit 2. The symbolization processing unit 2
The vector information of the input data is generated by the polygonal line approximation processing of the polygonal line approximation processing unit 1350 in the encoding apparatus shown in FIG. Compare and output the coded data with the highest similarity.

By the way, examples of symbols include rise, fall, step rise, rapid rise, gentle rise, pulse, etc., but any name (symbol) may be given. In other words, it is only necessary that a human can understand the meaning of the symbol here by listening / seeing the symbol.

[0168] The symbolizing dictionary 3 can directly store character strings, which are symbolized data, in the form of a table. Alternatively, a configuration may be adopted in which a character string is simply stored as a code and a table for associating the code with the encoded data is separately provided.

When converting time-series data, which is input data for symbolization processing, into vector information, two parameters, a time scale and a data size, are required for each process data as the above-mentioned conversion coefficients. That is, since the reaction and response speed vary depending on the type of process data (for example, temperature data, flow rate data, and pressure data), a slow response speed (a temperature that does not change easily) such as temperature data requires a large time scale. However, the vector information is made sparse with respect to the time scale, and the flow rate and pressure data have a fast response speed. Is used to ensure that the correct symbol is attached to

FIG. 20 shows an application example of the symbol conversion parameters. In the case of FIG. 9A, since the time scale is small, even a slight change (movement) appears as vector information. On the other hand, in the case of FIG. 3B, the time scale is large, so that even if the data fluctuates considerably, the change in the vector information is small. As described above, by determining this parameter depending on the nature of the data to be monitored, an appropriate symbol can be output.

In the symbolization process, the similarity is calculated internally before the conversion output result (symbolized data) is output, and the similarity is registered in the vector information of the input data and the symbolization dictionary. It shows the degree of similarity between the two in comparison with a standard pattern. In the present invention, by using this similarity in the following functions, a plant operation system having a wider range of functions than in the past can be obtained.

The coded data displayed on the screen is output together with the degree of matching. The inference engine 32 has a function of performing inference depending on the degree of matching. When a plurality of encoded data names are given to one input time-series data, the data having the highest matching degree is output as information.

(2) Output of encoded Fuzzy data The encoded data and similarity obtained by the encoding process of (1) are stored in the encoded data historical file 5 by the storage process of the encoded data storage processing means 4. Is stored. FIG. 21 shows the structure of the encoded data human file. Symbolized data historical file 5
Has a file configuration for each process data for each time. The components of the file are time,
There are four elements: process data value, symbolized data, and fitness. The matching degree stores the above-described similarity degree.

The symbolized fuzzy data output processing unit 31
If the encoded data for the various input data is different from the encoded data obtained at the previous collection timing, the operator is notified by the man-machine device 10 that there is an abnormality in the plant state. Therefore, as shown in FIG. 23, the symbolized fuzzy data output processing section 31 has a symbolized fuzzy data output display section 31a and an alarm registration information file 31b.

Here, an example of data output in the symbolized fuzzy data output processing unit 31 will be described with reference to FIG.

(A) Symbolized fuzzy data output processing unit 3
1 is, as shown in FIG. 22A, an occurrence time, a data name (F1024), encoded data (hunting),
The degree of conformity (0.9) is displayed as characters on the display 10a.
To be displayed. The procedure of the method of displaying a flash alarm of encoded fuzzy data shown in FIG. 23 is periodically executed.

As shown in FIG. 24, first, it is determined whether or not encoded data is newly generated, and if it is present, the data name, the encoded data, and the fitness are extracted (step 2).
401, 402). Next, for the extracted data, it is checked whether or not the data name is registered in the alarm registration information, and if it exists, it is determined whether or not the conformity exceeds a specified value.
0 (steps 2403 to 2406).

(B) Symbolized fuzzy data output processing unit 3
1 is a display 10a as shown in FIG.
, A pictorial display (symbol) representing symbolized data
, An occurrence time, a data name (F1011, F2011,...) Arranged below the symbol, and a degree of matching (0.8, 0.5,...). The symbols are registered in the dictionary in advance, and when a change occurs in the encoded data, the corresponding symbol, data name, and matching degree are displayed.

(C) Symbolized fuzzy data output processing unit 3
When the data change occurs, 1 displays an alarm symbol A1 indicating that the data change has occurred, as shown in FIG. 22C, and causes the alarm symbol A1 to blink. When the operator points the blinking symbol A1 on the touch panel 10b of the man-machine device 10, a window W showing the symbolized data status in detail is displayed. In the window W, a symbol meaning encoded data and generated data names arranged in chronological order for each encoded data are displayed.

A data change can be detected by a change in the expression of the encoded data. That is, the symbolized fuzzy data output processing unit 31 detects the presence / absence of a data change by comparing the encoded data output momentarily with the corresponding symbolized data of the previous time. In this case, for example, when something that was constant changed to rising,
It is sufficient to compare “constant” and “increase” of the encoded data, and it is not necessary to compare the numerical data. Therefore, it is possible to clearly and easily determine whether there is a difference.

(D) Symbolized fuzzy data output processing unit 3
1, the display items shown in (a) above can be set so that the latest encoded data is displayed in chronological order.

By such output processing, the operator can display the detailed data as necessary and know the plant state in the man-machine device 10 without always monitoring the detailed data of the input data. . Further, since the degree of conformity is displayed together with the encoded data, the degree of reliability for the encoded data can be known, and the information service can be improved. In other words, the symbolized data is expressed in such a way that the characteristics of the state of the data can be intuitively understood by a person, so that the state of the target can be understood in a very short time. On the other hand, since the state of the target is expressed in an abstract manner, it is not enough to know how reliable the expression is. Therefore,
If the matching degree indicating the reliability is also displayed, this point is complemented, and the reliability of the symbolized expression representing the target state is improved.

(3) Construction of Encoded Database The encoded data and the similarity output by the encoding process of (1) are defined in the encoded data historical file 5.
Is stored in This symbolized data historical file 5 adds input data (numerical data) to the symbolized data and similarity, and collects the input numerical data, symbolized data and similarity at that time as a time-series data file. Accumulate. This storage processing is executed by the symbolized data storage processing unit 4 as described below.

The processing of the historical collection procedure shown in FIG. 25 is periodically executed. First, a loop is performed for the number of input data, and for each input data type, the encoded data, the similarity, and the numerical data at that time (when this processing is executed) are extracted and stored in the encoded data historical file.

As described above, by adding the similarity to the conventional components to the encoded data historical file, the above-mentioned (2) and various functions described below can be realized.

FIG. 21 shows an example of the configuration of the encoded data historical file. In the example shown in the figure, a pair of numerical data, encoded data, and similarity are stored in chronological order at one-minute intervals.

(4) Knowledge Acquisition Input Processing In the present invention, data stored in a symbolized data historical file is used for inference of knowledge processing, and a rule description having more flexibility (ambiguity) than a conventional rule description. , Or a function that enables the execution of inference will be described below.

(4-1) Fuzzy Symbol Description This function is to directly describe the data of the above-mentioned symbolized data historical file 5 in the rules used for inference, so that numerical data, symbolized data, The similarity is collected, and the collected data is stored in the knowledge base 33.

The similarity in the symbolization process is considered as shown in FIG. In (1) of FIG. 26, Fuzzy
It shows the concept of membership in. When the change rate (dPV / dt) of the process data is applied to “slow rise”, “rise”, and “rapid rise” expressed by the membership function, the fitness of “rise” is 0.45 in the figure. , And the degree of fitness of the “rapid rise” is 0.3. In (2) of FIG. 26, it can be said that the same effect as Fuzzy can be obtained by catching the similarity between the symbolized data “rise” and “rapid rise” in the same meaning as the above-mentioned fitness. Therefore, in the present invention, the similarity of the encoding process is used as it is as the fitness of Fuzzy.

First, an example of a rule description by this function is shown in FIG. 27 and below. 1F (＠ trend of F1024 is sharp rise, ＠ fitness is greater than 0.8) (＠ trend of P2050 is hunting, ＠ fitness is greater than 0.7) THEN pump A abnormality This example shows process data ( The encoded data of F1024) rises sharply, the fitness at that time is greater than 0.8, and the encoded data of the process data (P2050) is hunting, and the fitness at that time is greater than 0.7. In this case, it means that the pump A is in an abnormal state.

As described above, the encoded data and the degree of conformity can be described in the rules used in the inference.

By performing the inference using such rules, it is possible to confirm the degree of conformity data to the symbolized data in the inference, and it is possible to increase the verification accuracy by the inference.

(4-2) Symbolized Directed Graph The symbolized directed graph is an FTA (FaultTrerancanaris) using the symbolized data and the fitness.
ys), wherein occurring events (indicated by symbolized data) are indicated by boxes, and the boxes are connected in a tree shape in the order of occurrence of the events according to the causal relationship, and the connection of each event is performed. In order to indicate the direction and the direction of change, these are connected by an arrow as a connection symbol and displayed. FIG. 28 shows an example of the description.

In the graph shown in FIG. 28, events (events represented by a directed graph) propagate from left to right in the graph. In the example shown in the figure, the process data (F100
1) starts the development of this event. Starting from the trigger generation of F1001, the state of the process data (F1002, PV) 0 minutes after the trigger generation is constant, that is, the symbolized data of the instantaneous value of F1002 0 minutes after (that is, in real time) is “constant”. , The degree of conformity of which is larger than 0.8 and the process data after 0 minute (F2931,
PV) is "sloping down" and its fitness is greater than 0.7 and the process data (P2031, P203
If the symbolized data of V) is “sloping down” and the degree of conformity is greater than 0.8, and if the symbolized data of the process data (T2031, PV) is “downward” and the degree of conformity is greater than 0.4,
A conclusion is reached that “the device 2031 is abnormal”.

Hereinafter, the same consideration will be given. This drawing is drawn and input from the man-machine device 10, necessary information (process data name, symbolized data name, conformity condition) is input, and each block is connected by an arc.

In the present embodiment, the format of the above-mentioned directed graph is defined on a machine different from the processing device on which the encoding device is mounted. Note that the configuration may be defined on the same machine.

Each data input for the format of the created directed graph is performed by FIF input into a box (blank space filling method). This can be performed, for example, by using the man-machine device 10 and the device shown in FIG.
In other words, the information input person (1) box generation,
With regard to (2) input of box information and (3) connection between boxes by an arc, a graph can be completed by drawing while repeating input of each data.

Now, the rules shown in FIG. 29 can be generated from the data input for drawing. In the figure, four rules are registered in the knowledge base 33.

In FIG. 28, a horizontal connection between boxes corresponds to an AND condition. A vertical connection between boxes indicates an OR connection. In the figure,
From the “F1001 trigger generation” box, the “F10” which is an IF box (white box) is connected by OR connection.
02PV constant> 0.8 "box (No. 001),
“L1031PV hunting> 0.9” box (N
o. 011), “T2045SV constant> 0.7” box (No. 021), “P1001PV sudden rise> 0.
8 "box (No. 031) by OR connection. The OR connection means that the IF... THEN... Rules are independently generated. Therefore, FIG.
That is why the rule expansion example in the figure has four rules.

[0200] An example of rule expansion at the left end is "IF1002PV constant>0.8" which is an IF box (box with white frame)
Box (No. 001), “F2031PV sudden drop> 0.7” box (No. 002), and “P2031
PV sudden drop> 0.8 "box (No. 003)
“T2031PV drop> 0.4” box (No. 00)
And 4) are all connected by an AND connection. Therefore, the contents obtained by connecting these box contents under the AND condition are expressed as IF items in the rule.

The THEN item is a “box 2031 abnormal” box (N) which is a THEN box (shaded box).
o. 005) is expressed.

Next, information registered in the box will be described. The information in the IF box is as follows: (1) Process data name (Tag name) Example: F10
02, (2) Process data item (data type) Example:
PV (PV means measured value, SV means set value, MV means manipulated variable), (3) state (= encoded data name) Example: constant, (4) condition (similarity, conformity (Regulated condition) Example:>, (5) Similarity default value Example: Consists of five data of 0.8. The information in the THEN box is only (1) conclusion.

Next, the correspondence between the information in the IF box and the information in the rule at the time of rule expansion will be described.

Here, "(" to ")" of the IF condition item of the rule corresponds to one box. That is, ([Process data item] of [Process data name] is [Status], and [Confidence] is [Default] [Condition]).

In the THEN term, the conclusion of the THEN box is [conclusion].

Is represented by

Next, a method for registering a symbolic directed graph will be described with reference to FIG.

FIG. 30 shows a screen display example used for registering a symbolic directed graph. This screen is displayed, for example, on the display 10a of the man-machine device 10.

As functions F1 to F5 of the registration screen, (1) IF BOX (F1), (2) THEN box (F2), (3) AND connection (F3), (4) O
There are five types: R connection (F4) and (5) layout (F5).

(1) The “IF BOX” function first determines the box generation position by pointing to the display 10a. This operation can be performed by touching a corresponding portion of the touch panel 10b with a finger. When the position is determined, the window W for defining detailed information
Is opened and detailed information (process data name, item,
A screen for setting the symbolized data name, condition, and default value is displayed in the window W. As shown in FIG. 30, the setting is made in the blank space corresponding to the item in the window W,
This is done by inputting symbols and characters. When the setting is completed, registration is instructed by touching the registration function in the window W with a finger, and the information is stored. Once registered, the data will be displayed in the corresponding box.

(2) The “THEN box” function is first performed by a pointing operation on the display 10a, like the “IF BOX”. When the position is determined, a window (not shown) for defining detailed information is opened, and detailed information (conclusion) is set. When the setting is completed, the information is stored by the registration function in the window. Once registered, the data will be displayed in the corresponding box.

(3) The “AND connection” function
Box information generated by the “IF BOX” and the “THEN box” is connected by an arc of a connection symbol. The connection method specifies from a box to a box by a pointing operation in accordance with a from ... to ... format. By this operation, the connection relation of the boxes is determined, and the direction of the change of the event is determined.

(4) "OR connection" is also specified in the form "from ... to ..." similarly to "AND connection".

(5) “Layout” is a function for showing the entire configuration of the registered symbolized directed graph. Thus, the layout is laid out so that the entirety can be seen in one screen. This operation can be performed automatically, but may be instructed by an operator.

[0215] The symbolized directed graph thus constructed is stored in the knowledge base 33 from the knowledge acquisition input processing unit 34. Then, when an event actually occurs in the process, the symbolized data and the fitness are applied to the symbolized directed graph stored in the knowledge base,
The cause of the abnormality and the abnormal device can be inferred.

(5) Abnormal output guide This function outputs the result obtained by inference as an abnormal output guide. The abnormal output includes the result of the inference using the Fuzzy symbol description of each event and the inference of the symbolized directed graph. At the time of abnormal output, it is preferable to show not only the inference result but also the process leading to the conclusion. In the former case, for example, by showing the inference rule in a format as shown in FIG. 27 and also showing the degree of conformity of the actual data, the degree of conformity to the rule can be determined, and the reliability of the inference is improved. .
Similarly, of the inference of the symbolic directed graph and the inference of the symbolic directed graph as shown in FIG. This is the inference engine 32
What is necessary is just to output the rule that fits among the rules of the knowledge base 33 used in. In addition, it is preferable that the degree of conformity of the symbolized data of the event to be inferred is also indicated. The display format can be, for example, as shown in FIG.

When the degree of conformity is low and the source of the abnormality cannot be narrowed down, a plurality of possible inferences may be described.

(6) Plant Status Inquiry Function This function is for inquiring the operator about process data other than the sensor information obtained from the process, receiving input of the corresponding non-sensor information from the operator, and supplementing it. It is. As shown in FIG. 31, information that is difficult to obtain with the sensor of the process, for example, non-sensor information such as smell, color, abnormal noise, and on-site patrol result, has conventionally been difficult to reflect on the control elements of the process. However, it is well known that even this kind of information is an important signal that informs the state of the process.

Therefore, in this embodiment, the operator inquires directly about this kind of non-sensor information from the inference, and reflects the inquiry result in the inference. That is, as shown in FIG. 31, the information from the sensor is subjected to the above-described encoding processing and input to the knowledge base 33, and the required non-sensor information is inquired from the system side to the operator to prompt the input. To collect. Then, the input result is used again for inference to determine an abnormality or the like.

The inquiry items can be described as rules in the knowledge base 33. That is, FIG.
As shown in (A), the corresponding non-sensor information item is described in the THEN section for the IF section described by the encoded data based on the sensor information.

An operator inquiry is made using the man-machine device 10. In this case, the inquiry is P
It is composed of RINT function and INPUT function.
The T function is an output to the man-machine device 10, and FIG.
As shown in (5), an inquiry corresponding to the THEN unit is displayed on the display 10a. The INPUT function is an input function from the man-machine device 10. FIG.
As shown in FIG. 3, in response to a question displayed on the screen of the display 10a, the keyboard 10c and the touch panel 10 are used.
By using b, the presence or absence of a certain state can be input.

The input result is passed to the inference engine 32 and used for inference. Therefore, the knowledge base 33
Describes an inference rule using various non-sensor information as an IF unit. As shown in FIG. 32, as an example,
An example of inference based on the input to the inquiry shown in FIG.

[0223] Also, inquiries can be made directly in the rules.
It can also be performed by a method of describing an imaginary sensor name (TAG name).

In this method, a sensor name is described in an IF term in an inference rule, and the corresponding sensor data (process data) is taken in by process data, and a method of symbolizing and collecting the data is used. That is, the imaginary sensor name (T
AG name) is described and registered. At the time of data collection, the system determines whether the data is real data or fictitious data. If the data is fictitious data, the system is set to inquire an operator about the contents. Specifically, this is performed as in the following example.

(Example) IF (ΔPV of F1002 is larger than 12.0) (ΔMV of T2358 is smaller than 75.0) (ΔPV of V2347 is smaller than 45.8) Abnormality of THEN F1111.

Here, V2347 is an imaginary sensor name. In this case, the system first determines F1002, T2
Attempts to collect the data of each of 358 and V2347 from the process data capture processing unit. When actually collected, actual data can be collected for F1002 and T2358. However, for V2347, since no sensor exists, an error (ERROR) occurs and actual data cannot be captured. For this reason,
The system knows that this data is a fictitious sensor name. Therefore, the system of this embodiment determines that the inquiry is an inquiry to the operator.

[0227] When it is determined in this way, the system outputs the following message to the man-machine device 10, displays it on the screen of the display 10a, and waits for the information provided by the operator.

**** Inference XXXX Please input the next data V2347PV: The operator who saw this inputs the data of V2347. The operator operates the display 10a of the man-machine device 10.
May not be constantly monitored, so that a specific alarm may be output to notify that there is an inquiry. In this case, an alarm that informs of an abnormality may be a special alarm.

(7) Symbolized data integration function The symbolization integration function integrates a plurality of combinations of encoded data conventionally converted by a symbolizing device and combines them into new encoded data. This is a naming function. As a result, data of a longer time interval can be recognized by the symbolized data. In other words, this means that the interpretation of the data differs depending on whether the data in a certain time section is analyzed microscopically or macroscopically. For example, a series of micro-coded data (ascending-descending-ascending-descending-...) Becomes (hunting) when expressed in a macroscopic manner. As described above, according to the symbol integration function of the present embodiment, a plurality of views can be obtained as described above without changing the time scale in the symbol encoding device.

The function of integrating encoded data will be described with reference to FIG. In FIG. 34, there are three patterns of performance data. Each encoded data in the example shown in FIG.
(Constant-surge-increase), (constant-increase), (constant-increase-constant). These correspond to cases where a human reads the situation with reference to the performance data and says that any pattern is “rapid rise”. That is,
(Constant-Rapid rise-Rise) = (Rapid rise), (Constant-Rise)
= (Rapid rise), (constant-rise-constant) = (rapid rise). As described above, the integration of a plurality of encoded data into one encoded data is performed by the encoded data integration function.

The symbol integration is performed by the symbol integration information input unit 38, the symbol integration information conversion unit 39, and the symbol integration information file 40. It goes without saying that the man-machine device 10 is used to input data and instructions and output results.

The processing method of symbolized data integration will be described with reference to FIG. The actual data graph in the figure is symbolized by the symbolizing device as (constant-rapid rise (fitness 0.8) -constant). The details of the encoded data at this time are shown together with the registered contents of the encoded data historical file 5.

On the other hand, a combination of the teacher data and the unified symbolized data sequence is registered in the symbolized integrated information input unit 38 in advance as symbolized integrated information. The unified symbolized data string indicates a combination pattern of the symbolized data to be integrated. There is not always one pattern. In this embodiment, a plurality of different patterns correspond to the same teacher data. The teacher data integrates each pattern of the unified symbolized data string to give one symbol. The number of teachers of the symbolized integration information indicates the number of teachers of teacher data (symbolized data judged by humans) input from the symbolization integration input unit with respect to the teacher pattern extracted from the performance data. The numerical value written in the unified symbolized data string indicates the degree of conformity, and limits the certainty of the symbolized data.

[0234] The symbolized integrated information conversion unit 39 includes a symbolized data historical file 5 and a symbolized integrated information input unit 38.
And reference. First, the changed part of the symbolized data name and the matching degree at that time are extracted from the historical file 5. In the drawing, in the table of historical encoded data, the extraction result is indicated by hatching (constant (fitness 0.7) -rapid rise (fitness 0.9) -constant (fitness 0.6)).

Next, it is searched whether or not the extracted pattern exists in the symbolized integrated information table. At this time, if there is a fitness condition, the fitness is also compared with the fitness of the extracted data. If there is a corresponding pattern in the symbol integration information, teacher data for integrating the integrated symbol data string is obtained. In the example of FIG. 35, as shown by oblique lines in the table of the symbolized integrated information, a suitable one is obtained.

The symbolized integrated information obtained in this manner is stored in the symbolized integrated information file 40. The symbolized integrated information stored in this file is stored in the knowledge base 33.
In the same way as described above, it is used for describing inference rules used in the inference engine 32.

The data integration function in this embodiment is different from, for example, the integration processing using vector data in the integration processing section 160 shown in FIG. 1, and performs integration using symbolized data (character data). By integrating the symbolized data (character data), data integration in a range (time range) larger than the vector integration range (time range to be integrated) can be performed.

For example, if the time axis range of the pattern to be detected is expanded, for example, to 8 hours, 1 day, 1 month, or 1 year, it is necessary to hold long-time vector information. This requires a large storage capacity and increases the amount of data to be processed when inputting / outputting and comparing data. On the other hand, by integrating long-term data, this difficulty can be reduced.

In the embodiment described above, the embodiment in which the present invention is applied to the process operation support of the plant has been described. However, the present invention is not limited to this. Various events can be symbolized and represented, and the system can be widely applied to a system in which events are inferred using the symbolized data. For example, an energy supply process such as a power system,
The present invention can be applied to a distribution process, a securities process, a financial process, a demand forecasting process for various events, and the like.

[0240]

As described above, according to the present invention,
The characteristics of the information indicating the state of the process can be extracted without being influenced by the personal condition of the operator who monitors the process, and the state of the process can be accurately grasped.

In addition, it is possible to create an inference rule for knowledge processing that is similar to the content that is conventionally considered by humans.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a symbolizing device used to convert process data into a symbolic expression in the present invention.

FIG. 2 is a flowchart showing an operation of a broken line approximation processing unit in the encoding apparatus.

FIG. 3 is an explanatory diagram (part 1) illustrating a process of symbolization in the symbolizing device.

FIG. 4 is an explanatory view (part 2) for explaining the process of symbolization in the symbolizing device;

FIG. 5 is an explanatory diagram showing an example of a feature extraction filter used in the encoding device.

FIG. 6 is an explanatory diagram showing an example of a standard vector used in the encoding device.

FIG. 7 is a flowchart illustrating processing in an integration processing unit in the encoding device.

FIG. 8 is an explanatory diagram illustrating an example of an integration process in an integration processing unit.

FIG. 9 is an explanatory diagram illustrating an example of an integration process in an integration processing unit.

FIG. 10 is a flowchart illustrating processing in a dictionary search processing unit of the symbolizing device.

FIG. 11 is an explanatory diagram showing an example of a standard pattern dictionary used in the dictionary search processing unit.

FIG. 12 is a flowchart illustrating a procedure of a search process in a dictionary search processing unit.

FIG. 13 is a block diagram showing the configuration of another embodiment of a symbolizing device used to convert process data into a symbolic expression in the present invention.

FIG. 14 is a flowchart showing the operation of a broken line approximation processing unit in the symbolizing device.

FIG. 15 is an explanatory diagram showing an example of a feature extraction filter used in the polygonal line approximation processing unit.

FIG. 16 is an explanatory diagram for explaining a process of symbolization in a broken line approximation processing unit;

FIG. 17 is an explanatory diagram for explaining the result of symbolization.

FIG. 18 is a block diagram illustrating an outline of a configuration of an embodiment of a process operation support system according to the present invention.

FIG. 19 is a hardware configuration diagram of a process operation support system of the present invention.

FIG. 20 is an explanatory diagram showing symbol conversion parameters used in the symbol processing.

FIG. 21 is an explanatory view schematically showing an example of a file configuration of a symbolized data historical file.

FIG. 22 is an explanatory view showing an example of a display screen for displaying a process state by treating symbolized data as fuzzy data.

FIG. 23 is an explanatory diagram showing an example of alarm display in alarm output processing by treating encoded data as fuzzy data.

FIG. 24 is a flowchart showing a processing procedure in alarm output processing by treating symbolized data as fuzzy data.

FIG. 25 is a flowchart showing a procedure for collecting a symbolized historical file.

FIG. 26 is an explanatory diagram showing an example in which measurement data is converted into symbolized data by the symbolizing process of the embodiment.

FIG. 27 is an explanatory diagram showing an example in which symbolized data of the present embodiment is treated as fuzzy data and inference rules are described.

FIG. 28 is a block diagram illustrating an example of a symbolized directed graph using the symbolized data according to the embodiment.

FIG. 29 is based on the symbolic directed graph shown in FIG. 28;
Explanatory drawing which shows the example which expands a knowledge rule.

FIG. 30 is an explanatory diagram showing an example of registration of the symbolic directed graph shown in FIG. 28;

FIG. 31 is a conceptual diagram showing an example of a function when an inquiry about a plant status is made in the embodiment.

FIG. 32 is an explanatory diagram showing an example of a rule for performing the plant status inquiry and a rule description based on the inquiry result.

FIG. 33 is an explanatory diagram showing an example of a plant status inquiry screen for making the above-mentioned plant status inquiry.

FIG. 34 is an explanatory diagram showing an outline of an integration process for converting encoded data into encoded data in the present invention;

FIG. 35 is an explanatory diagram showing an outline of an example of a search process of integrated coded data in the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Process data taking-in processing part, 2 ... Symbolization processing part, 3 ... Symbolization dictionary, 4 ... Symbolized data storage processing part, 5 ... Symbolized data historical file, 10 ... Man-machine device, 10a ... Display, 10c: keyboard, 11: symbol inverse conversion processing unit, 12: control data output processing unit, 15: time series data support promotion unit, 17: control arithmetic processing unit, 31: symbol Fuzzy data output processing unit, 32: inference engine, 33 ... Knowledge base, 34 ... Knowledge acquisition input processing unit, 35 ... Abnormal treatment guide output unit, 36
... plant status inquiry processing unit, 37 ... a little more signal reception processing unit, 38 ... symbolized integrated information input unit, 39 ...
Symbolized integrated information conversion unit, 40: Symbolized integrated information file.

Continuation of the front page (56) References JP-A-61-250518 (JP, A) JP-A-58-124911 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01M 19 / 00 G01D 21/00 G05B 23/02 301

Claims (8)

(57) [Claims] 1. A method for acquiring data representing a state of a process.
To convert the data into data that can be processed by the subsequent processing means.
Process data capture processing means and process status
Representation of process status symbolically based on data
Encoding processing means for generating encoded data to be encoded, and encoding
Means for storing encoded data, and a designated
The data in the specified range from the symbolized data storage means
And display processing means for outputting as display data,
Display data on the screen and receive external instructions.
Input / output means for receiving, and the symbol processing means includes data representing a state of the process;
Compatibility with symbolized data that symbolically represents the state of the process
It has a function to determine the degree of conformity, and the encoded data storage means checks the degree of conformity together with the encoded data.
Process operation support characterized by accumulation
system. 2. The display processing device according to claim 1, wherein
Display the degree of relevance together with the symbolic representation of the digitized data.
The process operation support system. 3. The method according to claim 2, wherein the symbol processing means includes a standard vector data and a characteristic of the standard vector data.
Symbolizing process having symbolic representation data for symbolically representing a symbol
Generates a vector sequence by performing a linear approximation on the input data and a scientific dictionary
Then, for this vector series, a dictionary for symbolization processing is used.
Search for similar standard vector series
Extract and record the symbolic expression data corresponding to the standard vector
A symbol processing unit that performs the coding process, a linear approximation vector sequence of the input data, and a standard vector system.
A similarity calculation processing unit that calculates the similarity with the column, and matches the similarity to the input data of the symbolized data.
And outputs it as a degree, process operations support system
M 4. The encoded data according to claim 2,
Knowledge acquisition to generate inference rules described using degree
A process operation support system further comprising a profit input means. 5. The knowledge acquisition input means according to claim 4,
Fuzzy representation of the encoded data representing the state of the process
Corresponding to the membership function of inference
Correspond to the membership function, and
Accepts input from the
Process operation support system that has the function of creating
Tem. 6. The method according to claim 4, wherein the knowledge acquisition input means includes a box defining an event occurring in the process, and a box defining the box.
And the connection symbol that defines the connection relationship between the boxes.
It has the format of the graph to be displayed in advance, and specifies that the box belongs to the IF section,
And accepts designation of belonging to THEN section,
For boxes belonging to the department, indicate the site where the event occurred
Name, encoded data indicating the content of the event, and its conformance
And at least input of fitness conditions
In the box belonging to the THEN section, the inference result
For boxes that accept input and for which events are defined,
Specify the connection relationships and the propagation direction of those events
Has a function to accept at least the input of the connection symbol
Process operation support system. 7. The method according to claim 6, wherein the direction of the event is specified.
The connection symbol to connect the boxes in parallel
As an OR connection and connecting the boxes in tandem
Are defined as AND connections, respectively.
Process operation support system. 8. The inference rule according to claim 5, 6 or 7,
Knowledge base that stores the
Engine that infers the state of a process using logic rules
And information indicating the state of the process.
A table that asks for information that cannot be detected and prompts for the information.
Process for accepting input of corresponding information
Means for inquiring about the situation, wherein the knowledge base contains encoded data and
And the IF section described by its conformity and the corresponding
Represents the state of the process necessary to determine the state
Inquiry of information that cannot be detected by the sensor
Process status with a THEN section where the matter is described
The query rules are stored and the inference engine is based on the knowledge-based query rules.
Machine that performs inference and outputs necessary inquiries
Has a capacity, process status inquiry means are output from the inference engine
Inquiry items displayed via input / output means
A process operation support system having functions .