JPH0493612A - Trend graph display device with guide - Google Patents

Abstract

PURPOSE:To enable the operation condition of a plant to be objectively judged without reference to an operator's judgement criterion by quantifying the aforesaid criterion, automatically inserting comment data and displaying the operation pattern of the highest similarity and a succeeding operation pattern during the operation of the plant. CONSTITUTION:The operation condition of a plant 1 is detected with a detection device G14 constituted with various detectors. Output signals from the aforesaid plurality of detectors are inputted to a feature judgement device G13. This device G13 is constituted with an input section for recognizing the signals as patterns, a memory section for storing the patterns, a processing section for collating the entered patterns with a plurality of previously stored patterns and outputting the similarity thereof, and an inference mechanism 12 for determining a comment corresponding to the patterns on the basis of the similarity, and sending a character to a display device G6. Furthermore, there are added a function for causing a feature judgement mechanism to operate, and confirming the output of a proper judgement result, and a teaching device G16 for changing pattern memory data, when the plant 1 is modified or the arrangement of the detectors is altered.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is a man-machine device for plant monitoring and control equipment that automatically provides guidance regarding the plant operating status when displaying detected data from the plant in a trend graph. Concerning methods for inserting data and predicting future trends.

[Conventional technology]

In the conventional apparatus, as described in Japanese Patent Application Laid-Open No. 63-221218, character data is manually input as a comment along with process data and displayed on the screen. Further, as described in Japanese Unexamined Patent Application Publication No. 1-4791, a trend history of process data is saved in a file, and by redisplaying this, an operator can judge the current operating state.

[Problem to be solved by the invention]

In the above-mentioned conventional technology, comment data must be input manually, which increases the burden on the operator. Furthermore, when inputting the plant operating state as a comment, there is a problem in that the criteria for judgment vary depending on the operator.

The purpose of the present invention is to be able to insert comment data representing plant operating status without increasing the burden on operators, to predict and display future plant operating data, and to objectively display future plant operating data without increasing the burden on operators. An object of the present invention is to provide a trend graph display device capable of determining operating conditions.

[Means to solve the problem]

The above objective is achieved by quantifying the operator's judgment criteria and automatically inserting comment data. Also,
For predictive display of driving data, a certain driving pattern and a driving pattern that appears following it are stored as a pair, and during actual driving, the driving pattern with the highest degree of similarity is searched for, and the driving pattern that follows this pattern is displayed. This is achieved by

[Effect]

Skilled operators extract characteristic patterns from control variables and make ambiguous decisions. Similarly, the confidence of the characteristic pattern is determined by calculating the sum of products of the control variables and passing the result through a nonlinear circuit, and the most suitable driving state pattern is inferred from the confidence of each characteristic pattern. Output as guide characters. Thereby, the inference method operates like a skilled operator and can obtain good inference performance.

Further, even if the know-how of a skilled operator is stored as knowledge as it is and inference is made using the knowledge, the same inference performance as described above can be obtained.

〔Example〕

An embodiment of the present invention will be described below.

FIG. 1 is a block diagram of the present invention. The operating state of the plant I is determined by output signals from a plurality of detectors from a device detection device G14, which is composed of various detectors respectively arranged, and is input to a characteristic determination device G13. The feature determination device G13 includes an input unit that recognizes the plurality of input detector output signals as a pattern, a memory function that stores the input pattern, and a plurality of patterns that are stored in advance of the input pattern. and a processing unit that outputs the degree of similarity with the previously stored pattern, and an inference mechanism 12 that determines a comment corresponding to the pattern from the degree of similarity and sends the character to the display device G6. .

In addition, before storing the pattern in the pattern storage function of the feature determination device G13, the operator sets a comment output for the input in advance, operates the feature determination mechanism, and recognizes that a correct determination result will be output. and the ability to
The structure is such that a teaching device G16 can be added to change the pattern storage contents in advance when the plant 1 is changed or the arrangement of the detectors is changed.

On the other hand, the conventional system does not have the feature determination device G13, and generally only the detector IG14 sends comments to the display device G6, and comments regarding the plant operating status are
Operators had to make decisions based on changes in detector signals and enter them manually. Feature determination device G as in the present invention
By adding 13, it is possible to insert comments on the plant operating status without increasing the burden on the operator.
This also has the effect of eliminating variations in judgments among operators.

Next, an example in which the present invention is applied to a trend graph display device will be explained with reference to FIG.

The plant 1 has a plurality of detectors 6 for communicating the operating status.
2 is attached, and the output of the detector 62 is periodically taken in by a process signal input device 71 and stored in a trend data storage device 72 in chronological order. The trend data storage device 72 holds data from the latest data to the past n points, for example, up to 500 points, regarding output signals from a plurality of detectors. In the guide storage device 73, comments corresponding to the trend data storage device 72 are stored as characters, and as shown in the screen display example of FIG. 75. Here, the trend graph 721 displays the contents of the trend storage device 72.

The method of storing comments in the guide storage device 73 includes an inference mechanism 12 that generates comments corresponding to a pattern, and a decision made by the inference mechanism 12 to which pattern the operating state of the plant belongs among a plurality of pre-stored patterns. Pattern recognition mechanism 1 that outputs the certainty of the pattern
3. A storage mechanism 15 that stores and outputs the contents of the trend storage device 72 at a certain moment for the pattern detection mechanism 13.
, and the information in the storage mechanism 15, the pattern recognition mechanism 1
It is composed of a learning mechanism 16 that changes the parameters of No. 3 through learning.

FIG. 4 shows a detailed diagram of the shape pattern processing mechanism 13. The outputs of the trend data storage device 72 and the storage mechanism 15 are input to the input cells 17 and 18 of the pattern recognition mechanism 13, and the input cells 17 convert the input signals into function values and output them to the intermediate layer 19. The output of the input cell input to the intermediate layer 19 is input to the cell 20.21 of the intermediate layer 19. The signal inputted to the cell 20 as the output of the input cell 17 is multiplied by WLI by the weighting function 23 and inputted to the adder 24, and the output of the input cell 18 is inputted to the adder 24 via the weighting function 26 and added. The function unit 24 adds the outputs of the weighting functions 23 and 26, and inputs the result to the function unit 25.
The function unit 25 performs a linear or nonlinear function operation, and outputs the result to the next intermediate layer 27. Note that the cell 20 is composed of the weighting functions 23 and 36, an adder 24, and a function unit 25.

Similarly, the outputs of input cells 17. It is output to the next stage intermediate layer 27 via the function unit 30.

The intermediate layer 27 has the same structure as the intermediate layer 19, and the output of the intermediate layer 19 is used instead of the output of the input layer 17.48.

Here, the weights of weight functions 23, 26.28 are W 2 J
Expressed as , W″1J is the weight to be applied to the i-th output of the k-1th hidden layer (input cell when k=1) in the i-th cell of the second hidden layer. show.

As described above, the signals input to the pattern recognition mechanism 13 are transmitted to the input cell 17.18, the intermediate layer 19.27.2 of multiple stages.
9, and an output layer 30 which is obtained by removing the weighting function and adder from the cells of the intermediate layer. Note that the input layer 31 represents a collection of all input cells 17 and 18.

The characteristics of this pattern recognition mechanism 13 are that it only requires a simple product-sum operation, and there is no repetitive operation such as feedback, and that each product-sum term in the intermediate layer can be processed in parallel when implemented in hardware. High-speed calculation is possible.

Next, the processing result of the pattern processing mechanism 13 is outputted to the guide storage device i73 via the processing mechanism shown in FIG.

That is, the output of the pattern recognition mechanism 13 is the output of the inference mechanism 12.
The guide character determination means 32 provided at The guide character determining means 32 selects the most effective processing mechanism for processing the input signal from among the plurality of internally prepared processing mechanisms, executes the process, and outputs the guide character.

FIG. 6 shows the configuration of the guide character determining means 32. Guide character determining means 32 receives a signal from pattern recognition mechanism 13 and activates control mechanism 141 .

The control mechanism 141 uses the knowledge base 36 to determine which inferences to activate, depending on the type of problem. That is, the control mechanism 141 activates the production inference mechanism 142 when it is necessary to find a cause using a syllogistic method, activates the fuzzy inference mechanism 143 when there is an ambiguous factor, and solves problems that have a certain degree of framework. For frame inference mechanism 14
4, the semantic net inference mechanism 145 is activated for problems in which relationships such as causal relationships and device configurations are network-like, and the semantic net inference mechanism 145 is activated for problems in which the objects to be diagnosed operate in a temporal order. , the script inference mechanism 146 is activated. Furthermore, the control mechanism 141 activates the optimization arithmetic mechanism 111 to quickly find an optimal solution for empirical problems that cannot be solved by the various reasoning mechanisms described above, can memorize patterns, extract features, and provide answers. The feature extraction/answer mechanism 110 (consisting of a Rumelhart-type neuron computer) is activated to solve a problem requiring the following. The processing results of the guide character determining means 32 are outputted to the inference mechanism via the control mechanism 141.

FIG. 7 shows the structure of the knowledge base 36, which is the knowledge necessary for inference. The knowledge base is knowledge input from the outside based on the experience of a control expert, etc. 1!106 is a production rule 147 for performing reasoning in a syllogistic manner;
Fuzzy rules 148 are knowledge for making inferences based on ambiguous information, knowledge frames 1492 can be described in a certain framework such as the configuration of parts to be diagnosed, etc. A network is created by summarizing relationships between parts and common sense relationships. Semantic network 1502 organized in the form of a script 151 that organizes and memorizes certain tasks when the diagnosis target sequentially performs them, and other knowledge 154 that cannot be described in the above knowledge 147 to 153. is memorized.

FIG. 8 shows an explanatory diagram of the operation of the guide character determining means 32. The processing of the control mechanism 141 is performed by the pattern recognition mechanism 13. Processing step 200 for organizing information from the storage mechanism 15 and converting it into data that can be used for the following processing, step 2 above.
A repetitive process step 201 in which the data prepared in step 00 is retrieved until it is exhausted and passed to step 202, a determination step 202 for determining the inference mechanism and process to be activated from the information collected in step 201, and various inference mechanisms 142 to 146. A feature extraction response mechanism 110, an optimization calculation mechanism 111, a general control mechanism 203 that executes algorithms for classical control such as PID control and modern control such as multivariable control, and a It consists of a termination processing step 204 in which flags are reset and the like.

Here, the role of each processing mechanism will be described. The production mechanism 142 is suitable for inference in which an expert operator assembles logical relationships using fragmentary production rules. Fuzzy reasoning 143 determines the amount of operation by quantifying the operator's ambiguous knowledge that cannot be quantified, such as the operator moving the actuator slightly if the state of interest of the controlled object changes, so that it can be processed by a computer. suitable for.

The frame inference mechanism 144 uses the knowledge of frames that describe relationships between control devices, and performs operations based on the relationships between these devices when returning to the original state when the state of the controlled object of interest changes. It is suitable for determining the amount of processing to be performed for each related device.

Since the semantic inference mechanism 145 organizes and systematizes the frames, which are the fragmentary knowledge, to create a network, it can determine the influence of the operation result of a specific actuator, and can calculate the compensation system. Suitable for assembling.

Since the script inference mechanism 146 makes inferences based on procedural knowledge when a specific state occurs, it is suitable for sequence control-like control where a fixed procedure must be followed in the event of a failure or the like.

Further, the feature extraction/answer mechanism 110
R mechanism 13. When the input pattern of the storage mechanism 15 and the input pattern are input, the inference mechanisms 142 to 146
If the relationship between the outputs produced by the inference mechanisms 142 to 146 is learned in advance, the same result can be outputted at high speed, unlike when the inference mechanisms 142 to 146 perform inference to determine the output.

FIG. 9 shows an explanatory diagram of the operation of the production mechanism 142. The production inference mechanism 142 activated by the control mechanism 141 retrieves from the control mechanism 141 the input process 34 stored in the memory at the time of activation, and the information stored in the input process 34 one by one, and if there is no pattern information in the memory. At times, the termination determination mechanism 35 is executed to terminate the processing of the production inference mechanism 142. Using the type of pattern extracted by the termination determination mechanism 35 and its confidence level, rules are retrieved one by one from the knowledge base 36, and in process 37, the type of input pattern and the antecedent part of the rule are compared. . Using the comparison results, step 38 executes the next process 39 if they match, and executes step 37 if they do not match. Step 39 replaces the input with the conclusion part of the rule when there is a match. At this time, the confidence level is handled using the mini-max theory, and is replaced with the minimum value or maximum value before replacement. In step 40, if the conclusion part of the replaced rule is an operation command, step 41 is executed, and if the conclusion part does not match, step 37 is executed to further perform inference.

FIG. 10 shows an input switching device 125 necessary for the learning.
The configuration is shown below. The input switching device 125 outputs either the output of the trend data storage device 72 or the output of the learning mechanism 16 to the input layer 31 using a switch mechanism 156 controlled by a learning mechanism. The state of the switch mechanism 156 in FIG. 10 indicates a state in which learning is performed.

FIG. 11 shows the configuration of the learning mechanism 16. The learning mechanism 16 includes an input pattern generating mechanism 45. Output pattern generation mechanism 4
7. Output matching mechanism 46. and a learning control mechanism 48. The output matching mechanism 46 outputs the output 01y oat On of the distributor 139 for outputting the output of the output layer 30 to the inference mechanism 12 and the matching mechanism 46, and the output 071 g OT i of the output pattern generation mechanism 47.
Adders 161, 162, and 163 calculate the difference from T OT n as deviations el and eI.

en, and outputs it to the learning control mechanism 48.

Note that the output Q 11 Q i p OLl of the distributor 139
is generated by the output of the input pattern generation mechanism 47 being output to the input layer 19 of the pattern recognition mechanism 13 (Itumelhart type neurocomputer). At this time, the input pattern generation mechanism 45 and the output pattern generation mechanism 47 are controlled by the learning control mechanism 48.

Figure 12 shows the weight function W I J2 in the learning process.
3 shows the relationship of the learning control mechanism 48 of No. 3. The adder 161
In response to the deviation ek, which is the output of
The value of J23 is changed in a direction that reduces the deviation.

FIG. 13 shows a processing overview 170 of the learning control mechanism 48. When the learning mechanism 16 is activated, a process 170 of the learning control mechanism 48 is activated. The processing 170 includes the input pattern generation mechanism 45. Pre-processing 171 for activating the output pattern generation mechanism 47 and generating the input as the teacher signal and the desired output, and performing the following steps 173 and 174 until the value of the deviation eh or the sum of the squares of the deviations falls within the allowable range. Step 172 of repeating ゜175. A step 174 of sequentially extracting the target intermediate layer from the intermediate layer close to the output layer 30 toward the input layer 31, a step 174 of sequentially extracting the target cells in the intermediate layer, and cells extracted in the direction where the deviation ek becomes smaller. It consists of step 175 of changing the weight function wtr23 of , and step 176 of ending the learning process.

By providing such a learning mechanism, a new phenomenon that has not been considered until now occurs, and once a countermeasure against it has been determined, it has the characteristic of being able to reflect that knowledge.

FIG. 14 shows the configuration of the storage mechanism 15 of FIG. 2. The storage mechanism 15 includes the inference mechanism 12. A memory element 49 to which the output of the trend data storage device is input; a memory element 50 to which the contents of the memory fi 49 are transferred after a certain period of time; and a memory element 51 to which data is sequentially transferred to the memory elements and reached after a certain period of time has elapsed. configured, each memory element 49
．． The contents of 50.degree. 51 are input to the pattern recognition mechanism 13.degree. learning mechanism 16 via an arithmetic mechanism 501 for performing pattern differentiation, integration, etc.

With this storage mechanism 15, the trend data storage group W72
．． Changes in the reasoning mechanism 12 over time can be taken into account.

The operation of the trend graph display device using the configuration described above will be described below using a specific example.

The initial value of the weight function W≦-28 of the intermediate layer 19, 27, 29 of the neurocomputer that constitutes the pattern recognition mechanism 13 is initially set to a random number or an appropriate value, such as a value that the weight relationship can take (0 to 1. If it is 0, set it to half (0, 5) of ). At this time, for example, even if an A-type time series data pattern generated by the input pattern generation mechanism 45 in FIG. Also, the output layer 3
The probability that the output of the 0 output line 71 is of type B is not zero.

Therefore, the learning mechanism 16 corresponding to the output line 70 of the output layer 30
The output line 72 of the output pattern generation mechanism 47 is 1, and the output line 7 of the cover pattern generation mechanism 47 corresponding to the output line 71 is
Output the output of 3 to zero.

Upon receiving these outputs, the output matching mechanism 46 receives the deviation between the ideal output (output pattern generation mechanism 47) and the output of the pattern recognition mechanism 13, and the learning control mechanism 48 calculates the weight function of the pattern recognition mechanism 13. The magnitude of W L J is changed in a direction in which the deviation decreases in proportion to the magnitude of the deviation. A typical example of this algorithm is the steepest slope method.

The weight of the weight function is sequentially changed according to the process shown in FIG. 13, and when the sum of squares of eh shown in FIG. 12 falls within the allowable range, the operation of the learning mechanism 16 ends.

After learning is completed, when the same waveform as the output pattern of the input pattern generation mechanism 45 of FIG. 17 is input from the trend data storage device 12 of FIG. and outputs zero from the output line 71 of the output layer 30.

Next, if the waveform shown in FIG. 16, which is named type B, is input, and learning has not yet been completed, the pattern recognition mechanism 1
The output of the output line 71 expressing the convex shape of 3 is 1, and the other output cuffs 0 do not become 0. Therefore, as described above, when a typical convex pattern is used as an input signal, the output of the output pattern generation mechanism 47 is connected to the output line 71.70.
The values corresponding to the outputs of are respectively set to 1.0. The learning mechanism 16 changes the weight function W I J, and when the learning is completed, the pattern recognition mechanism 13 receives the 16th
When the waveform of type B in the figure is input, the output line 71 of the output layer 30 in FIG. 15 becomes 1, and the output line 70 becomes 0.

As a result, the waveform shown in FIG. 17(a) is the pattern recognition mechanism 1.
3, and its output is the B-type waveform input in advance from the output layer 30 as described above.
1, the degree of resemblance to the waveform is output with a certainty of 40%, and at the same time, the degree of similarity to the waveform is output with a certainty of 50% from the output line 70 indicating that it is an A-type waveform.

FIG. 18 shows an example of a production rule or a fuzzy rule (corresponding to production rule 47 and fuzzy rule 48 in FIG. 7).

When the pattern recognition mechanism 13 obtains an output with a certainty factor of type A 50%, it is compared with the premise of the production rule and matches the type A rule 80.

Furthermore, if the output is obtained as a confidence level of type B 40%, it matches the type B rule 82. The conclusions drawn from the rules, such as "small vibrations" and "undulations", represent the state and are not conclusions as a guide. Then, if we further check the conformity level based on the certainty level, it will match the rule 84 of ``Small shift worker and undulations'', and we can obtain a guide of ``Unstable'' in the conclusion section. The obtained guide is stored in the guide storage device 73 and displayed on the display device together with the trend graph as shown in FIG.

The above explanation has been about the case where the signal from a single detection end is input as a time series pattern, but it also applies to the case where the signals from multiple detection ends are input as a time series pattern while interacting with each other. This method can be applied similarly.

Next, a method of displaying a trend graph of a time series pattern that will occur next to a certain time series pattern will be described.

In the learning mechanism 16 shown in FIG. 15, the number of output lines of the output layer 30 is the same as the number of input lines of the input layer 31.

The learning operation is repeated for a certain time series pattern generated by the input pattern generation mechanism 45 so that the actual value of the next generated time series pattern appears in the output layer 30. This is similarly performed for various time-series patterns that may appear.

After learning is completed, when time series data from the present to the past n points (0 is 500, for example) is input from the trend data storage device, the pattern recognition mechanism 13 calculates the time that will occur next on the output line of the output layer 30. Output series data. FIG. 19 shows the configuration of a trend graph display device for the predictive trend graph display technique.

The time series data output from the pattern recognition mechanism 13 is
The predicted trend data storage device 176 stores the predicted trend data. Further, the output of the trend data storage device 72 becomes the input of the pattern recognition mechanism 13. The trend data storage device 72 and the predicted trend data storage group [76 contents are displayed in the display control device 7
4 is displayed on the display device 75. Fig. 2o shows an example of displaying a trend graph. Trend data storage device 1
The contents of 72 are trend graphs 721. The contents of the predicted trend data storage device are displayed as a trend graph 761. Here, the data in the b section of the trend graph 761 is obtained from the time series data in the a section of the trend graph 721 through the pattern recognition mechanism 13.

As explained above, according to this embodiment, by learning or setting one time series pattern and corresponding judgments and rules in advance, comments expressing the plant operating status are automatically displayed together with the trend graph. Therefore, the burden on the operator does not increase. Also.

The above comments are not based on the judgment criteria of individual operators,
This has the effect of being able to objectively understand the operating status of the plant. Furthermore, since the plant operating state is predictively displayed, it is possible to take appropriate measures before an abnormal state occurs in the plant, for example.

〔Effect of the invention〕

According to the present invention, it is possible to categorize time-series data from a plurality of detection terminals that represent the plant operating state, determine the characteristics of the waveform, and display appropriate comments regarding the plant operating state together with the trend graph. It becomes easier to understand the status.

Moreover, this does not increase the burden on the operator, and has the effect of allowing objective judgments to be made without depending on the operator's individual judgment. Furthermore, since trend graphs can be displayed predictively, appropriate measures such as stabilizing plant operation and avoiding abnormalities can be taken in advance.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2 is an overall block diagram of a trend graph display device, Fig. 3 is a display example, Fig. 4 is a pattern recognition mechanism diagram, and Fig. 5 is an inference mechanism diagram. , FIG. 6 is a configuration diagram of the guide character determining means, FIG. 7 is a knowledge base configuration diagram, FIG. 8 is an explanatory diagram of the operation of the operation amount determination means, FIG. 9 is an explanatory diagram of the operation of the production mechanism, and FIG. 10 is an explanatory diagram of the operation of the production mechanism. The configuration of the input switching device, Figure 11 is the configuration of the learning mechanism, Figure 12 is a diagram of the relationship between the learning control mechanism and node weight functions, Figure 13 is a basic processing diagram of the learning control mechanism, and Figure 14 is the storage mechanism. 15 is an explanatory diagram of the operation of the pattern recognition mechanism.
The figure shows an example of an input pattern, Figure 17 is an explanatory diagram of the output of the pattern recognition mechanism, Figure 18 is a diagram showing an example of a production rule, Figure 19 is an overall configuration diagram of a trend graph display device with prediction display, Figure 20 is a display example. 1... Plant, 12... Inference mechanism, 13... Pattern recognition mechanism, 16... Learning mechanism, 19.27.2
9... Middle layer, 30... Output layer, 31... Input layer,
72...Trend data storage device, 73...Guide storage device, 75...Display device, 76...Prediction trend data storage Fig. 1 Fig. 16 Fig. 17

Claims (1)

[Claims] 1. A process signal input device for inputting a detector signal that conveys the operating status of the plant, a data storage device for storing the signals in time series, and a device for displaying the stored data as a trend graph. A trend graph display device comprising a display device, a feature determination mechanism that determines the characteristics of the plant operating state from data stored in time series and outputs it as characters, symbols, etc., and a guide storage device that stores the characters, symbols, etc. , a trend graph display device comprising a display device that displays characters, symbols, etc. together with a trend graph. 2. In claim 1, the feature determining mechanism includes means for storing a plurality of combination patterns of a plurality of detection signals of the detection device and a determination result of a plant operating state corresponding to the plurality of combination patterns. , characterized in that a new output pattern of a plurality of detection signals of the detection device and a previously stored output pattern are stored, and characters, symbols, etc. corresponding to the storage pattern with the greatest degree of similarity are stored in a guide storage device. Trend graph display device. 3. In claim 1, the feature determination mechanism is comprised of a pattern recognition mechanism that determines the similarity of the pattern of the output signal from a combination of a plurality of detection signals from the detection device. Trend graph display device. 4. In claim 3, the pattern recognition mechanism comprises:
a first stage consisting of an input layer that captures a combination pattern of a plurality of output signals of the detection device; and a plurality of nodes that add weights to the output signals of the input layer and map the results with a specified function; and the output signals of each node of the first intermediate layer are input signals, multiplied by weights, and added;
It has multiple stages of another intermediate layer composed of multiple nodes that map the result using a specified function, and the final stage of the other intermediate layer is an output layer that outputs the pattern similarity as a judgment result. A trend graph display device characterized in that characters, symbols, etc. corresponding to a pattern having the greatest degree of similarity are stored in the guide storage device. 5. Claim 1 provides an instruction device having a mechanism for determining whether the determination result of the feature determination device is good or bad, a means for notifying the determination result to the outside, and a means for changing the content of the feature determination device. A trend graph display device characterized by: 6. In claim 1, a storage mechanism is provided to store the output signals of the detection device in a chronological order, the output of the detector and the output of the storage mechanism are input to the feature determination device, and the characteristics A trend graph display device characterized in that the judgment device is capable of making comprehensive judgments that also take into account temporal changes. 7. The trend graph display device according to claim 3, wherein a Rummelhart neurocomputer is used for the pattern recognition mechanism. 8. Claim 7, further comprising a learning mechanism for applying an input pattern to the pattern recognition mechanism and changing the weight of each node of the pattern recognition mechanism in advance so that the pattern recognition mechanism has an ideal output. A trend graph display device characterized by: 9. In claim 8, the learning mechanism comprises an input pattern generation mechanism inputting to an input layer of a pattern recognition mechanism, an output pattern generation mechanism generating an ideal output of the pattern recognition mechanism, and an output pattern of the pattern recognition mechanism. , a comparison mechanism for determining a deviation of the output pattern generation mechanism, a learning that generates a command to change the weight of a node in the pattern generation mechanism based on the comparison result, and an operation command for the input pattern generation mechanism and the output pattern generation mechanism. A trend graph display device comprising a control mechanism. 10. Claim 3, further comprising a storage mechanism for storing the detection signals of the plurality of detectors in a chronological order, and inputting the output of the storage mechanism to the pattern recognition mechanism in parallel with the detection signals of the detectors. A trend graph display device characterized by having a configuration in which: 11. In claim 6, the feature determining mechanism has means for storing time series data in a certain time range of the detection signal of the detection device and the next time series data following the time series data. Then, the new time-series data pattern of the detection signal of the detection device is compared with the previously stored time-series pattern, and the next time-series data for the stored pattern with the highest degree of similarity is displayed as predicted time-series data in a trend graph. A trend graph display device characterized by displaying a trend graph. 12. The trend graph display device according to claim 11, wherein a prediction warning such as an upper limit value and a lower limit value for the predicted data is output based on the predicted time series data.