JPH0464014A - data monitoring equipment - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a data monitoring device for monitoring and analyzing the operating status of a process.

[Conventional technology]

As described in Japanese Patent Application Laid-Open No. 62-254278, the conventional device sequentially processes time-series data and stores the result as follows.

In the latest first period, time series data is stored for each sampling. The second period that occurred in the past is divided into intervals with the same time width as the first period, and for each interval, the maximum value is calculated from the time series data of multiple samples included in that interval. Extract and store representative values such as minimum value and average value. Further, the third period measured in the past is divided into sections having the same time width as the second period, and for each section, the maximum value is determined from among the plurality of representative values included in that section.

Extract and store representative values such as minimum value and average value.

Similarly, representative values of time-series data are sequentially stored in the past.

Furthermore, in Japanese Patent Laid-Open No. 140384/1984, time series data is stored as follows.

A process is performed to determine whether the input data is a peak value, and if it is a peak value, it is stored, and data other than the peak value is not stored.

[Problem to be solved by the invention]

In Japanese Patent Application Laid-open No. 62-254278, the prior art mentioned above stores less time-series data per time as the time goes further into the past. Therefore, when monitoring and analyzing the operating state of a process, when the analysis period is long, the closer to the present time the more data can be restored per hour, which allows for a more accurate understanding of changes in the operating state of the process over time. However, there was a problem in that the processing time for monitoring and analysis was long.

Additionally, the longer the analysis period goes into the past, the less data is restored per hour, so although the processing time for monitoring and analysis becomes shorter, it is not possible to accurately understand temporal changes in the operating state of the process. There was a problem.

Furthermore, in Japanese Patent Application Laid-Open No. 63-140384, the peak value and the time width between the peak values of the input data are stored, which saves storage space and makes it possible to grasp temporal changes with respect to the peak value. There is a problem in that it is not possible to grasp detailed temporal changes in the operating state of the process from peak value to peak value.

The present invention can monitor and monitor the operational status of processes regardless of the length of the analysis period and whether it is close to the present or far (back in the past).
It is an object of the present invention to provide a data monitoring device that can shorten analysis processing time and accurately grasp temporal changes in the operating state of a process.

[Means to solve the problem]

In order to achieve the above purpose, there is provided a means for storing all the raw time series data for each sampling, extracting and storing the representative value for each predetermined period including a plurality of pieces of raw time series data, and for a designated period. When restoring time-series data, a means is provided for selecting and restoring either raw time-series data or time-series data of one representative value, depending on the length of the specified period.

[Effect]

The means for storing all the time-series data for each sampling in the storage device correctly stores temporal changes in the operating state of the process, regardless of whether it is near or far from the present. The means for extracting a representative value of a plurality of pieces of raw time-series data every predetermined period and storing it in a storage device stores an approximate time trend of a temporal change in an operating state of a process. When restoring time-series data for a specified period from a storage device, there is a method for selecting and restoring raw time-series data or representative value time-series data depending on the length of the specified period. When the specified period to be restored is long, time-series data of representative values are restored to reduce the number of restored data, shorten the processing time, and grasp the general trend of temporal changes in the operating state of the process. From among these, the period for which you want to understand the details is limited, and when that limited period becomes shorter, the raw time-series data is restored to accurately understand temporal changes in the operating state of the process. At this time, since the specified period is limited, even if the raw time series data is restored, the number of restored data will be small and the processing time will be shortened.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG.

In FIG. 1, process data such as pressure and temperature of a process 1 is input to a process data input section 6 in a computer 3 by a process input/output device 2 every second.

The process data input unit 6 stores the process data in the raw process database 9 in the storage device 4 as data every second like the raw process database 1 in FIG. 2, and extracts representative values. input into section 7.

The representative value extraction unit 7 extracts a representative value containing a plurality of pieces of raw process data every predetermined period, and stores it in a representative value process database 8 in the storage device 4 like the representative value process database 2 in FIG. memorize in a format. A specific example of representative value extraction will be explained with reference to FIG.
March 2, 1989 1:00:1 to March 2, 1989
The maximum and minimum values of the 3599 pieces of raw data up to 1:59:59 on March 2, 1989, and the raw data at 2:00:00 minutes and 0 seconds on March 2, 1989 are stored as instantaneous values in the order of their occurrence. Next, the daily representative value 4 is, for example, the representative value for March 2, 1989.
The maximum value of 71 representative values up to 4 o'clock, excluding the last instantaneous value at 24 o'clock.

The minimum value and the instantaneous value at 24:00 on March 2, 1989 are stored as instantaneous values in the order of occurrence. Similarly, monthly representative value 5
For example, 1989 is the representative value for March 1989.
The last 3 from March 1st to March 31st, 1989
The maximum and minimum values of the 92 representative values excluding the instantaneous values of one day and the instantaneous values of March 31, 1989 are stored as instantaneous values in the order of occurrence.

Next, in order to monitor and analyze the process state, the data selection unit 10 retrieves the representative value process database 8 from the storage device 4 when the designated period of monitoring and analysis is long, and when it is short, the data selection unit 10 stores the representative value process database 8 from the storage device 4. The process database 9 is selected, the process data is restored, and the process data display unit 11 displays the process data on the CRT display device 5. As a specific example, a case will be described with reference to FIGS. 3 to 6, in which it is estimated whether or not there is any abnormality at measurement points 1 and 2 that are related to each other for a certain year. First, FIG. 3 shows a display example of 36 pieces of data from January to December 1989 from monthly representative value 5 in the representative value process database of FIG. 2. 1 and 2 in FIG. 3 represent graphs of measurement points 1 and 2, respectively. Here, both measurement points 1 and 2 have increased in March. Figure 4 shows the increase in March 1989 from the daily representative value 4 in the representative value process database 2 in Figure 2.
An example of displaying 93 pieces of data from March 1, 2013 to March 31, 1989 is shown. 1 and 2 in FIG. 4 represent graphs of measurement points 1 and 2, respectively. Here, measurement points 1 and 2 are both 2
It is rising by the day. Figure 5 shows the rise in 19893.
An example of displaying 72 data from 1:00 on March 2, 1989 to 24:00 on March 2, 1989 from time representative value 3 in representative value process database 2 in FIG. 2 for the 2nd day of the month is shown. . 1 and 2 in FIG. 5 represent graphs of measurement points 1 and 2, respectively. Here, both measurement points 1 and 2 rise between 1 o'clock and 2 o'clock. Figure 6 shows the rise from 1:00 to 2:00 on March 2, 1989, from raw process data 1 in Figure 2, from 1:00:1 on March 2, 1989 to March 2, 1989. An example of displaying 360o pieces of data up to 2:00, 0:00 and 0:00 seconds is shown. 1 and 2 in FIG. 6 represent graphs of measurement points 1 and 2, respectively. Here, it can be seen that measurement point 1 starts to rise from around 4 minutes, and measurement point 2 starts to rise from around 8 minutes.

From the above, it can be inferred that measurement point 1 was the cause of the problem, which spread to measurement point 2.

According to this embodiment, the processing time for monitoring and analyzing the operating state of a process can be reduced. For example,
When estimating whether there are any abnormalities in one year's worth of data at measurement point 1, which is sampled at a 1-second cycle, 31,536
,000 pieces (=60 pieces/min x 60 minutes/hour x 24 hours/
I need to search for data (days x 365 days/year),
In the present invention, 3,801 pieces (= 3 pieces/month x December/year + 3 pieces/day x 31st July + 3 pieces/hour x 24 hours/
It is only necessary to search for data (day + 60 pieces/minute x 60 minutes/hour), which can reduce the time to about 1/10,000. In addition, detailed data can be monitored using raw process data with a 1-second cycle.
Because of the analysis, it is possible to accurately understand temporal changes in the operating state of the process.

〔Effect of the invention〕

According to the present invention, in the process of monitoring and analyzing the operating state of a process, a raw process database is used.

By having two types of databases, a representative value process database, it is possible to shorten the processing time for monitoring and analysis, regardless of whether it is close to the present or far away (measured in the past), and to understand temporal changes in the operating status of the process. This has the effect of allowing accurate understanding.

[Brief explanation of the drawing]

FIG. 1 is a system configuration diagram of an embodiment of the present invention. Figure 2 is a diagram showing a specific example of the data arrangement of the raw process database and the representative value process database, and Figures 3 to 6 are measurement points 1,
Example of trend graph display in 2. Figure 3 shows trends based on monthly representative values for one year.
Figure 4 shows the trend based on the daily representative value of the abnormality occurrence sequence, Figure 5 shows the trend based on the time representative value on the day of the abnormality occurrence, and Figure 6 shows the raw process of the 1-second period during the abnormality occurrence time period. FIG. 3 is a diagram showing trends based on data. 1... Process, 2... Process input/output device, 3.
...Electronic computer, 4. Storage device, 5. CRT display device, 6. Process data input section, 7. Representative value extraction section, 8. Representative value process database, 9.
. . . raw process database, 10. . . data selection section, 11. . . process data display section.

Claims (1)

[Claims] 1. Data consisting of a process input/output device that samples data that changes over time, an electronic computer that calculates and processes the time series data, a storage device that stores the time series data, and a CRT display device that displays the time series data. In the monitoring device, means for storing all raw time series data for each sampling in a storage device, and extracting a representative value thereof for each predetermined period including a plurality of pieces of raw time series data and storing it in the storage device; When restoring time-series data for a specified period from a storage device, a means is provided to select and restore either raw time-series data or representative value time-series data depending on the length of the specified period. Characteristic data monitoring device.