|
Process Data Acquisition – CHPDA -- Detailed
analysis of quality manage&industrial big data sources |
|||||
|
12 HDC - Hot
rolled high-frequency high-density Digital steel Coil |
|||||
|
6 DCC - Digital Coil
Conversion and full process quality management |
|||||
|
21 Research Form for PDA System Configuration in the
Steel Industry |
|||||
|
11
HDP - High frequency density and speed Data Platform construction |
|
||||
3 LTA - Long historical Trend Analysis
system
In addition to millisecond level data records, slowly changing production processes require monthly, quarterly, and annual curve analysis. The PDA system can collect multiple frequency reduction data at once, allowing for rapid analysis of long-term trends (Long Trend Analysis). It can be widely used in process industries such as blast furnaces, heating furnaces, petrochemicals, beer, and the analysis of parameters such as liquid level and temperature. The main significance of long-term trend analysis is twofold: firstly, to understand the trend and regularity of phenomena over time; The second is to predict the future development prospects and trends of phenomena. The reason why temporal data has long-term trends is because it is influenced by certain basic and decisive factors. The stronger the influence of these dominant factors, the more obvious the long-term trend. Therefore, by analyzing the long-term trends of time series data, the internal mechanisms of phenomenon development and changes can be grasped, and the effectiveness of policies and measures taken in the past can be evaluated; The third is to remove long-term trend components from the time series, in order to facilitate the decomposition of other types of influencing factors, such as seasonal changes, cyclical changes, and irregular changes. The main methods for measuring long-term trend values include: extended time interval method, moving average method, and least squares method. The extended time interval method refers to the method of eliminating fluctuations in the values of various indicators caused by accidental factors due to the short time interval by expanding the time of each indicator in the dynamic sequence, so that the smoothed dynamic sequence can significantly reflect the overall trend of phenomenon development and changes. The moving average method refers to the method of moving a dynamic sequence period by period to expand the time interval, while calculating the time series average for each indicator value of a new dynamic sequence that has already expanded the time interval, thereby forming a derived dynamic sequence from the moving average. The series of moving time series averages obtained through moving average are the trend values of their corresponding periods. The least squares method, also known as the least squares method, is a commonly used method for estimating regression model parameters. The basic principle is to require the sum of squares of the deviations between the actual value and the trend value to be the smallest, in order to fit an excellent trend model and determine long-term trends.
Figure 3.1 PDA Long term historical Trend Data Frequency Reduction A high-speed acquisition can obtain data of multiple frequencies, and changing the backup point to a formal point or adding or reducing points at the end of the connection does not affect the normal conversion work. LTAServer.exe is located in the PDA system file directory, and is converted based on the data files generated by PDA and saved in the directory specified by BigDataDir in Config.csv. The generated data files are saved in the directory specified by BigDataDir. LTAServer scan and convert at startup and every 12 hours interval thereafter. 3.1 Millisecond
level - hourly data analysis
Collect data every 10 milliseconds and generate a data file every 10 minutes. 3.2 Second
level - monthly data analysis
Collect data once a second and generate a data file once a day. 3.3 10
second level - quarterly annual data analysis
Collect data every 10 seconds, generate one data file every week, and 52 data files every year. 3.4 60
second level - annual data analysis
Collect data every 60 seconds, generate one data file in January, and 12 data files in a year. 3.5 7-days
data curve at the second level for a certain project
Figure 3.2 7-days data curve with reduced frequency to second level 3.6 Converting
long-term historical trend files of a blast furnace AB to PDA format
AB company has a .dat file that records long-term historical trends, but its opening speed is slow. HistorianToPDA.exe can convert this dat file into a PDA format. dat file, which can be used to quickly open trend charts for several months using PDAClient. The following figure shows the original data file for a day, with a sampling cycle of 1 second.
Figure 3.3 Long term raw data of AB company The following figure shows a data file converted to PDA format with a compression rate close to 100 times.
Figure 3.4 Conversion of AB Company's Long Term Raw Data to PDA Format The following figure shows the trend curve.
Figure 3.5 Analysis of AB Company's
Long Term Raw Data Conversion Using PDA |
|||||
|
|
|||||
|
Apparatus test&Fault diagnosis&Quality analysis |
Millisecond data sampling Real-time data compression Capture signal instantaneous mutation |
||||
|
友情链接 百度 腾讯 新浪 网易 搜狐 凤凰 淘宝 京东 中国自动化网 中国工控网 西门子 罗克韦尔 华军软件园 天空软件站 非凡软件站 多多软件站 携程 知乎 中国五矿 中冶集团 中冶赛迪 中冶南方 中冶京诚 中冶华天 中冶长天 中冶北方 中冶焦耐 中国宝武 宝信 宝钢 武钢 鞍钢 河钢 首钢 沙钢 山钢 涟钢 鄂ICP备2025092850号 版权所有©Copyright:2025-2035. 经纬铭月科技（武汉）有限公司 |
|||||
Develop communication protocol, Customized
analysis function, Open data interface, XinChuang domestic obsession
PDAServer
PDAClient