SOE (sequence of events) for AI
Thread Content
This post was last edited by jacob451 on 2010-6-3 01:14. Hey guys, what about SOE (sequence of events) with a timestamp resolution of 1ms for 8 binary signals? SOE is generally used for DI/DO signals; but for AI signals, if timestamp support is not available (like in Zhejiang Zhongkong), how can one achieve a resolution of 10ms?
In engineering practice, generally when it comes to handling AI quantities with SOE functionality, comparison modules are often used to set upper and lower alarms, and SOE is employed for collecting these alarm signals.
Could Tiger** please explain a bit more?
The AI node can also collect SOE values; after AI conversion, it is possible to set output H/HH/L/LL alarms, which are of BOOL type, thus allowing for data collection
After all, it still ends up being SOE in the form of binary points. Analog values are generally not available; don’t analog values have historical trend records?
AI generally includes comparison modules to achieve HH/LL, etc.; the data is converted to BOOL values, which then trigger SOEs!
First, you need to set the alarm points using the upper and lower limits of the AI module; the AI input module of Zhejiang University’s DCS should be capable of performing this function. Then, introduce the AI upper and lower limit alarm points into the SOE of the SIS system. Regarding the issue of changing the time marker to 10ms, if it’s not possible to set time markers for the upper and lower limit alarm points in Zhejiang University’s AI system, then you can modify the \"input delay\" value in the DI hardware configuration screen of your SIS system – change it from 1ms to 10ms.
My personal understanding is that the main feature of SOE is its fast sampling speed; therefore, SOE devices generally come with precision specifications of 1ms or 3ms. Therefore, the functionality of SOEs is generally hardware-based; if it is software-based, such speed cannot be guaranteed. Furthermore, SOEs are timestamped, and the relevant action logic is based on sequence, which is what gives them their value. Therefore, I cannot figure out the specific meaning of AI’s SOE. Because in typical control processes, there seem to be no sensors with a measurement precision on the order of milliseconds. Additionally, the process of sending data to the controller (module) for A/D conversion after it has been transformed by a transmitter is too time-consuming for such short time intervals; therefore, I think the significance of SOE recordings in this type of AI lies more in their symbolic value. A trend chart is more straightforward to view. Additionally, the entry of HH or LL signals into SOE is achieved through software, so real-time performance cannot be guaranteed. Moreover, it seems that alarm signals also hold little significance for the analysis of SOE.
Only BOOL values can be recorded
AI is generally in the order of seconds, while DI/DO is in the order of milliseconds. Therefore, the SOE of AI has no practical meaning.
In the Zhejiang University Central Control System I’ve worked with, the analog signal acquisition doesn’t seem to be accurate to the level of 1–2 ms; the trend graphs usually have a resolution of at least 1 meter. As a result, even the use of alarm functions isn’t very useful in such cases