A basic of process control is to understand process conditions with measuring instruments and to operate processes so as to realize target conditions. If input measured values were inaccurate, output of manipulated variables would become improper and, as a result, it would be difficult to bring the process to the desired condition. In the wastewater treatment process, thanks to the latest progress in sensor technology, numerous automatic measuring instruments have been introduced. However, because of adverse environmental conditions peculiar to the wastewater treatment process such as slime-contaminated sensing elements, long-term continuous measurement is rather difficult. We believe such disadvantages in the measurement are making automatic control of the process very difficult to achieve. Under such circumstances, we have developed a detection system for unusual data which automatically checks six items of deviation from upper and lower limit values, rate of change (too much or too little), collating data from similar measuring instruments, etc. based on the measuring data of the last 30 days. With this system, validity of the accumulated data is being checked using measuring data. Accordingly, it enables us to deal with characteristics of measuring instruments, situations of wastewater treatment plants, seasonal changes, etc. automatically. In this report, automatic methods to establish judgement criteria, structure of this detection system and logic of detection of unusual data are introduced. Furthermore, test results with the data collected from actual wastewater treatment plants are covered.

This content is only available as a PDF.