Research on the online monitor of BOD based on process neural network

For the online monitor of BOD in sewage treatment process, a improved process neural network algorithm, included function momentum adjustment item and learning rate automatic adjustment, was introduced, which was used to establish the BOD soft sending model. The main controller about ATMEGA1280 singlechip was designed, the software was programmed by using the blocking software designing method and the integrated development environment of AVR Studio. The online monitor of BOD has many functions such as data acquisition, soft sensing computing, LCD display, data storing, printing, et al. The instrument is being applied on the industrial working field, and the result indicates the average relative forecasting error is less than 4.1%.