Fuzzy patterns allow better process control

Processes that take a long time to complete can be controlled more efficiently by applying fuzzy logic in a new way. Tom Shelley reports A strategy to improve the control of waste water treatment aerators uses fuzzy pattern classification - a very advanced technique developed to attempt the recognition of faces and the likely behaviour of groups of terrorists. Whether it will ever be successful in the latter two applications remains to be seen, but applied to aeration in waste water treatment plants, it is able to reduce operating costs, lower energy consumption, increase capacity and drastically improve nitrogen elimination. The object of the control system is to supply sufficient oxygen to allow micro organisms to convert dissolved ammonia to nitrates and to further convert the nitrates to gaseous nitrogen gas which can return to the atmosphere. The approach embodied in the TAT 2000 system, developed by the German company Envicon, is to make measurements of how the Redox potential - a measure of the amount of dissolved oxygen - changes with time. During the learning process, the system learns how the dissolved oxygen level increases as aeration is applied, as a function of time and aeration intensity under different conditions. Owing to the variable nature of the feedstock, the curve shape remains approximately the same, but it may be stretched out or compressed sideways according to the load on the system imposed by a particular batch. The system is therefore set up to generate an envelope of standardised reference performance curves. As the process completes its cycles, the standardised curve of the signal measured in the treatment tank comes to more and more resemble the learned characteristic pattern. When the measured, standardised real world signal maximally matches the learned, fuzzy, characteristic pattern the system engages, and is able to choose the optimum point to turn off the blowers. The company claims that the system lowers energy consumption for aeration by up to 30%, increases nitrogen elimination by up to 60%, and improves biological elimination of phosphorus, reducing the requirement for precipitating chemicals. It is also said to increase capacity by up to 15% and even out low level effluent values to a very high degree. The control strategy is applicable to any slow responding process with a varying input feedstock. The UK agent for Envicon is Naston, based in Weybridge. Envicon Naston Pointers * system lowers energy consumption for aeration by up to 30%, increases nitrogen elimination by up to 60%, and improves biological elimination of phosphorus, reducing the requirement for precipitating chemicals * Plant capacity is said to be increased by up to 15% * The control strategy is applicable to any slow responding process with a varying input feedstock