Keywords: activated sludge process, artificial intelligence, knowledge base, process control, database, rules, user interface.

Knowledge base systems are defined as computer programs that make use of logical relationships to incorporate knowledge and expertise about a specific problem area to perform specialised tasks that typically require human judgement. The activity of developing intelligent knowledge-base system employs proven mathematical principles, emphirical results, heuristic and pragmatic programming techniques. This work attempts to incorporate process schematics of knowledge base and integrating data management into the computer program. This paper's objective is to present the knowledge-base system possessing broad sector of the knowledge in respect of the Innovative Activated Sludge Process (IASP, an integrated version of the activated sludge process) to keep the system always in stabilised conditions during unsteady state conditions. IASP aims to simplify the process, and operational aspects of the latter. IASP relies on the unique geometrical shape of the unit to transfer the sludge to the aeration zone because the sludge recycle pumps are absent.

The architecture of knowledge-base system has a database, inference engine, process models, and a user interface. The knowledge base comes from the two categories: from the operator's experience, and from process simulation. The interactive database is also part of the knowledge base. The knowledge base is represented in the form of facts and rules that are structured for the purpose of diagnosis. The facts are the information and the rules manipulate the related facts. The rules are straightforward and designed to imitate the judgement of an expert when he solves a problem (heuristic knowledge). The knowledge base developed for IASP was classified into the categories pertaining to the problems associated with the aeration, settling zones and effluent quality. Based on the structure of the knowledge base about 28 rules have been framed for inclusion in the knowledge base. The Disolved Oxygen (DO) control and the waste sludge flow control are the only two parameters that were amenable for process control in the IASP because of the absence of the sludge recycle pumps. Food to Microorganism (F:M) ratio and Mean Cell Residence Time (MCRT) are considered as control variables and are computed from the values of influent 5-Day Biochemical Oxygen Demand (BOD₅), influent flow, Mixed Liquor Volatile Suspended Solids (MLVSS), volume of the reactor, waste sludge flow. The Java Expert Systems Shell (JESS) package was used for knowledge base development.

The integrated knowledge-base system was tested and found to be in good agreement with the human experts in most cases for a typical problem of the IASP loss of carbon-oxidation. A set of problems was selected from the pilot plant daily records for the validation of the developed knowledge-base system. The knowledge-base operator was asked to give an answer to each problem case. The suggestions obtained from the knowledge-base system were in very close agreement with the structure developed. In all the cases the field results were in close agreement with the IASP-KB findings. When the effluent BOD₅ is found to be higher than the normal, low MLVSS was determined to be the problem. The IASP-KB then begins to diagnose the process to find the causes of failure. If the MLVSS calculated is normal, knowledge base finds that the process problem was due to excessive influent substrate load. Diversion of part of the influent, decreasing the waste sludge flow rates were the corrective measures that the IASP-KB suggests quantitatively using simulation model. The model information is used along with the heuristic knowledge in parallel and the most appropriate one is suggested by the computer system for application.

The knowledge base system presented for the improvement of the operation of the pilot plant provides both qualitative and quantitative control advises to the user. As stated in the IASP-KB instead of IF-THEN statements normally adopted the provision has been made for entry of both qualitative and quantitative data directly at the interface point at one time, which is the deviation from other research works on this technique With IASP-KB process efficiency and reliability is enhanced substantially. It allows the direct entry of process data into spreadsheets. It is a powerful tool for improvement of operation and control of wastewater treatment systems. The structure of the knowledge base developed help specially when influent loading perturbations are going on and for the mechanisms, which do not have models to describe them. Process efficiency and stability can be enhanced with suitable integration of the knowledge base system with the process model. The case studies concluded that the process tool developed is an effective one for controlling the innovative activated sludge process to function at under unsteady state conditions.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)