Water. Desalination + reuse
Issue link: https://fhpublishing.uberflip.com/i/604344
TECHNOLOGY Adiquimica has developed software designed to diagnose and recommend measures to deal with membrane fouling problems. At its core is a knowledge- based system that includes an expert system and case-based reasoning. The expert system (ES) emulates the reasoning process of leading practitioners in membrane cleaning. The main modules of the ES are the knowledge base (KB) and the inference engine. The KB caries all the membrane cleaning process knowledge. The inference engine controls the reasoning taken by the ES to identify the cause or causes of membrane fouling and in recommending the best cleaning procedure. The heart of the ES is the knowledge included in the KB that is obtained from literature, recorded symptoms, behaviour of the normalized operational parameters, interviews with experts, Membrane Membrane 2016 CONFERENCE & EXPOSITION Membrane CONFERENCE & EXPOSITION Technology February 1–5, 2016 Henry B. Gonzalez Convention Center San Antonio, Texas Register by December 31 for best rates! The Membrane Technology Conference & Exposition is the premier showcase for the latest developments in water, wastewater, reuse, and industrial membrane systems and practices. Presented by America's Authority in Membrane Treatment Improving America's Waters Through Membrane Treatment and Desalting Dedicated to the World's Most Important Resource ® Check out these highlighted sessions: • Texas Perspectives: Desal to DPR • Carlsbad Seawater Desalination Project: Road to Operation • Featured Topics Given in Spanish • One Water, Many Uses: The Present and Future of Water Reuse • MF/UF Systems: Experience Leads to Improvement • The Knowledge of Hindsight: Plant Design, Start-up, and Operation • Improving MBRs: Operation and Practice www.awwa.org/amta/Membrane2016 | 1.772.463.0820 November-December 2015 | Desalination & Water Reuse | 31 | membrane autopsies, laboratory cleaning tests, and the effectiveness of cleaning procedures applied in full-scale plants. Environmental and economic criteria are also taken into consideration to build the ES. All symptoms, facts, reasoning strategies, procedures, relationships and action strategies used to diagnose and solve the membrane fouling problem are represented in a series of decisions that take the reasoning along a series of branches. This is known as a decision tree. The trees include diagnosis, identification of cause of fouling and strategies for troubleshooting, avoiding contradictions and redundancies. Once the ES has concluded the situation, it recommends a specific action strategy to clean the membrane. ES includes not only knowledge but also qualitative information processing to emulate the expert reasoning and to represent knowledge in an easily understandable form. But it has limitations, one of which is the static nature of the KB which, once developed, cannot learn from new experiences. A second knowledge-based tool, known as case-based reasoning system (CBR), was introduced to add continuous learning from new experiences to the Adiclean ES reasoning. The Adiclean CBR identifies similarities between the fouling condition under examination and those of recorded previous cases. It then draws on the measures taken in cleaning in the previous case. So each time the Adiclean system identifies a membrane fouling problem it proposes the specific causes, and suggests troubleshooting measures for the operator to consider. Its evaluation of the problem and cleaning results are added to the case-base library. Should the current case differ from all stored cases it will be added as a new experience. Reasoned response