Knowledge-based control and case-based diagnosis based upon empirical knowledge and fuzzy logic for the SBR plant

Area of Science:

• Environmental Engineering
• Process Control
• Artificial Intelligence in Water Treatment

Background:

• Biological wastewater treatment plants (WWTPs) face challenges due to time delays and disturbances, necessitating manual control by skilled operators.
• Current diagnostic methods rely on operators' past experiences, limiting systematic management and automation.
• Effective plant management requires automation based on established operating procedures.

Purpose of the Study:

• To develop an automated control and diagnosis system for WWTPs based on operator knowledge.
• To implement a fuzzy logic-based controller for managing influent and external carbon.
• To enable case-based diagnosis by analyzing historical operational data.

Main Methods:

• Fuzzy logic was employed to convert linguistic operator knowledge into actionable control rules.
• The controller utilizes dissolved oxygen (DO) lag-time as a proxy for loading rate, reducing the need for expensive sensors.
• Case-based reasoning was used for diagnosis, comparing new operational profiles with historical normal and abnormal patterns.

Main Results:

• The proposed fuzzy logic controller effectively manages influent and external carbon, considering the loading rate via DO lag-time.
• The system successfully replaces expensive sensors with a more cost-effective DO sensor for critical measurements.
• The case-based diagnosis effectively identifies normal and abnormal plant operating conditions by pattern matching.

Conclusions:

• The developed system offers effective and stable operation for WWTPs through automated control and diagnosis.
• The approach enhances operational efficiency and prevents issues like over-feeding.
• This knowledge-based system provides a robust framework for modernizing WWTP management.