I&T Wish - Automated Fault Detection and Diagnostics System for Seawater Pump 2024-07-25
Automated Fault Detection and Diagnostics System for Seawater Pump
I&T Wish
Automated Fault Detection and Diagnostics System for Seawater Pump (REF: W-0540)
Matched I&T Solution
Trial Project
Summary and Challenges
Maintenance of seawater pumps is crucial in MVAC systems because seawater pumps are in direct contact with seawater, leading to more severe aging and wear compared to other pumps. Seawater pump rooms are often located deep underground, and damage can cause serious consequences such as flooding of the plant room. Current monitoring systems can only provide incident alarms, typically after problems have existed for some time.
Additionally, existing automatic fault detection systems cannot accurately locate faulty components or quantify the severity and urgency of faults. Therefore, the project aims to develop a high-precision automatic fault detection system using artificial intelligence technology combined with various types of sensors, including vibration, temperature, and other sensors.
Expected Outcome
Accurate Prediction of Component Conditions and Data-Driven Decision Support: Utilize multi-source data (sensor data, operational history, maintenance records) and AI models to accurately predict seawater pump component conditions, including wear, aging, and potential failures. Utilizing physics-guided machine learning, or multimodal and attention mechanism-based technique to enhance accuracy. Provide suggestions on maintenance and replacement plans.
Quantification of Fault Severity and Urgency: Classify faults to quantify their severity and urgency. Use anomaly detection algorithms to dynamically assess fault risks based on historical and real-time data, helping maintenance personnel develop prioritized maintenance plans.
Early Warning Mechanism: Develop an early warning system based on time series analysis and prediction models to issue alerts as soon as problems begin, preventing deterioration. Achieve real-time monitoring and anomaly detection through automated system data flow monitoring, promptly identifying and reporting abnormalities.
Easy Installation and No Impact on System Operation: Design sensors and monitoring equipment that are easy to install, ensuring a simple process without affecting seawater pump operation. Develop a wireless sensor network to reduce wiring needs, enhancing flexibility and scalability.
Rapid Deployment: Use transfer learning and few-shot learning techniques to reduce the need for extensive fault data, enabling quick system deployment to target seawater pumps. Develop self-adjusting models optimized for different pumps' characteristics and environments, improving model generalization.
Visualization and Report Generation: Develop a visualization dashboard for intuitive data presentation and analysis results, helping maintenance personnel quickly understand and assess system conditions. Automatically generate maintenance reports.
Expected Trial Duration
12-month
Contact Information
I&T Wish Proposer
:
Electrical and Mechanical Services Department (EMSD)
