A5032891017- Water Quality Monitoring Technologies
- Fault Detection and Control Systems
- Hydrological Forecasting Using AI
- Neural Networks and Applications
- Domain Adaptation and Few-Shot Learning
- Water Systems and Optimization
- Data Stream Mining Techniques
- Time Series Analysis and Forecasting
- Fuel Cells and Related Materials
- Water Quality Monitoring and Analysis
- Microfluidic and Capillary Electrophoresis Applications
- Machine Learning and Data Classification
The University of Western Australia
2022-2024
Wastewater treatment plants are complex, non-linear, engineered systems of physical, biological and chemical processes operating at different timescales. Sensor used to monitor wastewater in order ensure public safety for efficient management the plants. However, parameters interest can require expensive or inaccurate sensors may off-site laboratory analysis. For example, ammonium is important as a prime indicator efficiency highly regulated discharge water. But also over <inline-formula...
Deep learning is being widely utilized in industrial process monitoring, control, and optimization. However, the wastewater industry, its applications are still underexplored. This because deep requires a large amount of labeled training data to induce effective predictive models. Owing high cost sensors frequency delay sampling laboratory analytics, treatment can be sparse with varying frequencies. One option address limitations use transfer learning. owing covariate shift between commonly...
Accurate deep predictive models of wastewater processing plants are important to ensure operational parameters safe and sustainable. Training such requires large volumes data that is hard find in the domain. Transfer learning addresses problem, by training using from an adopted domain, fine-tuning it on target However, due significant distributional shift between commonly source domains for transfer domain processes, rarely performs at acceptable level. This paper proposes a method generate...
Complex urban systems can be difficult to monitor, diagnose and manage because the complete states of such are only partially observable with sensors. State estimation techniques used determine underlying dynamic behavior complex their highly non-linear processes external time-variant influences. States estimated by clustering observed sensor readings. However, performance degrades as number sensors readings (i.e. feature dimension) increases. To address this problem, we propose a framework...