Skip to content

Home

Jump to bottom
Harsha Kumara edited this page Aug 14, 2023 · 8 revisions

LEAP (La Trobe Energy AI Platform)

Project Lead: Daswin De Silva ([email protected])

Technical Architect: Nishan Mills ([email protected])

Context

  • A recent study unveiled that 71% of Australians believe we should be a world leader in finding solutions to climate change and 83% of Australians support a phase-out of coalfired power stations. (Quicke, A., & Bennett, E. Climate of the nation 2020.)
  • As a large tertiary education institute, La Trobe is heavily dependent on the State electricity grid. In Victoria, brown coal is the primary source of energy, accounting for about 80% of electricity generation. It is also one of the largest contributors to domestic greenhouse gas emissions.
  • La Trobe University’s Net Zero Program aims to reduce the university carbon footprint to net zero emissions by 2029.
  • This will be achieved by decreasing our dependence on the state power grid (which is 80% brown coal based), increasing the use of renewable energy generated by rooftop PV, investing in battery storage, and optimising the overall energy usage across the university campuses.

Project

  • LEAP, the La Trobe Energy AI Platform, is the flagship Artificial Intelligence (AI) initiative of La Trobe’s Net Zero Program.
  • LEAP is being designed, developed and deployed as a “Living AI Lab” project, by the Centre for Data Analytics and Cognition (CDAC), La Trobe’s leading research centre for fundamental and applied AI research.
  • The CDAC team building LEAP consists of four full-time academics, ten PhD students and two Master's by research students.
  • The CDAC Ethics committee provides complete ethical, governance and compliance oversight and advice at each stage of LEAP development and deployment.
  • LEAP is also leveraged in La Trobe’s Master of Business Analytics and Bachelor of Business Analytics degrees, by approximately 150 students to, 1) gain an awareness on the net zero program and the role of technology for sustainability, 2) learn time-series forecasting algorithms and predictive analytics techniques in a real-world, multi-campus setting, 3) develop AI libraries and analytics dashboards using cutting-edge technologies.
  • LEAP has been an enabler of international research collaborations with Lulea University of Technology, Sweden and Aalto University, Finland. With Lulea, CDAC is working on spatiotemporal AI for energy market price dynamics, with Aalto, CDAC is working on adaptive multi-energy virtual power plants for complex of buildings.

Technology

  • LEAP can be described by its key technological innovations, 1) all-data lakehouse, 2) data quality framework, 3) AI and automation engine, 4) multi-stakeholder analytics dashboards.
  • The all-data lakehouse generates a single-version-of-truth, centralised repository of internal data streams on consumption, generation, utilisation, events, and external data streams of weather, climate, tariff, and energy market metrics.
  • The data quality framework validates all data feeds to standards and formats persisted by the AI and automation engine. It also uses AI for predictive coding during dropouts, missing and erroneous sensors reads, so that the all-data lakehouse is inclusive and complete.
  • The AI and automation engine consists of multiple intelligent and automated functions. To name a few, baseline prediction, short-term and long-term consumption and generation forecasting, anomaly detection using intra-day and inter-day consumption signatures, tariff simulation, metering and bill validation.
  • The analytics dashboards visualise energy metrics, forecasts, fluctuations, trends and outliers that can be sliced, diced, rolled up, drilled down by dimensions of interest by multiple types of stakeholders, such as users, operators, managers, providers and advocacy groups.
  • LEAP utilises full-stack development, staging and production environments, composed of cutting-edge technologies such as, Git, Jenkins, Terraform, logic apps, Synapse, Docker containers, React.js, Tensorflow and .NET services and Python.

Community

  • As an intelligent, automated platform for sustainable, future-ready facilities management, LEAP is shared with the local community, such as schools and community centres.
  • It is also a change agent that brings together local schools and residents to public events of awareness such as hackathons and seminars.
  • Results, feedback and lessons learned from delivery across the local community are then dispersed to the public for further emissions reduction and gains in operational efficiency.

Awards

  • LEAP was awarded the 2021 Clever Campus Innovation Award by the Tertiary Education Facilities Management Association (TEFMA)

Publications

  1. Sumanasena, V., Gunasekara, L., Kahawala, S., Mills, N., De Silva, D., Jalili, M., Sierla, S. and Jennings, A., 2023. Artificial Intelligence for Electric Vehicle Infrastructure: Demand Profiling, Data Augmentation, Demand Forecasting, Demand Explainability and Charge Optimisation. Energies, 16(5), p.2245.

  2. Moraliyage, H., Dahanayake, S., De Silva, D., Mills, N., Rathnayaka, P., Nguyen, S., Alahakoon, D. and Jennings, A., 2022. A Robust Artificial Intelligence Approach with Explainability for Measurement and Verification of Energy Efficient Infrastructure for Net Zero Carbon Emissions. Sensors, 22(23), p.9503.

  3. Mills, N., Rathnayaka, P., Moraliyage, H., De Silva, D. and Jennings, A., 2022, July. Cloud Edge Architecture Leveraging Artificial Intelligence and Analytics for Microgrid Energy Optimisation and Net Zero Carbon Emissions. In 2022 15th International Conference on Human System Interaction (HSI) (pp. 1-7). IEEE.

  4. Moraliyage, H., Mills, N., Rathnayake, P., De Silva, D. and Jennings, A., 2022, July. UNICON: An Open Dataset of Electricity, Gas and Water Consumption in a Large Multi-Campus University Setting. In 2022 15th International Conference on Human System Interaction (HSI) (pp. 1-8). IEEE.

  5. Wimalaratne, S., Haputhanthri, D., Kahawala, S., Gamage, G., Alahakoon, D. and Jennings, A., 2022, July. UNISOLAR: An Open Dataset of Photovoltaic Solar Energy Generation in a Large Multi-Campus University Setting. In 2022 15th International Conference on Human System Interaction (HSI) (pp. 1-5). IEEE.

  6. Rathnayaka, P., Moraliyage, H., Mills, N., De Silva, D. and Jennings, A., 2022, July. Specialist vs Generalist: A Transformer Architecture for Global Forecasting Energy Time Series. In 2022 15th International Conference on Human System Interaction (HSI) (pp. 1-5). IEEE.

  7. Haputhanthri, D., De Silva, D., Sierla, S., Alahakoon, D., Nawaratne, R., Jennings, A. and Vyatkin, V., 2021. Solar irradiance nowcasting for virtual power plants using multimodal long short-term memory networks. Frontiers in Energy Research, 9, p.722212.

  8. Kahawala, S., Haputhanthri, D., Moraliyage, H., Wimalaratne, S., Alahakoon, D. and Jennings, A., 2022, July. Comparative Evaluation of Gradient Boosting with Active Thresholding and Model Explainability for Peak Demand Forecasting. In 2022 15th International Conference on Human System Interaction (HSI) (pp. 1-6). IEEE.

  9. Gamage, G., Mills, N., Rathnayaka, P., Jennings, A. and Alahakoon, D., 2022, July. Cooee: An Artificial Intelligence Chatbot for Complex Energy Environments. In 2022 15th International Conference on Human System Interaction (HSI) (pp. 1-5). IEEE.

  10. Kahawala, S., De Silva, D., Sierla, S., Alahakoon, D., Nawaratne, R., Osipov, E., Jennings, A. and Vyatkin, V., 2021. Robust multi-step predictor for electricity markets with real-time pricing. Energies, 14(14), p.4378.

Clone this wiki locally