Data-driven strategies for flexible energy demand in business

Supervisors

Dr. Ibrahim Ibrahim, Dr. Hongda Tian and Prof. Stuart White (UTS)

Patrick Lay (Retragreen)

The project

There is now more than 20 GW of rooftop solar photovoltaic (PV) capacity across Australia, with 3 GW being added each year. While this rapidly increasing penetration of rooftop solar can reduce customers’ energy bills and carbon emissions, it creates significant challenges for maintaining grid stability. PV output can drive minimum demand so low that some synchronous generators must turn off, leaving insufficient inertia for a stable electricity system. One approach to managing this issue is to curtail the output of solar and wind generators; however, this wastes energy, degrades customer trust, and can significantly delay uptake of renewable energy, diminishing the associated emissions and cost benefits. And although batteries allow excess solar energy to be stored and used at a later time, the capital cost of batteries is often prohibitive.

An alternative to using curtailment or batteries to solve the mismatch between supply and demand is to incentivise customers to use more energy at times of high solar PV generation through activating flexible demand (FD). The C&I sector offers significant opportunities for FD, particularly through shifting of heating, cooling, ventilation, air conditioning and refrigeration (HVAC-R) loads. However, these opportunities must be balanced against other operational constraints such as maintaining productivity and product quality. Finding the most cost-effective strategy for managing electricity demand can therefore be a challenging problem.

Artificial Intelligence (AI) is being increasingly applied to manage flexible loads. With access to suitable data, AI algorithms can be trained to predict electricity prices and other variable incentives on the supply side, as well as opportunities for demand side flexibility within the operations of business customers. Hence there is an opportunity to use AI to both optimise business operations while increasing flexible demand capacity. By employing AI, both energy providers and business consumers can optimise energy use, reduce costs, and enhance the reliability and stability of power systems.

The project will be useful for four main groups of stakeholders.

1. Established C&I businesses—Through a continuous feedback loop of knowledge sharing, we will disseminate project findings and best practices in terms of FD potential, which will benefit participating businesses through lower electricity bills, lower emissions and better alignment with the UN Sustainable Development Goals.
2. Start-up businesses—The project provides an opportunity for new innovative businesses, such as Retragreen, to extend their presence in the market, to assist in opening up opportunities for more flexible demand capacity while improving their customers’ productivity, and ultimately overcoming the technical challenges of widespread renewable energy integration.
3. The electricity sector—The project findings are expected to be used by electricity networks, retailers and service providers to help identify, implement and better understand the impacts of FD opportunities, particularly in food sector businesses. By demonstrating tariff and other incentive models for unlocking FD capacity, transmission and distribution businesses will be better equipped to shift demand to help alleviate the so-called ‘duck curve’, enabling better and more consistent utilisation of existing electricity infrastructure to reduce costs. It will also contribute to better managing voltage and underfrequency load-shedding.
4. Policymakers—The project outputs will assist policymakers in identifying barriers to implementing FD strategies, leading to new policies that address these barriers and facilitate a faster, lower cost and more reliable clean energy transition.