Micro-grid design and dispatch co-optimisation considering uncertainties and demand response – Cases in New Zealand

One in eight people around the world, approximately one billion people, lack access to reliable electricity. Also, a great majority of people with access to electricity are experiencing some form of energy hardship – around a third (29%) of New Zealand households struggle to afford their electricity bills, spend a major part of their income on power, or often feel cold in winter. In this light, the ever-falling costs and continued efficiency improvements of renewable energy technologies are facilitating the ‘clean energy for all’ initiatives globally. Whilst considerable effort has been devoted to a range of interventions to address the underlying technological, institutional, and regulatory barriers, less attention has been given to address the glaring technical knowledge gaps in quantitative energy planning research, in terms of investment planning and capacity optimisation modelling; for the design of renewable energy systems, and specifically micro-grid systems. In response, this thesis addresses four notable gaps in the literature, namely: (i) the underrepresented usage of state-of-the-art meta-heuristic optimisation algorithms to determine the configurations of components, (ii) the lack of application of game-theoretic frameworks to the study of aggregator-mediated demand-side flexibility procurement, (iii) the limited number of approaches that quantify multiple parametric uncertainties simultaneously, and (iv) the narrow focus on joint micro-grid investment planning and energy scheduling optimisation.

To this end, the thesis introduces a novel strategic, meta-heuristic-based, demand response-integrated, uncertainty-aware, long-term micro-grid energy planning and capacity optimisation model, featuring the following key novel generalisations, each addressing one of the above-mentioned gaps: (i) utilising a state-of-the-art meta-heuristic optimisation algorithm, moth-flame optimiser, which is found to have superior performance to a wide variety of well-established and state-of-the-art meta-heuristics in minimising micro-grid life-cycle costs, (ii) characterising the utility-aggregator-customer interactions in interruptible load programmes using non-cooperative game theory in an equitable, market-based approach, (iii) expanding the number of model-inherent parametric uncertainties quantified concurrently without excessive computational demands, and (iv) integrating a dynamic, forward-looking scheduling design framework for the co-optimisation of investment and operational planning costs.

To demonstrate the effectiveness of the model in yielding the cost-minimal mix of candidate renewable energy technologies considered for integration into a micro-grid system, the model was applied to four previously unexplored test cases. Four on- and off-grid 100%-renewable and -reliable micro-grid systems were specifically conceptualised for the following cases in New Zealand: (i) the community of 400 permanent inhabitants on Stewart Island, (ii) a rural community of about 350 people near Feilding, (iii) the eight-lot Totarabank Subdivision located in the Wairarapa District, and (iv) a 1,000-strong community in Ohakune that swells to 8,000 people during skiing season. Crucially, the case studies, undertaken on different scales and with different degrees of topological complexity, provide a robust evidence base to support the main research proposition that not only is it technically feasible to implement the smart, integrated renewable energy systems optimised by the proposed model, but they also surpass unsubsidised retail parity.

In particular, the thesis demonstrates that using the moth-flame optimisation algorithm, capturing the real flexibility potential of small- to medium-scale end-users, characterising multiple sources of data uncertainty, and adopting look-ahead, predictive dispatch strategies during the investment planning phases of stand-alone and grid-connected micro-grid systems, can pave the way toward achieving greater energy independence, -democracy, -resilience, and -security in rural and semi-urban areas in a cost-effective and environmentally efficient way. Most of all, the developed model provides in-depth, accurate, and robust strategic infrastructure planning decision-making support by adopting a holistic and comprehensive approach to energy planning optimisation. The approach enables a high-level, realistic analysis of the financial implications of the clean energy transition, especially in community-scale installations, necessary to cost-effectively promote private sector investment in the green economy – in the efforts to advance global electrification and economy-wide deep decarbonisation.