Abstract:
Demand-side integration (DSI) refers to the overall technical area focused on advancing the efficient and effective use of electricity to support power systems and its customer needs. As smart grid (SG) development continues to transform the passive distribution network to the active network with increasing distributed energy resources (DER), the importance of the electricity distributors also shifts from being the network asset owner to more of the network operator role. To integrate the demand-side of electricity market having diverse DER ownership and varying customer needs and expectations, it is not enough just to rely on nodal price signalling that is based on the wholesale electricity market considering the transmission grid constraints. Instead customer choice, service quality differentiation, fashionable lifestyle choices, innovative technologies and new media methods are important factors of DSI price signalling on the retail market side, alongside the technical features of DERs and the distribution network regulatory policies. This thesis recognises the role of wholesale nodal prices as a proxy of real-time energy cost, but focuses more on the derivation and modification of the distribution network usage cost considering various DERs. The DSI approaches that have been investigated focus on the three aspects: distributor pricing structure, service quality differentiation and customer comfort with distribution network constraints. To investigate each aspect, the standardised distribution test networks have been modelled, on which the optimisation formulations are solved by the GAMS software and the network parameters are derived from the open-source distribution analysis software of GridLAB-D or OpenDSS. The software MATLAB or Excel VBA has been used to automate the software execution and data transfer. Firstly to refine the distributor price signalling, the controllable loads and photovoltaic (PV) units have been factored in to the distribution network planning based on the long-run average incremental costs (LRAIC). The avoided or deferred investment costs of reinforcing the distribution network have been used to determine the amount of discount or rebate payable to the DER owner. For the case of controllable loads, the rebates of offering the loads as controllable become less when more controllable loads are enabled. For the case of PV integration, PV growth has been planned similarly to the peak load growth. When PV penetration level is high, some of the network reinforcement costs need to be attributed to the PV owners. On the other hand, the benefit of PV support infrastructure, such as electricity storage systems (ESS), can be separately recognised. Secondly, the delivery service quality differentiation is made possible with the advance of SG technologies, such as using controllable loads, distribution automation or ESS. However, it is best to conduct the network planning and operation according to the individual customer needs and aspiration of a highly reliable network. The proposed reliability premium (RP) based on load point reliability indices is the payment adjustment to the normal distribution network tariff, which also reflects the customer reliability preference at a more granular level. The logic behind this is that customers who pay for more RP are expected to receive better network supply quality and higher compensation; while the extra reliability built in to the network by the distributor can also be `traded' to the wholesale market using automated load control if not desired by the `free-rider' customers. Thirdly, to facilitate the role of distributor as the network operator, a DER scheduling approach has been proposed to use the `time value' of deferrable loads in the objective formulation instead of the typically used monetary terms. The DER scheduling approach recognises the operational benefit of deferrable loads in the system and also the importance of safeguarding customer comfort. The network asset has shown to be fully utilised by operating all the network voltages close to the lower voltage constraints. The scheduling intervals are adjusted to be more granular when closer to the present time but more coarse in the distant future, so that higher time resolution and longer scheduling horizon can both be achieved. The computation performance has been tested for a practically large distribution test network with numerous DER decisions, in which scheduling results have also indicated the possibility of quantifying the individual DER contribution during the final reconciliation process. Overall, the thesis has led to the discussions on an emerging topic of transactive energy, which is the first of its kind addressing the aspects of New Zealand Electricity Market, power systems and regulatory environment altogether.