Suppliers in competitive electricity markets regularly respond to prices that change hour by hour or even more frequently, but most consumers respond to price changes on a very different time scale, i. e. they observe and respond to changes in price as reflected on their monthly bills. This thesis examines mixed complementarity programming models of equilibrium that can bridge the speed of response gap between suppliers and consumers, yet adhere to the principle of marginal cost pricing of electricity. It develops a computable equilibrium model to estimate the time-of-use (TOU) prices that can be used in retail electricity markets. An optimization model for the supply side of the electricity market, combined with a price-responsive geometric distributed lagged demand function, computes the TOU prices that satisfy the equilibrium conditions. Monthly load duration curves are approximated and discretized in the context of the supplier's optimization model. The models are formulated and solved by the mixed complementarity problem approach. It is intended that the models will be useful (a) in the regular exercise of setting consumer prices (i. e. , TOU prices that reflect the marginal cost of electricity) by a regulatory body (e. g. , Ontario Energy Board) for jurisdictions (e. g. , Ontario) where consumers' prices are regulated, but suppliers offer into a competitive market, (b) for forecasting in markets without price regulation, but where consumers pay a weighted average of wholesale price, (c) in evaluation of the policies regarding time-of-use pricing compared to the single pricing, and (d) in assessment of the welfare changes due to the implementation of TOU prices.

Suppliers in competitive electricity markets regularly respond to prices that change hour by hour or even more frequently, but most consumers respond to price changes on a very different time scale, i. e. they observe and respond to changes in price as reflected on their monthly bills. This thesis examines mixed complementarity programming models of equilibrium that can bridge the speed of response gap between suppliers and consumers, yet adhere to the principle of marginal cost pricing of electricity. It develops a computable equilibrium model to estimate the time-of-use (TOU) prices that can be used in retail electricity markets. An optimization model for the supply side of the electricity market, combined with a price-responsive geometric distributed lagged demand function, computes the TOU prices that satisfy the equilibrium conditions. Monthly load duration curves are approximated and discretized in the context of the supplier's optimization model. The models are formulated and solved by the mixed complementarity problem approach. It is intended that the models will be useful (a) in the regular exercise of setting consumer prices (i. e. , TOU prices that reflect the marginal cost of electricity) by a regulatory body (e. g. , Ontario Energy Board) for jurisdictions (e. g. , Ontario) where consumers' prices are regulated, but suppliers offer into a competitive market, (b) for forecasting in markets without price regulation, but where consumers pay a weighted average of wholesale price, (c) in evaluation of the policies regarding time-of-use pricing compared to the single pricing, and (d) in assessment of the welfare changes due to the implementation of TOU prices.

This thesis proposes equilibrium models to link the wholesale and retail electricity markets which allow for reconciliation of the differing time scales of responses of producers (e.g., hourly) and consumers (e.g., monthly) to changing prices. Electricity market equilibrium models with time of use (TOU) pricing scheme are formulated as large-scale variational inequality (VI) problems, a unified and concise approach for modeling the equilibrium. The demand response is dynamic in these models through a dependence on the lagged demand. Different market structures are examined within this context. With an illustrative example, the welfare gains/losses are analyzed after an implementation of TOU pricing scheme over the single pricing scheme. An approximation of the welfare change for this analysis is also presented. Moreover, break-up of a large supplier into smaller parts is investigated. For the illustrative examples presented in the dissertation, overall welfare gains for consumers and lower prices closer to the levels of perfect competition can be realized when the retail pricing scheme is changed from single pricing to TOU pricing. These models can be useful policy tools for regulatory bodies i) to forecast future retail prices (TOU or single prices), ii) to examine the market power exerted by suppliers and iii) to measure welfare gains/losses with different retail pricing schemes (e.g., single versus TOU pricing). With the inclusion of linearized DC network constraints into these models, the problem size grows considerably. Dantzig-Wolfe (DW) decomposition algorithm for VI problems is used to alleviate the computational burden and it also facilitates model management and maintenance. Modification of the DW decomposition algorithm and approximation of the DW master problem significantly improve the computational effort required to find the equilibrium. These algorithms are applied to a two-region energy model for Canada and a realistic Ontario electricity test system. In addition to empirical analysis, theoretical results for the convergence properties of the master problem approximation are presented for DW decomposition of VI problems.

In this dissertation, the transformer management system has been developed to monitor transformer over loading and generate warning message in conduit mapping management system (CMMS) of Taipower company. The transformer over loading prediction is performed by both offline and online modes. Performs the transformer loading estimation by using the customer monthly energy consumption in customer information system (CIS) and the connectivity attributes of transformer and customers served in CMMS system of Taipower company. The daily load curve of distribution transformer is derived considering the typical daily load patterns which have been developed in load survey study. The warning message will be generated when the peak loading estimated is lager then the transformer rated capacity. To enhance the accuracy of transformer attributes in CMMS system, the transformer phasing measurement system (TPMS) and the connectivity identification system to identify all of the customers served by each transformer are developed. It is difficult to receive the 1 pulse per second signal form global positioning system for timing synchronization of TPMS measuring units for phasing measurement of transformers located in basement, the temperature compensated crystal oscillation with Fuzzy calibration algorithm is used to maintain the timing synchronization within 10o deviation for measurement period of 2 hours. To solve the incorrect problem of transformer and customer connectivity in CMMS, the power line carrier technology is applied in the design of connectivity measurement system for the identification of customers served by the transformer. The peak loading of transformer is estimated by including the temperature effect and the overloading flag of transformer is displayed on the CMMS automatic mapping system. For the online TLM system, the embedded transformer terminal unit is developed for the real time measurement of transformer loading and insulation oil temperature. For the transformer with abnormal operation condition, the alarm signals will be generated and transmitted to the TLM master station via hybrid communication system for the activation of demand response function to execute the load shedding control of customer loads.

The fact that household customers are central in the discussion of future sustainable energy systems compels the Swedish electricity retail companies to provide strategies in order to successfully follow the trends on the electricity market. The purpose of this thesis is to complement the electricity retail companies’ understanding of how they are able to enter a sustainable and close business relationship with these customers. The purpose is fulfilled by the analysis of how relationship concepts impact the development of demand response specifically. The information has been collected through a case study of Smart Customer Gotland, via interviews with people who have great experience from the field and via a survey directed towards the customers. The results indicate that there is a major difference in both of the actors’ visions regarding the relationship. The company desires to enter a position where less support and high customization is available. Customers instead want more personal support because of their lack of knowledge and uncertainty of new systems. To succeed, the companies have to consider the fundamental influencing incentives, economy and comfort, while also maintaining the customers’ trust. The most essential parts of the context specific relationship exchange are product exchange, information exchange, and social exchange. These should be directed towards maintaining and increasing the trust from customers. Focusing the resources earlier used for marketing, on these exchanges to make the current customers more satisfied will open up for using word-of mouth primarily from early adopters. Continued work with these early adopters will also increase companies’ knowledge, which is important for the companies to strategically develop its business towards the market of smart energy solutions.

This thesis comprises three papers examining the impact of competitive pricing or competition on participants in energy markets. The scope of each paper is narrow but focused, dealing with one particular aspect of competition in each market under study. It is hoped that results from these three studies could provide valuable policy lessons to public policy makers in their task to create or maintain competition in different energy markets, so as to improve efficiencies in these markets. The first and second papers examine the load shifting behavior of industrial customers in Ontario under real time pricing (RTP). Using Hourly Ontario Energy Price (HOEP) data from 2005 to 2008 and industry-level consumption data from all industrial customers directly connected to the transmission grid, the first paper adopts a Generalized Leontief specification to obtain elasticities of substitution estimates for various industry groups, while the second paper adopts a specification derived from standard consumer theory to obtain price elasticity estimates. The findings of both papers confirm that in some industries, industrial customers who are direct participants of the wholesale market tend to shift consumption from peak to off-peak periods in order to take advantage of lower off-peak prices. Furthermore, in the first paper, a demand model is estimated and there is evidence that the marginal effect of hourly load on hourly price during peak periods is larger than the marginal effect during off-peak periods. An important policy implication from the results of these papers is that while RTP is currently limited to industrial customers, it does have positive spillover effects on all consumers. The third paper uses a unique panel dataset of all retail gasoline stations across five Canadian cities from late-2006 to mid-2007 to examine the effect of local competition on market shares and sales of individual stations. The base empirical specification includes explanatory variables representing the number of same brand stations and the number of different brand stations within a 3km radius to identify brand affiliation effect. It is found that the number of local competitors is negatively correlated with market share and sales. More interestingly, a same brand competitor has a larger marginal impact on market share and sales than a competitor of a different brand. These findings suggest that additional local competition leads to cannibalization of market share among existing stations, rather than create new demand. Another implication is that relying only on the number of different brands operating within a geographic market could understate the competition intensity in the local market.

This thesis presents novel algorithms and a revised modeling framework to evaluate residential aggregate electrical demand response performance under scenarios with limited device-state feedback. These algorithms permit the provision of balancing reserves, or the smoothing of variable renewable energy generation, via an externally supplied target trajectory. The responsive load populations utilized were home heat pumps and deferred electric vehicle charging. As fewer devices in a responsive population report their state information, the error of the demand response program increases moderately but remains below 8%. The associated error of the demand response program is minimized with responsive load populations of approximately 4500 devices; the available capacity of the demand response system scales proportionally with population size. The results indicate that demand response programs with limited device-state feedback may provide a viable option to reduce overall system costs and address privacy concerns of individuals wishing to participate in a demand response program. / Graduate

Le réseau de distribution et de transport de l’électricité se modernise dans plusieurs pays dont le Canada. La nouvelle génération de ce réseau que l’on appelle smart grid, permet entre autre l’automatisation de la production, de la distribution et de la gestion de la charge chez les clients. D’un autre côté, des appareils domestiques intelligents munis d’une interface de communication pour des applications du smart grid commencent à apparaître sur le marché. Ces appareils intelligents pourraient créer une communauté virtuelle pour optimiser leurs consommations d’une façon distribuée. La gestion distribuée de ces charges intelligentes nécessite la communication entre un grand nombre d’équipements électriques. Ceci représente un défi important à relever surtout si on ne veut pas augmenter le coût de l’infrastructure et de la maintenance. Lors de cette thèse deux systèmes distincts ont été conçus : un système de communication peer-to-peer, appelé Ring-Tree, permettant la communication entre un nombre important de nœuds (jusqu’à de l’ordre de grandeur du million) tel que des appareils électriques communicants et une technique distribuée de gestion de la charge sur le réseau électrique. Le système de communication Ring-Tree inclut une nouvelle topologie réseau qui n’a jamais été définie ou exploitée auparavant. Il inclut également des algorithmes pour la création, l’exploitation et la maintenance de ce réseau. Il est suffisamment simple pour être mis en œuvre sur des contrôleurs associés aux dispositifs tels que des chauffe-eaux, chauffage à accumulation, bornes de recharges électriques, etc. Il n’utilise pas un serveur centralisé (ou très peu, seulement lorsqu’un nœud veut rejoindre le réseau). Il offr une solution distribuée qui peut être mise en œuvre sans déploiement d’une infrastructure autre que les contrôleurs sur les dispositifs visés. Finalement, un temps de réponse de quelques secondes pour atteindre l’ensemble du réseau peut être obtenu, ce qui est suffisant pour les besoins des applications visées. Les protocoles de communication s’appuient sur un protocole de transport qui peut être un de ceux utilisés sur l’Internet comme TCP ou UDP. Pour valider le fonctionnement de de la technique de contrôle distribuée et le système de communication Ring-Tree, un simulateur a été développé; un modèle de chauffe-eau, comme exemple de charge, a été intégré au simulateur. La simulation d’une communauté de chauffe-eaux intelligents a montré que la technique de gestion de la charge combinée avec du stockage d’énergie sous forme thermique permet d’obtenir, sans affecter le confort de l’utilisateur, des profils de consommation variés dont un profil de consommation uniforme qui représente un facteur de charge de 100%.

This thesis presents a residential thermostat that enables accurate aggregate load control systems for electricity demand response. The thermostat features a control strategy that can be modeled as a linear time-invariant system for short-term demand response signals from the utility. This control design gives rise to linear time-invariant models of aggregate load control and demand response, which is expected to facilitate the design of more accurate load-based regulation services for electricity interconnections and enable integration of more highly variable renewable electricity generation resources. A key feature of the new thermostat design is the elimination of aggregate short-term load control error observed with existing real-time pricing thermostats as they respond to price signals. / Graduate / 0548 / 0791 / 0544 / dchassin@uvic.ca

This thesis proposes equilibrium models to link the wholesale and retail electricity markets which allow for reconciliation of the differing time scales of responses of producers (e.g., hourly) and consumers (e.g., monthly) to changing prices. Electricity market equilibrium models with time of use (TOU) pricing scheme are formulated as large-scale variational inequality (VI) problems, a unified and concise approach for modeling the equilibrium. The demand response is dynamic in these models through a dependence on the lagged demand. Different market structures are examined within this context. With an illustrative example, the welfare gains/losses are analyzed after an implementation of TOU pricing scheme over the single pricing scheme. An approximation of the welfare change for this analysis is also presented. Moreover, break-up of a large supplier into smaller parts is investigated. For the illustrative examples presented in the dissertation, overall welfare gains for consumers and lower prices closer to the levels of perfect competition can be realized when the retail pricing scheme is changed from single pricing to TOU pricing. These models can be useful policy tools for regulatory bodies i) to forecast future retail prices (TOU or single prices), ii) to examine the market power exerted by suppliers and iii) to measure welfare gains/losses with different retail pricing schemes (e.g., single versus TOU pricing). With the inclusion of linearized DC network constraints into these models, the problem size grows considerably. Dantzig-Wolfe (DW) decomposition algorithm for VI problems is used to alleviate the computational burden and it also facilitates model management and maintenance. Modification of the DW decomposition algorithm and approximation of the DW master problem significantly improve the computational effort required to find the equilibrium. These algorithms are applied to a two-region energy model for Canada and a realistic Ontario electricity test system. In addition to empirical analysis, theoretical results for the convergence properties of the master problem approximation are presented for DW decomposition of VI problems.