Browsing by Subject "Demand response"
Now showing 1 - 10 of 10
- Results Per Page
- Sort Options
Item Demand side load control in residential buildings with HVAC controller for demand response(2015-05) Yoon, Ji Hoon; Baldick, Ross; Novoselac, Atila; Arapostathis, Aristotle; Liedl, Petra G; Kwasinski, AlexisDemand Response (DR) is a key factor to increase the efficiency of the power grid and has the potential to facilitate supply-demand balance. Demand side load control can contribute to reduce electricity consumption through DR programs. Especially, Heating, Ventilating and Air Conditioning (HVAC) load is one of the major contributors to peak loads. In the United States, HVAC systems are the largest consumers of electrical energy and a major contributor to peak demand. In this research, the Dynamic Demand Response Controller (DDRC) is proposed to reduce peak load as well as saves electricity cost while maintaining reasonable thermal comfort by controlling HVAC system. To reduce both peak load and energy cost, DDRC controls the set-point temperature in a thermostat depending on real-time price of electricity. Residential buildings are modeled with various internal loads using building energy modeling tools. The weather data in different climate zones are used to demonstrate that DDRC decreases peak loads and brings economic benefit in various locations. In addition, two different types of electricity wholesale markets are used to generate DR signals. To assess the performance of DDRC, the control algorithms are improved to consider the characteristics of building envelopes and HVAC equipment. Also, DDRC is designed to be deployed in various areas with different electricity wholesale markets. The indoor thermal comfort on temperature and humidity are considered based on ASHRAE standard 55. Finally, DDRC is developed to a hardware using embedded system. The hardware of DDRC is based on Advanced RISC Microcontroller (ARM) processor and senses both indoor and outdoor environment with Internet connection capability for DR. In addition, user friendly Graphic User Interface (GUI) is generated to control DDRC.Item Dynamic modeling and optimal scheduling of chemical processes participating in fast-changing electricity markets : a data-driven approach(2021-08-10) Kelley, Morgan Taylor; Baldea, Michael; Baldick, Ross; Edgar, Thomas F; Allen, David TCompetitive pressure from global market forces has placed a heightened emphasis on information exchange, coordination, and integration of all decision-making layers of the chemical supply chain. Significant developments in this area, supported by advances in computer hardware, data exchange, storage, and optimization algorithms, have potential for incurring substantial economic benefits for chemical operations. This coordination often extends to inclusion of power grid and power supply networks such as distributed energy systems into the operation and control of chemical systems. Two essential layers in the decision-making hierarchy of a chemical enterprise are production scheduling and process control. The interface between scheduling and control represents, in effect, an interaction between business-driven decisions (scheduling) with situation- and safety-driven decisions (control). Integrating these activities is therefore key in maximizing operational profits by meeting demand and ensuring that production targets (i.e., the setpoints transmitted to the control system) are met safely in the presence of disturbances and operational uncertainties. Integrating scheduling and control becomes particularly important when a chemical plant operates in fast-changing markets (e.g., markets with real-time electricity pricing). In order to operate more profitably under such circumstances, the plant must be able to quickly change production rates or product grades, taking advantage of excess production capacity and product storage facilities when available. This is reflected in fast and frequent changes in scheduled production targets, often over time intervals shorter than the (closed-loop) time constant of the process. As a result, the process may permanently operate in a transient mode (as opposed to being predominantly at or close to a steady state). Under these circumstances, it is crucial that the scheduling calculations account explicitly for the dynamics of the process and the performance of its control system, such that the aforementioned scheduled transitions are feasible and economically optimal. In the first part of this dissertation, I will focus on computational efficiency in solution frameworks for doing production scheduling under dynamic constraints. I will first present a mixed-integer linear program (MILP) framework using scale-bridging models (SBMs). To further increase efficiency, I implement a Lagrangian relaxation. I compare these modeling and optimization efforts to traditional reduced-order modeling methods and full-order modeling methods. In the next chapter, I apply the linear framework from the previous chapter to a nonlinear problem and determine the transferrability of the proposed method. In Part II, I apply the MILP framework to an Air Separation Unit (ASU) and utilize SBMs to represent the system. I provide a base-case scheduling problem where I examine the benefits of transient operation of ASUs for use in load shifting for fast-changing electricity markets. From here, I expand the discussion to considerations necessary for accounting for uncertainty, presenting two methods: (i) optimization under uncertainty and (ii) a feedback/ moving horizon framework. I dedicate a chapter to each of these methods and then compare the two to close out Part II of this dissertation. In Part III, I extend my work on scheduling in fast-changing electricity markets in three directions. In the first, I adapt the scheduling framework and models from Part II to the reduction of grid-side greenhouse gas emissions and quantify the cost of prioritizing emissions-reduction over cost-reduction for operation of an ASU. In the second, I identify SBMs for capturing the dynamics of a multiproduct ASU, utilizing a years' worth of historical data. I implement these SBMs in a scheduling problem to quantify the economic benefits of operating an industrial ASU in a fast-changing electricity market. In the third, I examine the potential for transient operation of an ammonia synthesis process in a fast-changing electricity market. In the final chapter, I provide broad conclusions for each part of the dissertation, commenting on the economic savings potential of transient operation of chemical processes in fast-changing electricity markets and identify areas for future work in both model identification, computational efficiency, and potential future applicationsItem Economic forecasting and optimization in a smart grid built environment(2013-08) Sriprasad, Akshay; Edgar, Thomas F.This Master’s Report outlines graduate research work completed by Akshay Sriprasad, who is supervised by Professor Tom Edgar, in the area of modeling and systems optimization for the smart grid. The scope this report includes the development and validation of strategies to elicit demand response, defined as reduction of peak demand, at the residential level, in conjunction with collaborative research efforts from the Pecan Street Research Institute, a smart grid research consortium based in Austin, TX. The first project outlined is an artificial neural network-‐based demand forecasting model, initially developed for UT’s campus cooling system and adapted for residential homes. Utilizing this forecasting model, a number of demand response-‐focused optimization studies are carried out, including optimization of community energy storage for peak shifting, and electric vehicle charging optimization to harness inexpensive night-‐time Texas wind energy. Community energy storage and electric vehicles are chosen as ideal dynamic charging media due to increased proliferation and focus of Pecan Street Research Institute on critical emerging technologies. As these two technologies involve significant capital investment, an alternative mobile application-‐based demand response strategy is outlined to complete a comprehensive portfolio of demand response strategies to suit a variety of budgets and capabilities.Item Employing chemical processes as grid-level energy storage devices : demand response and frequency regulation(2020-02-10) Otashu, Joannah Igbe; Baldea, Michael; Edgar, Thomas F; Truskett, Thomas M; Bakolas, EfstathiosReliable operation of the electric grid which comprises of instantaneously matching electricity demand and supply is a tier one security concern and has received increased attention in recent years. This is a challenging task as both electricity demand and supply are characterized by variability and uncertainty; the volatility on the supply end originating from increased integration of renewable energy in power generation. Electricity demand typically peaks in the late afternoons whereas wind and solar generation peak in the early mornings and mid-day respectively. To reconcile this electricity supply and demand mismatch, energy storage is employed for storing excess power generated during off-peak demand periods. Correspondingly, stored power is discharged during peak electricity demand hours. Electricity storage is however presently expensive and current grid storage devices are insufficient to ensure reliable electric grid operation. Standby generators with fast ramp rates which are generally more expensive to run are also employed when needed. A solution to this problem involves motivating end-users of electricity to modify their demand patterns in manner that enhances grid stability. This activity, referred to as demand response (DR) can help relieve the strain on the grid by tailoring end-user power consumption to match the time varying supply of electricity. Chemical industries that are power-intensive can provide valuable DR as they represent localized grid loads with end products that are generally easier and cheaper to store compared to electricity. These industries can ramp production during off-peak power demand hours storing excess product and reduce production rates when power demand peaks while depleting inventory to meet any required product demand. However, the operation of chemical industries are subject to strict constraints including; economic (e.g., reliable product delivery), feasibility (e.g., limits on process variables) or safety concerns. Additionally, the dynamics of chemical processes are typically described by high-order non-linear models and ramp rates and capacity limits alone, which are typically employed on the grid side to manage demand response activities, are not sufficient to characterize the transient nature of chemical processes. In this dissertation, we employ modeling techniques and production scheduling to evaluate the provision of demand response by chemical industries in highly volatile electricity markets. We develop and illustrate the use of suitable dynamic models to represent the chemical plant operation while providing fast-paced DR. As fast DR contribute significantly to the solution pool available for handling power grid contingencies, this study will aid the smoother deployment of industrial chemical loads as grid support devices while ensuring safe and feasible operation of the chemical plant.Item Interruptible load program for the Argentinian electricity market : an economic instrument(2023-04-14) Nava, Sabrina Solange; Spence, David B.; Zarnikau, Jay William, 1959-; Tuttle, DavidThis research proposes a new regulatory framework for an interruptible load program aimed at large users in the Argentine electricity market. A Regulatory Impact Analysis (RIA) approach is utilized to assess the impact of this new regulation across different perspectives: technical, economic, energy and environment, and regulatory. The technical analysis provides insights into the alignment of peak demand hours of the system and large users. It identifies 266 MW as the technical potential for load interruption, which could meet almost 1% of the peak demand in 2021. Only 15 large companies are expected to contribute 35% of the interruptible load, and the chemical, rubber, plastic, and other non-metallic mineral materials industries show the most promising potential. The economic approach suggests an initial price cap of 106 USD/MWh (3-hour events) for interruptible load payments. The program would have generated greater monetary savings during the winter periods of 2021, indicating that seasonal variations in energy demand may impact the program's effectiveness. The section on energy and environment underscores that the geographical distribution of potential providers and costly generation is a significant challenge. While some regions could be balanced, others, such as the Patagonic region, may present a barrier to fully implementing reductions. Gas and steam turbines are the technologies affected by the program, with natural gas and gas oil being the most frequently reduced fuels. The characteristics of the electricity matrix at the time of program application are crucial for achieving emission savings. The regulatory analysis emphasizes the importance of designing an effective regulatory framework for the success of the program. It identifies the shortcomings of the current regulation and confirms the necessity for significant changes or the creation of a new one. To mitigate risks associated with the implementation, three essential approaches are identified: leveraging best practices from successful programs, applying transparent stakeholder consultation processes, and developing detailed and comprehensive rules. These findings can provide valuable insights for policymakers seeking to improve the regulatory framework for interruptible load programs in the electricity market. Finally, the study limitations and potential areas for future research are discussed to offer guidance for further investigation.Item Multi-agent reinforcement learning for demand response and load shaping of grid-interactive connected buildings(2020-09-14) Vázquez Canteli, José Ramón; Nagy, Gyorgy ZoltanIncreasing electrification, integration of renewable energy resources, rapid urbanization, and the potential shift towards higher integration of electric vehicles and other distributed energy resources, such as batteries, represent challenges that the energy sector will have to face in the future. As these changes begin to happen, buildings will have more energy storage and generation capacity, which can create additional uncertainty and volatility, but also more flexibility if these systems are appropriately controlled. Demand response can enable consumers to reduce their energy consumption through load curtailment, shift their energy consumption over time to periods of higher renewable energy generation, or generate and store energy at certain times to provide the grid with more flexibility. Model-based control approaches, such as model predictive control, can provide near optimal control policies for demand response. However, they require to develop accurate models of the systems to be controlled, which is not always a cost-effective scalable option in complex, unpredictable or non-stationary systems. On the other hand, model-free control algorithms, such as reinforcement learning, have the potential to provide good control policies in a cost-effective manner. In this dissertation, I review the literature on reinforcement learning for demand response and control of urban energy systems, and introduce a novel multi-agent reinforcement learning algorithm specifically designed to be decentralized, scalable, and implemented in a demand response setting. This multi-agent reinforcement learning algorithm outperforms its single-agent counterparts and a baseline rule-based controller by more than 15% on an average of 5 different metrics: load factor, net electricity consumption, peak demand, average daily peak demand, and ramping. I test the controllers in CityLearn, an Open AI Gym environment I created for the implementation of single and multi-agent reinforcement learning for demand response. CityLearn is also intended to be used by other researchers and tackle another problem I observed in my review of the literature: lack of standardization and reproducibility of most research in RL applied to energy management in urban settings.Item SCALEX: SCALability EXploration of multi-agent reinforcement learning agents in grid-interactive efficient buildings(2023-08) Almilaify, Yara Sami; Nagy, Gyorgy ZoltanThe transition to renewable energy sources and the decarbonization of buildings bring new challenges to grid-interactive efficient building (GEB) communities such as grid stability issues and energy system integration. The dynamic and stochastic nature of intermittent renewable energy production makes it challenging for conventional building control systems to maximize them. To overcome this, advanced control architecture and energy storage are utilized to achieve energy flexibility. Reinforcement learning (RL) offers potential solutions, but its scalability and computational demands in large-scale settings remain unclear. This thesis examines the scalability of Soft-Actor Critic (SAC) in multi-agent systems, comparing decentralized-independent SACs and centralized SACs using CityLearn, an OpenAI Gym environment. We consider neighborhoods consisting of 2 to 64 single-family residential buildings, each equipped with cooling and heating storage devices, domestic hot water storage devices, electrical storage devices, and solar PV systems. In this work, we delve into the challenges faced by these controllers when scaled in a multi-agent system. Our findings suggest that decentralized-independent controllers outperform the centralized controller with an increasing number of buildings. We also show that the performance on the building level can differ from the aggregated performance.Item Sustainable energy roadmap for Austin : how Austin Energy can optimize its energy efficiency(2010-12) Johnston, Andrew Hayden, 1979-; Oden, Michael; Spelman, WilliamThis report asks how Austin Energy can optimally operate residential energy efficiency and demand side management programs including demand response measures. Efficient energy use is the act of using less energy to provide the same level of service. Demand side management encompasses utility initiatives that modify the level and pattern of electrical use by customers, without adjusting consumer behavior. Demand side management is required when a utility must respond to increasing energy needs, or demand, by its customers. In order to achieve the 20% carbon emissions and 800 MW peak demand reductions mandate of the Generation, Resource and Climate Plan, AE must aggressively pursue an increase in customer participation by expanding education and technical services, enlist the full functionality of a smart grid and subsequently reduce energy consumption, peak demand, and greenhouse gas emissions. Energy efficiency is in fact the cheapest source of energy that Austin Energy has at its disposal between 2010 and 2020. But this service threatens Austin Energy’s revenues. With the ascent of onsite renewable energy generation and advanced demand side management, utilities must address the ways they generate revenues. As greenhouse gas emissions regulations lurk on the horizon, the century-old business model of “spinning meters” will be fundamentally challenged nationally in the coming years. Austin Energy can develop robust analytical methods to determine its most cost-effective energy efficiency options, while creating a clear policy direction of promoting energy efficiency while addressing the three-fold challenges of peak demand, greenhouse gas emissions and total energy savings. This report concludes by providing market-transforming recommendations for Austin Energy.Item Techno-economic methods for analyzing the energetic and economic effects of solar, storage, and demand response(2020-09-11) Bandyopadhyay, Arkasama; Webber, Michael E., 1971-; Leibowicz, Benjamin D.; Hall, Matthew; Baldick, RossGrowing population, changing climate, urbanization, and rising economic activities have led to an overall increase in electricity demand. Maintaining the balance between supply and this increasing demand often necessitates the usage of old, inefficient, and environmentally-polluting generators as well as the construction of expensive generation, transmission, and distribution infrastructure. Demand response initiatives (e.g. time-varying electricity prices) and distributed energy resources (DERs), like solar photovoltaic panels and onsite energy storage systems, can help offset a portion of this demand while simultaneously reducing harmful emissions. DERs additionally provide a variety of value streams including peak load reduction, energy arbitrage, real time price dispatch, demand charge reduction, congestion management, voltage support, etc. The impact of price-based demand response and DERs at the electricity distribution level is assessed in this dissertation through the following three studies: (1) quantifying the reduction in 4 coincident peak (4CP) loads and Transmission Cost of Service (TCOS) obligations of electric utilities using local distributed solar and storage, (2) evaluating the peak load reduction/shift potential of time-varying electricity pricing in the residential sector, and (3) investigating the combined energetic and economic potential of DERs and time-varying electricity pricing in the residential sector. When the Electric Reliability Council of Texas (ERCOT) peaks for a single 15-minute interval during each summer month between June and September, the loads of individual Distribution Service Providers (DSPs) in the same time interval are recorded. The averages of these DSP loads, defined as 4CP loads, are used to calculate TCOS obligations that each DSP must pay Transmission Service Providers (TSPs) in the next calendar year as compensation for using their transmission infrastructure. First, a generalized tool is built to forecast the change of 4CP loads and corresponding TCOS obligations for electric utilities within ERCOT based on varying amounts of solar and storage capacity. The tool is illustrated by using empirical electricity demand data from the municipally-owned utility in Austin, TX (Austin Energy) and solar generation data from the PVWatts calculator developed by the National Renewable Energy Laboratory. TCOS obligations can be on the order of tens of millions of dollars. Results indicate that solar and storage capacity can substantially lower these payments. For example, a 20 MW increase in local solar capacity in 2018 would reduce Austin Energy’s payment by an estimated $180,000 for each subsequent year. By using the novel approach of incorporating coincident peak demand charge reductions at the distribution level, the economic value of local generation and storage is highlighted. Next, a convex optimization model is developed to analyze the potential for time-varying electricity rate structures to reduce and/or shift peak demand in the residential sector. In this model, a household with four major appliances minimizes electricity costs, with marginally increasing penalties for deviating from temperature set-points or operating appliances at inconvenient times. The four specific appliances included are: heating, ventilation and air-conditioning (HVAC) systems, electric water heaters (EWHs), electric vehicles (EVs), and pool pumps (PPs). The study incorporates a one-parameter thermal model of the home and the electric water heater, so that the penalties can apply to the room and water temperatures rather than the total appliance loads. Analysis is performed on a community of 100 single-family detached homes in Austin, TX. These homes each host a combination of the four end-use devices while some also have onsite solar panels. Results show that dynamic pricing effectively shifts the residential peak away from the time of overall peak load across the electricity system, but can have the adverse impact of making the residential peak higher. The energy consumption does not differ significantly across the different rate structures. Thus, it can be inferred that the time-varying rates encourage customers to concentrate their electricity demand within low-price hours to the extent possible without incurring significant inconvenience. By incorporating the novel approach of including monetary value of customer behavior in price-based demand response models, this study builds a tool to realistically quantify peak load reduction and shifts in the residential sector. Finally, the convex optimization model is extended to consider larger sets of distributed technologies that might be deployed in homes and investigate how different combinations of these technologies affect peak grid load, energy consumption from the grid, and emissions in the residential sector under time-varying pricing structures. In the model, households with varied amalgamations of distributed energy technologies minimize electricity costs, amortized capital, and operational costs over a year, with marginally increasing penalties for deviating from room temperature set-points. The four technologies considered are: solar photovoltaic (PV) panels, lithium-ion batteries, ice cold thermal energy storage (CTES), and smart thermostats. Results show that from an economic perspective, it is optimal for residential customers to install solar panels under tiered rates, time-of-use rates, and critical peak prices while it is cheapest to own a combination of solar panels and smart thermostats when real-time prices and demand charges are in effect. The capital and installation costs of both storage systems are still too high to make them economically profitable investments for typical residential customers. Additionally, solar panels are the main instruments to reduce energy purchased from the grid and carbon dioxide emissions under all pricing schemes. Adding smart thermostats can reduce these metrics to a greater extent by making the home energy-efficient. Further, while the energetic effect of the two storage systems can be favorable or detrimental depending upon the load profile of the particular household and the pricing structure, lithium-ion batteries are the main instruments to avoid high demand charges by spreading the demand in the home (and power bought from the grid) evenly to the extent possible without incurring significant customer discomfort. Thus, this study recommends that residential customers invest in solar panels and smart thermostats to minimize overall annual expenditure and make their homes environmentally efficient. Further, as an effective peak load control mechanism, electric utilities should offer significant rebates to encourage residential customer investment in storage systems in addition to subjecting them to demand charges. Electricity generation from intermittent renewable energy sources has grown rapidly worldwide. DER installation levels continue to rise with the decline in capital costs of energy storage systems and local renewable generation assets, the growth of supportive government policies, and rising concerns about climate change among the masses. Additionally, electric utilities are increasingly employing demand response initiatives to curtail and/or shift peak demand. As a whole, the body of work developed in this dissertation can be used by electric utilities to make optimal decisions about dynamic rate design and policies for increased DER adoption. It can also be used by residential electricity customers to maneuver their own energy consumption patterns and assess the economic viability of investing in DERs.Item Utility management of plug-in electric vehicle residential charging(2014-05) Hernandez, Guillermo, active 21st century; Baldick, Ross; Webber, Michael E., 1971-The purpose of this study is to identify realistic opportunities and barriers regarding PEV charge management by analyzing real-world PEV data from customers in the Austin Energy service area and evaluating direct, quantifiable economic value benefits as it relates new revenue, cost avoidance, CO2 reductions, and MW potential for peak shaving. The main objective is to provide business analysis to support the strategic road-map for Austin Energy PEV home charging programs. Three main charge program implementations are considered: Uncontrolled Charging, Time of Use Rates, and One Way Utility Control. The data used for the analysis includes 45 households with PEVs from Mueller area; 24 were under a Time of Use trial with pricing incentives to charge at night, and 21 receive normal Austin Energy rates. Data analysis shows that 66% of Time of Use trial group successfully shifted PEV load to Off Peak hours (10:00PM to 6:00AM). The potential of One Way control, based on load availability for interruption, shows that it will not be possible to implement until there are 37,000 PEVs in the Austin Energy area. Uncontrolled Charging represents a risk by increasing load during the residential peak. Time of Use Rates program will incentivize load shifting, reduce wholesale energy costs for Austin Energy while allowing customers to reduce their overall electricity bill.