Interior space conditioning means heating or cooling building interior space to pro- vide comfort to occupants. In the modern world, the thermostat is a popular form utilized in residential and commercial buildings. Although the thermostat industry has recently matured, the development of new technology provides new opportunities to interior space conditioning. Motivated by the energy crisis, a demand response- enabled interior space conditioning system is designed for residential users. The feature of completely autonomous controls improves the acceptability and usability of the system.

Built on low-cost, low-power wireless technology, the system uses a disaggregated set of sensors and actuators. The software adopts a hierarchical layered structure, providing modularization of functions and semi-independent design. User interfaces provide easy and instructive interaction to users. The system interacts with the public utility, houses and their HVAC systems, users and outside climates. Robust adaptive control is used to address system uncertainties. Validation tools were developed to evaluate the system.

As the major contribution of this research to interior space conditioning, super- visory controls were developed to locate the optimal control settings. Adopting a hierarchical structure, supervisory controls determine control modes, control strate- gies/states, and control settings. To meet users' various requirements on utility cost and thermal comfort, four control strategies/states were designed: the normal strat- egy, the pre-cooling/pre-heating strategy, the pre-conditioning strategy and the over- lapping strategy. The supervisory control strategies were realized by hybrid methods. Expert systems were utilized to choose control mode and control state. Model-based methods or performance-based methods were adopted in each state to seek optimal control settings. Results from computer simulations and field tests indicate that the system responds automatically to price signals with appropriate behavior of energy saving and load shifting. By identifying dynamic signatures of individual houses, the system is able to adapt its control strategies to a house and its HVAC systems as well as to the ever-changing outdoor conditions.

In conclusion, the thesis successfully demonstrates an intelligent, adaptive and autonomous interior space conditioning system under the context of demand response for residential buildings.

Demand response—the management of customer electricity demand in response to supply—has emerged as a promising means of increasing grid reliability by reducing peak demand. One potential technology to enable residential Demand Response (DR) is a Programmable Communicating Thermostat (PCT) that receives price signals from the electrical utility. However, several issues preclude the widespread adoption of this technology and policy. One is the poor adoption and energy-conserving performance of a similar technology and policy—the programmable setback thermostat. Another is lukewarm customer response to residential DR air conditioning cycling programs. Finally, financial incentive alone may not suffice to persistently reduce peak electricity consumption.

A team at UC Berkeley developed an alternative model for a residential demand response enabling technology, called the Demand Response Electrical Appliance Manager (DREAM). The DREAM system acts as both an intelligent thermostat and in-home energy display. DREAM consists of a wireless network of data sensors, appliance actuators and a central controller that can communicate variable price signals. We tested the DREAM in two houses in the summer of 2007 for six weeks. The DREAM controlled the HVAC system, monitored electricity from appliances, sensed temperature and occupancy, and displayed temperature and energy consumption. This paper discusses potential issues for DR policy and technology, and describes potential solutions in improving user adoption.

A number of Demand Response (DR) technologies work by responding to variable electricity pricing, but have not yet been applied to control residential HVAC systems. An autonomous thermostat system, the Demand Response Electrical Appliance Manager (DREAM), provides possibilities to improve price-based demand responsiveness in residences.

Built on low-cost, low-power wireless technology, the system uses a disaggregated set of energy- and environmental sensors. Control strategies are implemented to optimize electricity cost and user’s comfort. To perform the optimization, the system starts from default values and learns the dynamic behavior of a house and HVAC system. A graphic user interface provides easy interaction with the system. Computer simulation, lab tests and field tests have been used to validate the system infrastructure and control strategies. These tests indicate that the DREAM responds automatically to price signals with appropriate energy saving behavior. The system can reduce electricity consumption during peak price hours without significantly decreasing comfort.