A practical cost and energy efficient model predictive control (MPC) strategy is proposed for HVAC load control under dynamic realtime electricity pricing. The MPC strategy is built based on a proposed model that jointly minimizes the total energy consumption and hence, cost of electricity for the user, and the deviation of the inside temperature from the consumer’s preference.
An algorithm that assigns temperature set-points (reference temperatures) to price ranges based on the consumer’s discomfort tolerance index is developed. A practical parameter prediction model is also designed for mapping between the HVAC load and the inside temperature. The prediction model and the produced temperature set-points are integrated as inputs into the MPC controller, which is then used to generate signal actions for the AC unit.
To investigate and demonstrate the effectiveness of the proposed approach, a simulation based experimental analysis is presented using real-life pricing data. An actual prototype for the proposed HVAC load control strategy is then built and a series of prototype experiments are conducted similar to the simulation studies. The experiments reveal that the MPC strategy can lead to significant reductions in overall energy consumption and cost savings for the consumer.
Results suggest that by providing an efficient response strategy for the consumers, the proposed MPC strategy can enable the utility providers to adopt efficient demand management policies using real-time pricing. Finally, a cost-benefit analysis is performed to display the economic feasibility of implementing such a controller as part of a building energy management system, and the payback period is identified considering cost of prototype build and cost savings to help the adoption of this controller in the building HVAC control industry.
REVIEW OF LITERATURE
Temperature Regulation in Residential Buildings:
Many people spend most of their days indoor and each day they spend about half of the day in their residences. Therefore the thermal comfort of the residential building is very important and the control of the heating, ventilation and cooling (HVAC) system should satisfy the thermal comfort and energy efficiency requirements.
It is also a very important issue to reduce and optimize the HVAC energy consumption in the residential sector in the context of the global warming effect, since HVAC is the largest contributor to a home’s energy bills and carbon emissions, accounting for 43% of residential energy consumption in the U.S. and 61% in Canada and U.K., which have colder climates (Energy Information Administration, 2009), (Energy, E. P. B. E. S., 1997) and (Rathouse & Young, 2004).
Temperature Regulation in Commercial Buildings:
It is common to have a single HVAC unit and controller to control multiple spaces or rooms in commercial buildings. The controller supplies heating or cooling to all rooms proportionally depending on the temperature reading it gets from one room, assuming that all rooms have the same load and temperature. Lin et al. developed a sensor feedback structure for multiple sensors with single HVAC system control (Lin, Federspiel, & Auslander, 2002).
Temperature Regulation using Model Predictive Control (MPC):
MPC for HVAC systems control has been investigated by several researchers in recent years. Majority of the research in this area focus on increasing energy efficiency using the advantage of time-varying constraints such as allowed room temperature variations.
MPC-BASED HVAC LOAD CONTROL
Model Settings for Energy Efficient Control of HVAC Systems:
In this section a description of the proposed model for control of HVAC systems is provided from the perspective of the consumer. The model has two conflicting goals: 1) minimizing the total energy consumption by the HVAC system and hence the associated costs for the consumer and in parallel, 2) minimizing the deviation of the indoor temperature from the consumer preference. These two goals are combined by a single objective via a weighted squared sum of energy consumption as a function of HVAC
usage and the deviation between the indoor temperature and a reference temperature
Model Predictive Control (MPC) Process:
First, the algorithm for determining the temperature set-points for price ranges is
presented. Later the MPC based prediction approach that is used to map the control
action to the average inside temperatures is discussed.
Experiment for Parameter Identification:
To illustrate the use of the prediction model, an experimental validation that utilizes the above procedure is carried out. In the experiment, data needed for parameter identification is collected from a typical house in Coral Gables, Florida.
MPC CONTROLLER PROTOTYPE
In this chapter, a prototype for the proposed HVAC load control strategy is developed to validate the simulation results through prototype experiments. A cost-benefit analysis is also conducted to identify the payback period considering cost of prototype development and cost savings.
Although the performance of the proposed control strategy is confirmed through simulation studies in the previous chapter, building an actual prototype would not only further validate the simulation results but would also help in assessing the economic feasibility of introducing such a controller as part of a building energy management system. The low cost of build and ease of operation for the controller shown in this chapter will allow the adoption of this controller in the building HVAC control industry.
In this research, a practical cost and energy efficient model predictive HVAC load control (MPC) strategy is proposed for buildings facing dynamic real-time electricity pricing. The proposed MPC strategy aims to reduce the total energy consumption and hence, cost of electricity for the user, while considering the thermal comfort of the consumers by concurrently minimizing the deviation of the inside temperatures from the consumer’s choice of reference temperatures. To achieve this, the model assigns temperature set-points (reference temperatures) to price ranges based on the consumer’s discomfort tolerance index and accordingly generates efficient signal actions for each time period for the AC unit.
Since the proposed MPC strategy is tailored for the energy user, its benefits could be best realized when a component that can effectively capture the consumer’s discomfort tolerance index is introduced and integrated into the controller. Such a component would be able to deduce the tolerance index via direct input, past usage behavior or both for a particular user. An interesting research extension is to develop an effective and practical component that can accomplish this task.
Source: University of Miami
Author: Mesut Avci