(32g) Operation of Residential Hybrid Renewable Energy Systems: Integrating Forecasting, Optimization and Demand Response | AIChE

(32g) Operation of Residential Hybrid Renewable Energy Systems: Integrating Forecasting, Optimization and Demand Response


Wang, X. - Presenter, University of California, Davis
El-Farra, N., University of California, Davis
Palazoglu, A., University of California, Davis

Hybrid renewable energy systems, which consist of solar, wind and other energy generation and storage units, have been widely researched in recent years owing to the reduction in traditional fossil resources and increase in the world-wide energy demand. This paper focuses on the optimal design and operation of a standalone community-scale hybrid renewable energy system comprising photovoltaic (PV) cells, wind turbines, diesel generators and storage units such as batteries. A detailed modeling of each system component is presented. The demand-side management as well as the performance of the generation system are considered to guarantee that the electricity demand of the consumer is satisfied and the overall system economic and environmental cost is minimized.

To optimally allocate electricity generation among different units, a precise prediction of the future electricity consumption and the available solar and wind resources is necessary. Both the day-ahead and real-time predictions are made to provide sufficient information for further optimization. For the day-ahead prediction, historical demand as well as solar radiation and wind speed data are used as the basis for the classification of day types, which are determined by the temperature, humidity, seasonality and other related characteristics. During the day-ahead operational management, a general electricity usage and a generation profile are developed to realize the preliminary dispatch strategy. For the real-time prediction, autoregressive and polynomial fitting methods are adopted to facilitate time-series analysis and forecasting. The online forecasting model is able to consider the updated data and modify the overall prediction and optimization.

On the demand side, a demand-responsive scheme is integrated with the operation of power generation. Among all the residential electricity usage, some are rigid in demand which cannot be changed or shifted such as lighting and cooking while others are flexible loads including washing machines, clothes dryers and vacuum cleaners, which are the key factors to be considered to manage the demand side. These loads can be arranged to reshape demand profiles to match renewable output better.

In the optimization stage, the method of model predictive control (MPC) is adopted by considering real-time optimization to minimize multiple objectives over a receding horizon, the most significant of which is to minimize the generation cost within the time window considered. The environmental influences are also accounted for by the indicators of equivalent cost. As constraints, each energy resource has its maximum generation capacity, as well as minimum output level, which decides the operation range of the generation unit. Limits of ramping rate of each resource are also posed to avoid damage caused by excessively rapid change of power output. The formulation is simplified to a convex optimization problem. At each time interval, only the first set of power references calculated by the optimization are implemented. The optimal allocation between different resources drives the generation to meet electricity requirement of the demand side, with minimum operating cost and environmental impact. Different strategies to realize the receding-horizon method are applied and the results are compared. It turns out that the length of the moving horizon has a critical impact on the optimal allocation of electricity generation among the available resources. Proper selection of the horizon allows for an efficient and optimal operation of the energy system.

The strength of the proposed strategy is that it is able to achieve global optimization over a certain time period and at the same time the users are able to obtain sufficient information of the future conditions and thus make rational decisions about electricity management. Compared with traditional operation of energy systems, the on-line process that combines real-time prediction, optimization and demand-responsive scheme will improve system efficiency and take better advantage of all available resources.

We will show several simulation and optimization case studies that demonstrate the effectiveness and applicability of the proposed receding horizon scheme. Future work to connect the current system to the electrical smart-grid to realize large-scale application will also be discussed.


This paper has an Extended Abstract file available; you must purchase the conference proceedings to access it.


Do you already own this?



AIChE Pro Members $150.00
AIChE Graduate Student Members Free
AIChE Undergraduate Student Members Free
AIChE Explorer Members $225.00
Non-Members $225.00