BTRL ribbon

Optimal Control of District Cooling System

Optimal Control of District Cooling System with Variable-Speed Plant and Distribution and Stratified Cool Storage

Principal Investigators:
Dr. Afshin Afshari, Professor of Practice, Engineering Systems and Management, Masdar Institute
Dr. Peter Armstrong, Associate Professor in Mechanical Engineering, Masdar Institute

An optimal control strategy will be developed for one of Tabreed’s district cooling plants which can later be implemented in other plants by inserting the proper chiller, pump and cooling tower parameters. It is expected that the new control scheme will result in net energy savings of about 20% or more (if chillers/pumps retrofitted) compared to business-as-usual. This project’s main deliverable is a control algorithm derived within an off-line simulation environment which accurately models the plant. During the second year we will implement the advanced control scheme in a real-time environment at the plant and assess/improve it based on actual measurements. We will also implement suggested retrofits (variable speed compressors and pumps) in order to maximize plant efficiency.

The research project will start with one-time measurements of power, flow and head of each type of pump and cooling tower fan in the plant. We are currently monitoring four chillers to characterize part-load performance over a range of outdoor temperature. We are also creating a detailed plant model which will be validated against the measurements. With the model in hand and historical load data for the plant we will develop optimal control strategies for the plant and estimate their annual impact. We will also develop optimal control strategies for the plant when the fixed-speed pumps and fans are retrofit with variable-speed drives. Annual energy savings of 20-30% are expected.

In the second year we will implement the retrofits and controls and demonstrate the real world savings by careful monitoring of the plant for at least six months. In addition we will develop and implement optimal control for charging thermal energy storage. The storage control problem is one of the most interesting challenges one can pursue in the building technology sector and has direct relevance to demand response (DR) and smart buildings/smart grids research.

  • Develop an in-house capacity to implement and maintain model-predictive control of large cooling plants with and without thermal energy storage
  • The plant model built up from the component models must be able to predict capacity and input power given motor speeds and operating conditions
  • The plant model can then be used to determine the optimal combination of speeds for any load and operating condition
  • A grid pattern search or particle-swarm method will be formulated for 24-hour-ahead optimal charging control
  • Various methods of on-line implementation will be explored