MBCx at UC Santa Barbara Chilled Water Loop Optimization

While monitoring-based commissioning (MBCx) efforts most frequently focus on individual buildings, central plant projects provide opportunities for energy savings and benefits that extend throughout a campus. At UC Santa Barbara, commissioning optimization reduced chiller system energy use by 1.97 GWh annually, and provided additional benefits at the level of individual buildings through improved reliability of the chilled water supply.

The campus chilled water loop at UCSB is supplied by eight chiller plants. The original design concept was based on functionality; energy efficiency was less of a concern. The staging of the chillers was performed manually, and failed to take advantage of the advanced automated control approaches that are now commonly used. One of the primary goals of this MBCx project was to reduce energy use through improved control of existing equipment, keeping capital costs to a minimum. An additional goal was to maintain more constant and reliable chilled water temperature and pressure throughout the loop.

Using optimizing control algorithms and automated processes, the chilled water system has been optimized from the level of individual chillers to the system as a whole.

The project first required getting accurate baseline operational and energy use data for each chiller on the loop. However, before this project the error margins for these measurements were as high as 20 percent. Therefore before establishing the baseline, the project team made improvements to the metering of the chilled water temperature and flow rate. Once this was complete and the data was collected, missing data was filled in using regression models, and weather data for Santa Barbara was utilized from the MesoWest national weather database.

Optimization of the chilled water loop was approached by first operating each chiller in the most efficient manner. To accomplish this, the engineering team focused on developing methods for balancing power between chiller compressors and peripheral components, such as pumps and cooling towers.

Next, the overall system was optimized by improving the sequences of starting and stopping each chiller in response to overall demand. For example, constant speed chillers are used during times of peak loads when they are most efficient, while variable speed chillers are used during times of low loads. The control sequencing was further improved through predictive staging, controlling chilled water production to meet demands expected a few hours in the future, using weather forecast data. The optimization program will also take advantage of the thermal capacity of the loop for peak load shedding, and will improve the ability for facility staff to detect and locate leaks.

To meet an annual cooling load of 11 billion tons of refrigeration, the chiller loop total electrical consumption had been 8.8 GWh before the MBCx program. The optimization program reduced this to 6.93 GWh, or savings of 1.87 GWh that provided $225,000 in annual utility cost saving. In terms of efficiency, the previous baseline of 0.8 kWh per ton of refrigeration was reduced to 0.63 kWh per ton, an annual overall savings of 21 percent.

The project team estimates that this savings represents 70 percent of what can ultimately be achieved, and they hope to build on this success to reach full optimization of the control strategies. For example, the computer running the JCI Metasys control system is unable to do the complex numerical processes needed for full real-time optimization, so the campus hopes to integrate a more powerful one in the future, potentially running advanced software such as MATLAB. The project also identified equipment upgrades that would improve chiller loop performance and efficiency, for example, by replacement or refurbishment of cooling towers.