St. Joseph's Hospital
  Tampa, FL

EnergyWitness™ Customer Profiles University of Arkansas St. Joseph's Hospital Iowa State University

Managing a Chiller Plant based on Hourly Operating Costs

St. Joseph's Hospital is changing the culture of how they operate their facility. They're changing to a culture of data, of informed decision-making, and of understanding the cost implications of how the facility is operated. It's started in their central plant with the chilled water system.

The Need for Better Facilities Operational Data

The hospital had come to suspect that there were significant issues within its chiller plant, and that perhaps engineering design issues were at the root of the problems. If true, a significant investment would be needed to re-engineer, and upgrade the system. Additionally there were space comfort issues in several areas of the hospital, possibly related to the chilled water system.

Before seeking funding, the facilities staff needed to prove that engineering design problems did indeed exist. Further, if they could quantify the ongoing wasted cost of current operations, it would provide the necessary justification.

Everything Changes Once You Have the Data

The first step with any EnergyWitness installation is to begin data collection. It started with about 800 points from the JCI system covering the chiller plant and later added some air handlers, growing to about 3,000 points. After collecting data for a few weeks, IDS began to perform a diagnostic review. The first set of problems found were a variety of sensors that were not working or were out of calibration. This directed St. Joseph's to redirect some of its technician's time and prioritize the repair/calibration of these sensors to ensure accurate readings. (This is not unique to St. Joseph's. Every site we've seen has its share of bad sensors.)

With accurate data to review, the next priority became obvious. The chiller plant was operating in a very unstable manner. To examine just one example of what was found, the output of the chillers was cycling by 500 tons across 15 minute intervals. That's a third of the 1,500 ton chiller's capacity.

The tons output load was cycling widely (1–green line) until the setpoint (2–magenta line) was fixed at 42°F, when as you can see, the trashing stopped and the chillers returned to more stable operations.

Because EnergyWitness was collecting data from all points within the plant, it was a quick process to diagnose the cause of the instability—a setpoint program that was fluctuating by nearly 3°F in some conditions. Fixing the setpoint at a constant 42°F immediately stabilized the system. EnergyWitness was able to verify the results beginning at the next 15 minute interval, as shown in the chart to the left.

The Power of Data

With the chiller plant stabilized, we could address the original concerns about engineering design. Over the course of a couple months we were able to identify $226,000/year in lost savings—savings that would be realized by addressing operational issues, but couldn't because of the design of the chiller plant and chilled water loop. We'll take a look at one of the situations found where a low-cost chiller was underutilized.

Chiller 1, the gas-fired chiller (1–dark green line), operates well below its 800 ton capacity (2).

Off-peak the cost difference between the gas and electric chillers is very small, but on-peak the gas chiller (1–blue line) is one-third the cost of the electrics (2).

The chiller plant's three electric chillers, need to maintain minimum loads to prevent surging. The gas-fired chiller is used during on-peak hours when the electric chillers are more expensive to run. Despite its 800 ton capacity, the gas chiller rarely outputs more then 600 tons, and its output trails off during the course of the day. That leaves, on average, 236 tons of unused capacity.

The unused capacity is important because this chiller operates at a much lower cost per ton than the electric chillers do on-peak. The gas chiller (chiller 1) costs an average of $0.07 per ton to operate, while during on-peak hours the electric chillers cost $0.22 per hour. By running below 75% of capacity, there is a lost savings of $83,000 per year.

What's the moral of the story so far? Data = Knowledge.

St. Joseph's facilities director was able to document the various operational, design, and comfort issues with indisputable facts. With that information in hand, he was able to present his case to the hospital CEO for a $2.5 million engineering redesign of the chiller plant and distribution loops.

The entire $2.5 million budget request was approved. Data = Credibility.

Managing by Cost

St. Joseph's is migrating towards a cost-based management approach. In order to do this, they have installed EWChillerPlant, EnergyWitness' chiller plant optimization module that provides hourly operating cost data recalculated at every 15 minute interval. The plant total view of costs ensures that actions that save money in one area but increase costs in another are quickly caught.

St. Joseph's staff can move through time (1), and see data for each system and the entire plant (2). Bypass (3), energy cost (4), and weather data (5) are shown. A tab (6) allows the user to view the underlying models. Total costs in $/hr and $/ton-hr are shown as gauges (7).

The chiller plant speedometer provides a single screen view of the overall chiller plant and its component systems, and also has historical trend data for all of the cost calculations. It is custom configured for St. Joseph's equipment, utility rates, and any other local variables, such as average ground water temperature. Facilities staff can see what it is costing to run the plant or individual components on an hourly basis and a production basis ($/ton-hour).

One example of how this view helps the facilities team is how easy it is to see the impact of rising natural gas prices on running the gas chiller. This has eroded the advantage of using it during on-peak hours.

Another example was a situation where a single chiller was running, but not really able to meet the load demand. It was intentionally shortchanging the load to save money and not start a second chiller. However, taking a system-wide view of the costs, any savings gained in the chiller was lost in the cooling tower. Running the second chiller, both at partial loads, was shown to cost about the same while better maintaining space comfort.

Next Steps

The cultural change is really just beginning. More members of the facilities staff are looking at the available data and understanding the importance of a system-wide view and understanding. The engineering redesign project is underway, but the data is informing decisions as to what the correct course of action is. The chiller plant optimization modules will expand to include the boiler plant (also custom configured for St. Joseph's). And of course we expect to find and document more operational improvements and cost savings.

We'll update this case study periodically as St. Joseph's adds to its list of accomplishments. So, bookmark this page and come back in a few months to see the progress.