Team Projects

Team:

Lake Superior State University

Name of the energy conservation measure (ECM)

Intelligent Modeling and Control of Steam Plant Operations to Reduce Energy Consumption through Artificial Intelligence

Purpose of ECM:

Lake Superior State University (LSSU) experiences six to eight months of cold weather yearly. LSSU’s Central Heating Plant (CHP) constructed in 1969, generates saturated steam to provide heating needs to seventeen buildings across campus. CHP has three boilers with a total capacity of 3400 HP.  These are modern fire-tube non-condensing boilers and are operating at a peak efficiency of 84 – 87 percent. Currently, the boilers are manually operated. Heating needs in each building vary due to fluctuating weather, activity level, holidays, exam weeks, and many other factors. The on-demand energy needed by any building is not monitored, and as a result, the boilers operate at peak power to provide the necessary heat in the worst weather conditions. Therefore, when lesser heating energy is needed, a substantial amount of energy is delivered to the environment as waste. The purpose of the ECM was to look at AI to reduce unnecessary energy consumption.

Description of ECM:

The team identified ways of applying Artificial Intelligence to control of the boiler. As part of the project, numerous meters and sensors will be installed in the Library to monitor building activities and energy supplied from the CHP. This data, as well as weather patterns and campus-wide activity, will be fed into a neural network, which will predict impending energy requirements for the Library, which is expected to provide a better boiler operational sequence.

The results could be replicated in all major buildings across campus, allowing the boilers to be operated at the optimum sequence to meet campus wide energy needs. Since the meters and sensors are not installed at this time, a rough estimation of the natural gas savings, based on the current heating energy supplies and the fluctuating weather pattern, shows that it is possible to save about 36,000 MCF annually. With this savings, the payback period was estimated to be less than two years.