Horizon Information Portal

MSOE Library Home

MSOE Library Home

Home

Development of an energy optimized control system for a central plant using biquadratic modeling : a report submitted to the faculty of the Milwaukee School of Engineering in partial fulfillment of the requirements for the degree of Master of Science in Engineering / by David Carrimon.

by Carrimon, David, author., Turney, Dr. Robert, thesis advisor , Acharya, Dr. Kishore, committee member , Shimek, Gary , Milwaukee School of Engineering

[Milwaukee School of Engineering], [2017]

Electric power transmission -- Mathematical models.

Air conditioning -- Efficiency.

Air conditioning -- Control

67 leaves : illustrations, some of which are in color ; 29 cm

Introduction -- Background -- Review of literature -- Methods -- Testing and validation -- Results -- Conclusion -- Recommendations -- Appendix A: MATLAB functions.

For this project, different models for chiller plants and cooling towers were explored in an effort to develop an optimization algorithm that would control the systems to meet specific loads while using the lowest amount of energy possible. The Gordon-Ng universal chiller model and a biquadratic regression model were examined for three chillers connected in parallel, while Braun's NTU-effectiveness model and Lu and Cia's universal cooling tower models were examined for controlling three cooling towers.

The biquadratic regression model for chillers was selected and solved as a mixed integer linear programming, using a slack variable to approximate the quadratic elements. An alternative control scheme was created for comparison featuring two chillers for loads exceeding 400 tons, and three chillers turned on for loads exceeding 900 tons, with the loads divided equally among the chillers in each scenario. A simulation was created to show the load profile for a week. The optimization control resulted in a 4.9% power savings over the week over the alternative scheme.

The cooling towers were arranged such that each tower was dedicated to one chiller. The fan affinity law prevents using biquadratic models to find the optimum power consumption for towers directly. Instead, Braun's and Lu and Cia's models can be employed to find the optimum air flow from the towers, which can then be used in the fan affinity law to find the amount of power needed by the fan. Braun's model was used for a variable speed fan and compared to a fixed-speed fan with the same designated air flow and designed power. Braun's model saved 65.8% power over the course of a week.

Horizon Information Portal 3.25_9885

Powered by SirsiDynix

© 2001-2013 SirsiDynix All rights reserved.

Horizon Information Portal