Publication Date

2017-08-07

Availability

Open access

Embargo Period

2017-08-07

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PHD)

Department

Civil, Architectural and Environmental Engineering (Engineering)

Date of Defense

2017-06-15

First Committee Member

Gang Wang

Second Committee Member

Wangda Zuo

Third Committee Member

David Chin

Fourth Committee Member

Hongtan Liu

Abstract

Pumps and fans, which are driven by electrical motors, are the integral parts of heating, ventilating and air conditioning (HVAC) systems and a lot of research have been dedicated to improve the efficiency of them. Recent developments, increased the complexity of these systems, made them hard to study and model. To improve the energy efficiency of the system, variable frequency drives (VFDs) are widely applied on the electric motors of pumps and fans to reduce energy consumption. VFDs have two functions: reducing frequency to reduce the speed and match the reduced load and reducing voltage to reduce motor power. VFDs change voltage and frequency together, and they use some preset ratios to do that. The ratio of voltage to frequency squared, called the squared ratio, is applied for centrifugal fans and pumps, which are considered to have a cubic correlation between the motor load and speed. VFD manufacturers did not give any simulation or experimental supporting data for the recommended preset ratios. Faults are another great factor which causes extra energy consumption in the system and fan belt slippage and pressure setpoint override faults are two of the most common faults regarding pumps and fans. Currently, these faults have to be detected by either data driven, model-based or rule-based fault detection and diagnosis (FDD) approaches. For FDD, model based approach generally require high computational time, which makes it unsuitable for real time applications. Rule based approaches used physical flow meters which are either expensive or inaccurate. Accurate fault detection strategies require flow rate measurements. As mentioned, conventional physical flow meters are not usually accurate and have some installment conditions which makes them unsuitable to be installed right upstream of the pumps and fans. Consequently, virtual flowmeters have been emerged in the recent years but the calibration process for using them is still complicated. The first purpose of this study is to investigate energy efficient voltage-frequency ratios of VFDs driving the induction motors of fans and pumps. First, equivalent motor circuit is used to simulate the motor efficiency, then experiments are designed to validate the results. The second purpose of this work is to develop a hybrid (combination of rule based and data driven) FDD approach which can be implemented in building automation system (BAS). Experiential rules are developed based on affinity laws and head-flowrate relationship and used to identify faults and experimental data are used to calibrate the developed relation. Finally, the implementation of the existing virtual flowmeter technology is simplified and improved by making changes in calibration parameters which makes the calculation process easier.

Keywords

HVAC; Pump; Fan; VFD; Energy

Share

COinS