Date of Award

May 2013

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Yaoyu Li

Second Advisor

John E. Seem

Committee Members

Yaoyu Li, John E. Seem, Ronald A. Perez, Tien-Chien Jen, Guangwu Xu


(Adaptive) Extremum Seeking Control, Chilled Water System, Dynamic Modeling, Integrated Internal Resistance Detection, Photovoltaic, Self-Optimizing Control


Green buildings are sustainable buildings designed to be environmentally responsible and resource efficient. The Net-Zero Energy Building (NZEB) concept is anchored on two pillars: reducing the energy consumption and enhancing the local energy generation. In other words, efficient operation of the existing building equipment and efficient power generation of building integrated renewable energy sources are two important factors of NZEB development. The heating, ventilation and air conditioning (HVAC) systems are an important class of building equipment that is responsible for large portion of building energy usage, while the building integrated photovoltaic (BIPV) system is well received as the key technology for local generation of clean power. Building system operation is a low-investment practice that aims low operation and maintenance cost. However, building HVAC and BIPV are systems subject to complicated intrinsic processes and highly variable environmental conditions and occupant behavior. Control, optimization and monitoring of such systems desire simple and effective approaches that require the least amount of model information and the use of smallest number but most robust sensor measurements. Self-optimizing control strategies promise a competitive platform for control, optimization and control integrated monitoring for building systems, and especially for the development of cost-effective NZEB. This dissertation study endorses this statement with three aspects of work relevant to building HVAC and BIPV, which could contribute several small steps towards the ramification of the self-optimizing control paradigm.

This dissertation study applies self-optimizing control techniques to improve the energy efficiency of NZEB from two aspects. First, regarding the building HVAC efficiency, the dither based extremum seeking control (DESC) scheme is proposed for energy efficient operation of the chilled-water system typically used in the commercial building ventilation and air conditioning (VAC) systems. To evaluate the effectiveness of the proposed control strategy, Modelica based dynamic simulation model of chilled water chiller-tower plant is developed, which consists of a screw chiller and a mechanical-draft counter-flow wet cooling tower. The steady-state performance of the cooling tower model is validated with the experimental data in a classic paper and good agreement is observed. The DESC scheme takes the total power consumption of the chiller compressor and the tower fan as feedback, and uses the fan speed setting as the control input. The inner loop controllers for the chiller operation include two proportional-integral (PI) control loops for regulating the evaporator superheat and the chilled water temperature. Simulation was conducted on the whole dynamic simulation model with different environment conditions. The simulation results demonstrated the effectiveness of the proposed ESC strategy under abrupt changes of ambient conditions and load changes. The potential for energy savings of these cases are also evaluated. The back-calculation based anti-windup ESC is also simulated for handling the integral windup problem due to actuator saturation.

Second, both maximum power point tracking (MPPT) and control integrated diagnostics are investigated for BIPV with two different extremum seeking control strategies, which both would contribute to the reduction of the cost of energy (COE). In particular, the Adaptive Extremum Seeking Control (AESC) is applied for PV MPPT, which is based on a PV model with known model structure but unknown nonlinear characteristics for the current-voltage relation. The nonlinear uncertainty is approximated by a radial basis function neural network (RBFNN). A Lyapunov based inverse optimal design technique is applied to achieve parameter estimation and gradient based extremum seeking. Simulation study is performed for scenarios of temperature change, irradiance change and combined change of temperature and irradiance. Successful results are observed for all cases. Furthermore, the AESC simulation is compared to the DESC simulation, and AESC demonstrates much faster transient responses under various scenarios of ambient changes.

Many of the PV degradation mechanisms are reflected as the change of the internal resistance. A scheme of detecting the change of PV internal shunt resistance is proposed using the available signals in the DESC based MPPT with square-wave dither. The impact of the internal resistance on the transient characteristics of step responses is justified by using the small-signal transfer function analysis. Simulation study is performed for both the single-string and multi-string PV examples, and both cases have demonstrated successful results. Monotonic relationship between integral error indices and the shunt internal resistance is clearly observed. In particular, for the multi-string, the inter-channel coupling is weak, which indicates consistent monitoring for multi-string operation. The proposed scheme provides the online monitoring ability of the internal resistance condition without any additional sensor, which benefits further development of PV degradation detection techniques.