Design, analysis, operation, and advanced control of hybrid renewable energy systems
Date
2015
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Because using non-renewable energy systems (e.g., coal-powered cogeneration
power plants) to generate electricity is an unsustainable, environmentally
hazardous practice, it is important to develop cost-effective and reliable renewable
energy systems, such as photovoltaics (PVs), wind turbines (WTs), and fuel cells
(FCs). Non-renewable energy systems, however, are currently less expensive than
individual renewable energy systems (IRESs). Furthermore, IRESs based on
intermittent natural resources (e.g., solar irradiance and wind) are incapable of
meeting continuous energy demands. Such shortcomings can be mitigated by
judiciously combining two or more complementary IRESs to form a hybrid renewable
energy system (HRES). Although previous research efforts focused on the design,
operation, and control of HRESs has proven useful, no prior HRES research endeavor
has taken a systematic and comprehensive approach towards establishing guidelines
by which HRESs should be designed, operated, and controlled. The overall goal of
this dissertation, therefore, is to establish the principles governing the design,
operation, and control of HRESs resulting in cost-effective and reliable energy
solutions for stationary and mobile applications.
The first empirical part of this dissertation focuses on HRES equipment
selection and sizing using an economic and feasibility analysis. We determined that
HRES components and their sizes should be rationally selected using knowledge of
component costs, availability of renewable energy resources, and expected power
demands of the application. To demonstrate this statement, we ascertained the
economically preferred and feasible renewable energy system types and sizes for a
range of average annual wind speed, average annual solar irradiance, power demand
and minimum renewable fraction. We found that under some combinations of these
variables HRESs are less expensive than IRESs and grid-supplied energy. This result
has a significant implication for renewable energy systems becoming increasingly
ubiquitous.
After HRES type and sizes are selected, it is necessary to determined how the
components are arranged and what control systems are best for coordinating the entire
system to meet a set of operating objectives (typically satisfy a power demand while
storing excess power for later use). This was demonstrated by way of a case study
involving the design, control, and economics of a University of Delaware FC/battery
bus retrofitted with a roof-installed PV array. The preferred arrangement of HRES
components is the FC in series with the battery (the FC is committed to maintaining a
desired battery state of charge (SOC) while the battery meets the majority of the bus
power demand) while the PV array is used to assist the battery in meeting the bus
power demand, with excess PV power being used to charge the battery. Simulation
results indicate that under a variety of operating conditions, a PID control strategy is
best for enabling the bus to satisfy required power demands and maintain a desired
battery SOC. An economic analysis of the PV investment necessary to realize the
HRES design objectives indicates that the investment will pay for itself in Newark,
DE, establishing the economic viability of the proposed addition of a PV array to the
existing University of Delaware FC/battery bus.
Although the performance of standard controllers for HRESs is generally
satisfactory, operating objectives can be met more reliably and efficiently when
information-rich data is used to coordinate HRES component. Two such data-driven
control paradigms were developed in separate case studies. In the first case study, we
developed and evaluated a method for reconfiguration of control loops in
decentralized control schemes using directed spectral decomposition of collected
process data. This technique was applied to a stirred mixing tank process and was
shown to improve control performance under changing operating conditions. The
second case study, which involved data-driven centralized control, establishes a
method for adaptive data-driven MPC of a PV/WT/FC/battery/electrolyzer HRES for
a single-family home using measured disturbance prediction and model adaptation.
This procedure was shown to enable the HRES to better track a power demand
setpoint, resulting in increased system reliability and efficiency.
Finally, a PV/WT/battery HRES called an OMNi-Charger, which is intended
for small-scale (~1 W) remote power applications) was experimentally validated under
varying power demand, wind speed, and solar irradiance during controlled and field
tests. The OMNi-Charger controller—a simple state controller that enacts a prespecified
control action depending on the current state or operating conditions of the
system—was shown to enable the OMNi-Charger to meet varying power demands
nearly instantaneously and consistently despite rapid and unknown fluctuations in
wind speed and solar irradiance. Although the OMNi-Charger warrants some future
investigation, the device should prove useful as a cell phone charging station or a
power source for lighted ocean buoys.