Abstract:
Joule heating, a fundamental phenomenon, holds immense importance across diverse
practical domains in science and technology. Its relevance has been further accentuated by advancements in semiconductor devices, MEMS technology, and fabrication
techniques. Leveraging joule heating, microheaters can be meticulously designed to
cater to applications necessitating compact, portable heating solutions. These miniature heating devices boast attributes such as small size, low power consumption, and
the capability to attain exceptionally high temperatures with remarkable precision and
stability. Among the myriad applications, microheaters find extensive use in gas sensors, PCR (Polymerase Chain Reaction), wearable devices, DNA amplification kits,
and the synthesis of nano materials. However, regardless of the specific application,
precise temperature control of microheaters emerges as a pivotal aspect. The ability
to regulate temperature accurately not only influences the performance of the micro heater itself but also determines its efficacy and reliability across a wide spectrum of
applications. Microheater temperature control is dictated by not only the control al gorithm but also the design of microheater, materials used and fabrication techniques
employed.
This thesis studies the design of a single ring membrane type circular symmetric
microheater that is etched away on the substrate. In the literature, Finite Element
Method (FEM) simulations are used to study the temperature distribution of microheaters; however, they tend to be computationally very complex. Hence, in this thesis,
an accurate numerical model has been developed on the basis of heat balance equation
considering conduction and convection heat transfer and heat production in the single
ring microheater. The novelty of the thesis is that the numerical model of the microheater, verified against its COMSOL model, can accurately predict the temperature
distribution in a microheater along the radial length and dynamically over time. This
aspect of dynamic temperature prediction adds a layer of sophistication and practical relevance to the study of microheater behavior, making it a noteworthy contribution
to the field.
In existing literature, the design of microheater temperature control typically involves several approaches. One common method is the utilization of a hardware-in-loop
configuration, where the microheater system is integrated with hardware components to
test and refine the control strategy. Additionally, FEM simulations are often employed
to model and analyze the temperature control dynamics of microheaters. Another
approach, considered simpler yet effective, involves creating an electrical analogous
equivalent of the thermal system. This equivalent system is then used to develop state
space equations, which form the basis for designing and implementing control strate gies tailored to microheater temperature regulation. Indeed, while these approaches
are commonly used, they do have their limitations. Relying solely on real hardware can
be limiting, costly and time-consuming. Similarly, Finite Element Method simulations
can become computationally intensive, especially for complex microheater designs. The
creation of an electrical analogous equivalent, while simpler, may also introduce some
level of approximation or simplification that could impact the accuracy of the control
strategy. These limitations highlight the need for more efficient and accurate methods
for designing microheater temperature control systems. Hence, this thesis studies the
use of real time software implementation of control strategy with the accurate numerical model. A four wire TCR method is used for the sensing of temperature, where
the platinum ring heater acts as both a sensor and an actuator, hence, avoiding the
parasitic losses caused by having an additional sensor in the microheater configuration.
A comparative study is also carried out with the system identification approach for the
model identification of the ring type circular symmetric microheater for control design.
In the realm of microheater controls, literature predominantly discusses linear control methods such as on/off switching, proportional-integral (PI), and proportional-integral-derivative (PID) control. However, this thesis diverges from conventional approaches by acknowledging the inherent non-linearities in microheater thermal systems.
As a result, a novel non-linear gain-scheduled PI controller is developed specifically for
temperature regulation in a ring-type microheater.The thesis conducts a comparative
analysis between the outcomes of a traditional PI control and the newly designed
gain-scheduled PI control, showcasing the efficacy and advantages of incorporating
non-linear control strategies tailored to the complex dynamics of microheaters.
This thesis emphasizes the critical role of microheaters and control strategies in
their temperature management, paving the way for advancements in temperature control across their various applications