Research Overview

Control of thermofluid systems lies at the intersection of two traditionally disparate fields of mechanical engineering.  Many types of energy systems such as vapor compression cycles, solar boilers, fuel cells, and biomass reactors, use fluid phase changes and/or chemical reactions to transform or transport energy.  These systems often operate almost exclusively in transient, despite being designed for steady-state operating conditions. Although precise transient control of energy systems like fuel cells is critical to achieving high efficiencies, the traditional static viewpoint of more standard energy systems also results in a missed opportunity for efficiency improvement. This is particularly true for vapor compression cycles, which find wide use in refrigeration, air conditioning, and heat pump applications. Moreover, with legislation phasing out typical HFC refrigerants, more environmentally friendly refrigerants (e.g. CO₂) are being vigorously pursued.

The technological focus of current research projects focus primarily on vapor compression systems and related energy systems.  This necessarily includes research in the areas of dynamic modeling (model development, reduction, and validation) and nonlinear control design (gain-scheduling, Model Predictive Control, sets of stabilizing controllers). We ensure the viability of our research by developing software tools for industrial partners (e.g. The Thermosys Toolbox for MATLAB) and involving direct experimentation for virtually every project (Research Facilities).  Some of our current research projects are listed below.  Please visit our Research Opportunities page for more information on current opportunities for students.

Current Research Projects

Nonlinear Control of Distributed Chiller Systems

Student:         Matthew Elliott

Objectives:     This research consists of the construction of a chiller system with multiple evaporators with dynamic load emulation for the purpose of dynamic model development and controller evaluation. The experimental facility will be custom built and fully instrumented and interfaced to a Data Acquisition system using WinCon/Simulink. The experimental facility will be a platform for ongoing research of dynamic modeling, control and fault detection algorithms. The focus of this project will be to develop and evaluated nonlinear control architectures and strategies for multi-evaporator systems, meeting changing demands for cooling capacity at specified temperatures while optimizing system efficiency. The final stage of this research will seek to develop general strategies for control of massively distributed vapor compression systems.

Control-Oriented Dynamic Models of Distributed Parameter Systems

Students:        Abhishek Gupta and Bhaskar Shenoy

Objectives:     This research seeks to unify the disparately different modeling paradigms in order to capture the salient dynamic input-output behavior of multi-phase fluid heat exchangers. The PDE representation of most heat exchanger dynamics is discretized into a finite number of control volumes for feasible simulation. This approach has its theoretical roots in the Galerkin method for solving PDEs, and permits the user to simulate the time evolution of the entire spatial profile. However, for most control-oriented end-uses, the spatial profile is not needed and a parsimonious, but accurate, input-output model is desired. Indeed, when sufficiently discretized these finite-difference models cease to be minimal representations of the system dynamics, and are less attractive for control synthesis purposes. Recent lumped parameter approaches for modeling multi-phase fluid energy systems have proven sufficiently accurate for model based control design, but require accurate estimates of lumped heat transfer coefficients and other distributed parameters. This research will bridge the gap between these two modeling approaches by developing model reduction methods for constructing lumped parameter control-oriented models resulting in minimal input-output prediction error relative to the complex PDE models. The resulting models will be included in the Thermosys Toolbox for Matlab, a suite of tools for modeling, control, and fault detection of vapor compression systems. All modeling efforts will be validated with experimental data.

Gain Scheduled Control Using the Dual Youla Parameterization

Student:         Young Joon Chang

Objectives:     This research seeks to demonstrate that gain scheduling using the Youla parameterization results in greater stability and better performance. Gain scheduled control is generally approached as a control synthesis problem (i.e. synthesize an LPV controller for a given LPV plant, scheduled on the same set of varying parameters) or as a control blending problem (blend a set of predefined controllers and determine stability of the resulting closed loop system). While the former case has advantages in terms of guaranteed stability, the latter is most often used in industry under the term Local Controller Network. This approach uses a simple weighted average of all controllers’ outputs. However, recent research shows that this is merely a special case of a more generalized framework for gain scheduling using the dual Youla parameterization. Moreover, intuition and empirical evidence indicate that this more generalized formula is superior to the standard LCN in terms of both stability and performance. However, this remains to be formally proven. Furthermore, the LPV synthesis techniques could also be extended to the Youla parameterized framework with improved results expected. This research will seek to evaluate both of these opportunities and the resulting theoretical extensions.

Custom Dynamic Modeling of Land and Air Vehicle Vapor Compression Systems

Students:        Aarti Ramani and Bhaskar Shenoy

Objectives:     This research is responsible for developing custom dynamic models for the climate control systems used in various land and air vehicles. Particular focus is placed on developing models and software tools that execute in real-time and are compatible with embedded systems. This research also offers several unique challenges, such as dynamic modeling of vapor compression systems operating at ambient temperatures of -32° C. Nondisclosure agreements prevent dissemination of any further details.

Advanced PID Control Design and Evaluation for Building HVAC Systems

Student:         Sachin Pingle

Objectives:     The consensus of HVAC control engineers is that the majority of building HVAC control loops do not operate as intended. The problem can be traced to poor or limited tuning, or changing system dynamics. This research seeks to apply and extend some of the recent theoretical advancements made at Texas A&M University in the area of PID control synthesis to building HVAC systems. In cooperation with the Energy Systems Lab at TAMU, data from many different HVAC control loops will be used to assess controller effectiveness. Selected controllers will be redesigned, implemented, and evaluated for long term effectiveness.

Instrumentation of a Residential Heat Pump and Air Conditioning System

Students:        Byron Bolding, Sean Elliston, Shana Van Fleet

Objectives:     This undergraduate research project consists of instrumentation of a residential heat pump and air conditioning system. The appropriate sensors and control hardware will be identified, purchased, and installed. Software programs for data acquisition and control will be developed. The resulting experimental testbed will allow observation of all relevant dynamic phenomena and real-time control of compressor speed, valves opening, fan speeds, etc. and will be the basis for several subsequent research projects.

Start-up and Shut-Down Dynamics of Air-Conditioning Systems

Student:         Sean Elliston

Objectives:    This research seeks to characterize the dynamics of a typical air conditioning system during start-up and shutdown phases. Experimental evaluation of these transients will be made, and simplified models developed. Optimal feedforward control strategies will be developed and assessed.

Air-Conditioning System Control Strategies for Regulating Temperature and Humidity

Student:         Byron Bolding

Objectives:    This research consists of developing and evaluating control strategies for regulating temperature and humidity using air conditioning systems with multiple actuation devices. Limitations and potential efficiency tradeoffs will be experimentally evaluated.

Experimental Evaluation of Novel Valve Actuator

Student:         Zachary Walton

Objectives:    This research consists of design and experimental evaluation of a novel valve actuator. The objectives are improved control with greater component longevity. More details will be posted after an evaluation of patent potential has been completed.