Control Techniques in Heating, Ventilating and Air Conditioning (HVAC) Systems

Problem statement: Heating, Ventilating and Air Conditioning (HVAC) systems are among the main installations in residential, commercial and industrial buildings. The purpose of the HVAC systems is normally to provide a comfortable environment in terms of temperature, humidity and other environmental parameters for the occupants as well as to save energy. Achieving these objectives requires a suitable control system design.

Approach: In this overview, thermal comfort level and ISO comfort field is introduced, followed by a review and comparison of the main existing control techniques used in HVAC systems to date.

Results: The present overview shows that intelligent controllers which are based on the human sensation of thermal comfort have a better performance in providing thermal comfort as well as energy saving than the traditional controllers and those based on a model of the HVAC system.

Conclusion: Such an overview provides an insight into current control methods in HVAC systems and can help scholars and HVAC learners to have the comprehensive information about a variety of control techniques in the field of HVAC and therefore to better design a proper controller for their work.

HVAC systems characteristics: The most important specifications of HVAC systems -which are as inherent part of all thermal systems - are Time lags. There are several types of time lags in HVAC systems such as distance-velocity lag, exponential lag and capacity lag. The distance-velocity lag is the time between a signal being sent to an element and the element starting to respond, arising from the finite speed of propagation of the signal. An exponential lag occurs when the change with time in the output from an element or system (resulting from the application of a step change in the Input signal to that element or system), is of simple exponential Form. This lag may be defined as a capacity for storing energy which may be on the demand side of the process such as heated water in a tank, or on the supply side such as the hot water in the primary heating coils. HVAC process has several nonlinear components like temperature and humidity which are nonlinear and extremely interrelated. Moreover, actuators like valves and dampers which perform control actions are nonideal and nonlinear and these nonlinear factors must be identified and compensated for in the design of the controller. Another problem is variable condition of HVAC systems. These variations arise from changeable climatic conditions and variation in occupants' activities which change significantly and periodically from day to night and from season to season. An HVAC system is basically an MIMO system. However, sometimes it may be considered as a SISO system in the design of the controller, but if the aim was full control of the system, the interaction between temperature control and humidity control loops is important and must be taken into consideration.

Control modes in HVAC systems: Many control modes are used in HVAC systems which are categorized in three main groups: 1-Traditional Controllers, 2-Advanced Controllers, 3-Intelligent Controllers

Traditional HVAC controllers :This Controllers have historically relied on pneumatic and electromechanical systems. Pneumatic controls, which use compressed air to regulate the flow of heated or cooled air via mechanically controlled logic, were a foundational technology in HVAC systems.

Advanced Controllers: This Controllers utilize sophisticated control techniques and technologies to optimize system performance, enhance energy efficiency, and enable remote management. The tuning procedure of a PID controller can be a time-consuming, expensive and difficult task. Auto-tuning relieves the pain of manually tuning a controller. PID auto-tuning means automatically determining PID parameters without human intervention. Generally, auto-tuning PID controllers utilize two kinds of algorithms: Model-based algorithms in which the parameters of the PID controller are related to the parameters of a transfer-function model of the plant; Empirical rule-based algorithms in which the parameters of the PID controller are determined by a set of heuristic rules. Although Self-Tuning Control (STC) can offer many advantages and are normally superior to the PID control, this approach is limited to large range applications because model identification is required as initial step, together with model parameter ide

Intelligent Controllers :This category of controllers includes Neural Network based and Fuzzy Logic based Controllers. Since the HVAC systems are MIMO, nonlinear and time-varying systems, Intelligent Controllers seem to be the most proper choices for the control of these systems. Moreover, since the human sensation of thermal comfort is vague and subjective, fuzzy logic theory is well adapted to describe it linguistically depending on the state of the thermal comfort dependent variables. In the following, First Intelligent controllers based on the neural network are inspected and then intelligent controllers based on fuzzy logic are considered. Neural network (NN) as a whole is applied extensively in the control of HVAC systems, but there are two neural network based control systems used more commonly as Neural Network based Predictive Controller and Direct Neural Network Controller. In the Neural Network based Predictive Controller, the NN is exploited for the system model construction used in the control.

There are two system identification approaches: forward system identification and inverse system identification The procedure of training a neural network to represent the forward dynamics of a system is often referred to as the forward system identification approach. The neural network is placed in parallel with the system and the error e between the system outputs y and network outputs y is used to train the network.

Conclusion: The variety of control methodologies as applied in the control of HVAC systems were reviewed and investigated in the present study. Conventional control methods are still the first choice in HVAC systems today. Their traditional use is due to their simplicity of implementations and low initiation cost. However, these methods suffer from a high cost of maintenance and energy consumption. An alternative approach would be to revert to the more modern and intelligent methodologies. It is noted that thermal comfort is inherently a vague issue in humans since they have different definitions for comfort. Therefore, intelligent controllers which are based on the human mind are essentially more reliable and hence apt to providing thermal comfort in the control of HVAC systems. In addition, fuzzy-neural controllers are normally more energy efficient as well.