|
Article Excerpt
INTRODUCTION
The control of an air-conditioning system plays a major role in its performance and is one of the most challenging problems in the field of process control. The challenge in controlling air-conditioning systems lies in achieving maximum thermal comfort at minimum energy consumption. The design of an efficient air-conditioning controller that maintains a good thermal environment in a given space at minimum energy consumption largely depends on the availability of an accurate dynamic model of the system. Modeling of refrigeration and airconditioning systems has been the subject of many studies over the last few decades (Wu and Shiming 2006; Wang and Jin 2000; Shiming 2000; Yiu and Wang 2007). Several researchers have contributed to the development of modeling the dynamics of air-conditioning systems for control purposes. He et al. (1995, 1998) presented a state space multi-input multi-output (MIMO) dynamic model for a vapor-compression system. In their research, only the refrigeration circuit was considered and the effect of humidity in air on a dynamic model was not considered. Browne and Bansal (2002) gave a dynamic component-based model for vapor-compression liquid chillers for predicting performance and on-line fault detection and diagnosis. Huang et al. (2006) and Arguello-Serrano and Velez-Reyes (1999) derived a dynamic model for an air circuit, but the effect of condensation of moisture in air was not considered in their studies. Most of the work in the literature presented dynamic models for refrigeration circuits or air circuits separately. Wu and Shiming (2006) and Shiming (2000) presented a component-based dynamic mathematical model that takes into account the behaviors of a direct expansion (DX) refrigeration plant, i.e., the refrigeration circuit and the variable-air-volume (VAV) air distribution subsystem simultaneously. The model is component-based, and all the internal variables of the system that contribute to the performance of an air-conditioning system were not analytically expressed. The variables (length of two-phase section in the evaporator and condenser, evaporator and condenser pressure, evaporator temperature, duct air temperature and humidity, etc.) were not modeled in their work. Hence, it is difficult to use the Wu and Shiming (2006) and Shiming (2000) models for feedback control design.
In this paper, a state space based lumped parameter dynamic model is proposed for representing a DX air-conditioning system for control purposes. This model represents a full airconditioning system, i.e., the refrigeration subsystem or circuit as well as the air subsystem or circuit simultaneously. The cross-coupling between the refrigeration circuit and the air circuit was also considered. The proposed state space model considers twelve state variables, four regulated variables, four output variables, and five disturbance input variables. The state variables, such as temperature and humidity of air inside thermal space, have a direct effect on thermal comfort. Thus, all the state variables that contribute to energy consumption and maintaining good thermal comfort are considered. Some of these state variables cannot be measured directly, but with the help of this model the behavior of these variables can be estimated by designing an appropriate observer (like extended Kalman filter-based prediction). For example, the length of the two-phase section in the evaporator cannot be measured, but its behavior can be obtained using the proposed model.
The effect of condensation of moisture in air was considered in this model for calculating humidity of duct air and the temperature of the evaporator wall and duct air. If the temperature of the air passing through the evaporator goes below the dew-point temperature, then condensation of moisture in air takes place. The heat energy released due to condensation of moisture in air, specifically for hot, humid climates like most of India, is very substantial. Hence, the condensation of moisture in air must be considered in dynamic modeling of an air-conditioning system for humid climates to obtain better mathematical representation of an actual system.
In this paper, modeling has been done also for an air-conditioning system with ON-OFF control. Here, ON-OFF control implies the air-conditioning system remains on (blower, condenser, and compressor motor are on) until the temperature of the conditioned space is greater than the preset temperature and the system turns off (blower motor on and condenser and compressor motor off) when the temperature of the conditioned space is less than the preset temperature. When the air conditioner is off, then there is fast change in the air temperature and air humidity of a conditioned space. Feedback control is implemented on the basis of air temperature and humidity measurements. The timing of switching on and off an air-conditioning system can be optimized in terms of energy consumption and thermal comfort with knowledge of an ON and OFF model.
The air-conditioning system of a passenger coach of Indian Railway was used as a case study and for experimental validation of the proposed model. The design of a controller for mobile air conditioning, like air conditioning of a passenger railway coach, is more challenging compared to air conditioning of a stationary installation due to the complicating influences of highly varying thermal load, non-uniform air distribution, and reliability of air-conditioning equipment. The thermal load on mobile air-conditioning systems varies with changes in direction and location of the moving vehicle due to the change of exposure to the sun and due to varying wind effects. The ambient conditions also change more frequently in mobile air-conditioning systems than in stationary ones, which also leads to varying thermal load. The conditioned air speed distribution is not uniform due to partitions and different air speed obstruction materials inside the railway coach. The non-uniform air distribution gives rise to variations in temperatures in the thermal space. The efficiency of an air-conditioning controller and energy consumption are directly related, i.e., if the efficiency of the controller is good then energy consumption will be less and thermal comfort will be better.
The performance measure of any air-conditioning system depends on two factors: thermal comfort and energy consumption. So, the performance evaluation function in this paper was designed by taking into account the thermal comfort and energy consumption. Thermal comfort was calculated in terms of a well-known index, Fanger's (1970) predicted mean vote (PMV). The second part of performance evaluation function is energy consumption. In this part, energy consumed by a compressor motor, blower motor, and condenser motor were considered. Presently, the most widely used control in the industry is ON-OFF control, so its performance was analyzed by computing the proposed evaluation function evaluated for ON-OFF control of an air-conditioning system on a railway coach. This function will characterize thermal comfort and energy consumption of the air-conditioning system. This ON-OFF control can be taken as a reference for designing better control strategies, thereby leading to improved performance indices.
The main contributions of this paper are as follows:
1. A MIMO state space based dynamic model is proposed for a full air-conditioning system, i.e., the refrigeration circuit as well as the air circuit.
2. The proposed state space model includes the dynamic effect of moisture condensation in air, which has not previously been reported in the literature.
3. The proposed model was validated through simulation and experimental studies. The air-conditioning unit of a railway coach was taken as a case study for simulation and experimental validation.
4. A...
|
