Conference Year
2018
Keywords
Tube-fin Heat Exchanger, GA, Robust Circuitry Optimization, Airflow Maldistribution
Abstract
Tube-fin heat exchangers(HXs) are widely used in the HVAC&R industry. Studies have proved that by optimizing the refrigerant circuitry, heat exchanger performance can be significantly improved. Since air-to-refrigerant heat exchangers are typically confined in packaged units along with a fan, the airflow distribution on the face of the HXs is a dominant factor influencing its performance. During the operation of a heat exchanger as a part of the system, the air flow distribution changes continuously, especially as the fan speed changes during startup and shutdown cycles. This poses a design challenge as typically heat exchangers are designed using the assumption of uniform flow or for a single known flow distribution profile. For each profile and for the same flow rate, a typical circuitry optimization algorithm can generate a completely different optimal circuitry. Therefore, robust circuitry design that can always guarantee an acceptable minimum performance under various airflow distributions is required. In the field of optimization, this is referred to as robust optimization. This paper presents a robust circuitry design optimization approach. The formulation consists of an upper-level optimization problem and a lower-level finite search problem. In the lower-level problem, a finite number of typical airflow distribution profiles are imposed. These profiles are obtained from the literature, experimental measurements, and CFD simulations. The goal of the lower-level finite search problem is to obtain the worst case capacity degradation from different air flow profiles for a given circuitry. The objective of the upper-level problem is to obtain the circuitry that maximizes the worst case capacity subject to a set of operating constraints such as pressure drops and subcooling/superheat. In order to effectively obtain the optimal designs and guarantee manufacturable designs, an integer permutation based genetic algorithm (IPGA) developed in previous research is used to solve the upper-level problem. The optimized circuitry is then verified by using exhaustive search. The comparison between the solutions of the proposed approach and the optimal circuitries obtained under uniform airflow distribution shows that despite a 1.4%-3.7% decrease in capacity, robust circuitry designs are more resilient to multiple airflow maldistribution profiles. The proposed approach is applied to an A-type indoor unit which demonstrates its applicability in real-world design.