A PROCEDURE TO ANALYZE AIRFLOW IN AUTOMOTIVE AIR-CONDITIONING SYSTEMS IN AN EARLY STAGE OF DESIGN

 

Jörn Hofhaus, BMW AG, Climate Control Concepts, Simulation, 80788 Munich, Germany

Hooshang Didandeh, Technalysis Inc., 5856 W. 74th Street, Indianapolis, Indiana 46278, USA

 

Abstract

A procedure for a computer-aided analysis of the airflow in the entire automotive air-conditioning system is introduced. The system is modeled as a one-dimensional network of flow resistances. We analyzed four typical operating modes of the air-conditioning system at different flow rates. Only rough CAD data were necessary to build the models. Results are obtained in terms of resistance curves, operating points, flow distributions at the outlets, and local pressure losses. Once the models are set up, sensitivity analysis is used to examine critical components or design modifications very rapidly.

 

Keywords

Air-Conditioning Systems, Flow-Simulation, Network Analysis, Sensitivity Analysis

 

1 Introduction

In recent years the requirements to reduce the costs in the design process in automotive engineering have resulted in two major tasks. On the one hand, the overall time to develop a new car has been reduced by approximately 30 percent. On the other hand, fewer prototypes and less testing hardware are available in the comparatively late stage of the design. Yet, in order to cover all functions of the car at early stage, numerous computer models and programs are applied to simulate nearly all characteristics of a new car, including the air-conditioning system.

A common approach to simulate flow characteristics inside the air-conditioning system is the use of three-dimensional analysis tools by solving the full Navier-Stokes equations to obtain a detailed description of the flow field (1-3). The results are fairly accurate, however generating them is time consuming as:

• A complete and closed description of the geometry’s surface in CAD is necessary. This is not always available. In general it is necessary to modify the CAD data for the special requirements of the 3-D analysis.
• A computational mesh has to be generated, which is still an extensive amount of work, especially for complex geometries.
• The iterative solution of the 3-D conservation equations for turbulent flows requires hours and sometimes weeks of computing time on high performance computers.

This expense grows very rapidly with the size of the problem. For analyzing a 3-D flow in a single duct, the time to obtain the results is relatively acceptable. However, in predicting the overall performance of an air-conditioning system or the flow distribution at the several outlets in the passenger compartment, the time to build and compute the model is critical. Furthermore, the flexibility for a quick comparison between different design variants is, in general, not an easy task to achieve.

In the early stage of a design, frequent changes to the package are encountered and very little CAD data is available. Often, only 2-D geometry information is available. At this point, the accuracy of a 3-D analysis is generally not necessary. Only a rough approximation of the new design or qualitative comparisons between different design alternatives are needed. In this study, a procedure is introduced to analyze the flow in the air-conditioning system with the help of a one-dimensional network of flow resistances. The resulting pressure drop for a given flow rate and the flow distribution in the various branches of the network is the primary output of the program. Together with the performance curve of the blower, the operating point for the system can be determined.

Although methods for flow network analysis are certainly not new and the results are naturally approximate in comparison with a detailed 3-D analysis, the procedure has some decisive advantages which made it attractive to shorten the design process:

• The time for building up the model is substantially smaller.
• Detailed description of the surface geometry is not necessary. A few geometric dimensions are sufficient.
• The computational time is in the range of minutes on Workstations or PC’s.
• The effect of different component parameters such as area ratios or expansion angles are quickly evaluated. Without any CAD data, many different parameters can be investigated.
• The critical components of the entire system can be determined with the help of sensitivity analysis before a detailed 3-D simulation or testing is carried out.