Thermal comfort is an important concern for occupants in an enclosed environment such as the passenger compartment of a vehicle. Recent design trends, such as the increased use of glass in vehicle styling, tightening fuel economy constraints, and the move to environmentally safe refrigerants have made it more difficult to achieve occupant comfort. While the capacity of the HVAC system could simply be increased to respond to these trends, doing so would increase both the cost and fuel consumption of the vehicle. In order to avoid these negative repercussions, the task of optimizing the design of the HVAC system has become increasingly more important. In the past, HVAC systems were designed by using a series of physical tests run in dissipaters, wind tunnels, and proving grounds to evaluate the performance of proposed designs. These experiments were very expensive – it typically costs near the order of $500,000 to build a single vehicle prototype and about $400 per hour to operate a wind tunnel. It’s also important to note that, in the vehicle environment, many parameters, such as solar load for various solar incidence angles, glass properties, air temperature, and air velocity magnitude, are dependent on each other. Furthermore, their complex relationship is not known exactly. This makes it nearly impossible to perform a parametric experimental study.
Simulating passenger comfort
In an effort to help achieve improved thermal comfort while reducing development time and cost, Delphi engineers have worked with the University of California, Berkeley to develop the capability of predicting occupant thermal control to support the design of automotive climate control systems. First, the climactic conditions in which the vehicle is operating are used to develop a thermal model of the passenger compartment. The results are passed as input to a separate model of the human thermal regulatory system to predict human comfort. The resulting virtual thermal comfort engineering model makes it possible to explore different climate control strategies as they relate to human comfort in a relatively quick and inexpensive manner. This model is now being used to develop an actual vehicle climate control system at Delphi. The long lead times involved in HVAC system design make it impossible to estimate the cost savings involved by Delphi engineers but it can already be stated with assurance that the design will be optimized to a significantly higher level than was possible in the past, and that engineering costs will be significantly reduced.
The modeling process begins with Delphi Vehicle Interior Model (DVIM) that is developed by computer aided design (CAD) using Unigraphics (UG). Using this model, Delphi engineers simulated solar loading on the vehicle using custom modifications in FLUENT, CFD software from Fluent Incorporated, Lebanon, New Hampshire. A two-step process was used for this task. First, based on the geometrical location (such as Phoenix, Arizona) and the date and time (such as July 15 at 12 pm), the program automatically calculates the maximum solar radiation. Then, for a given cabin and glass geometry, the program calculates how the solar load affects the passenger compartment. The software keeps track of the reflection from the glass, absorption by the glass, transmission into the cabin and also the incident radiation on the occupants of the cabin. Radiation heat load is calculated using a linearized model based on mean radiant temperature specified for each body segment or an explicit model using the Stefan-Boltzmann law.
Performing the calculation
Next, a CFD mesh is created for the passenger compartment using a solid model provided by the automotive OEM. The model is imported directly into GAMBIT, the FLUENT preprocessor, which quickly prepares a clean surface geometry and generates a 3D mesh. The Reynolds-averaged Navier-Stokes equations for fluid flow are solved on this mesh along with conservation equations for energy and moisture concentration. Together, these equations predict the airflow, temperature, and humidity distribution around the occupants.
During the course of the solution, the CFD calculations are coupled to the Berkeley physiological model. Air temperature, humidity, and velocity information from the CFD code are transferred to the Berkeley calculation of occupant surface temperatures on the sixteen body zones. These surface temperatures are updated and transferred back to the CFD calculation, where they are used as revised boundary conditions for the next round of CFD calculations. The Berkeley code is based on a model developed by J. Stolwijk, and on work done by S. Tanabe. The Stolwijk model is based on six body segments: head, torso, arms, hands, legs and feet. The current Berkeley model can simulate an arbitrary number of segments. Each of these segments consists of four body layers – core, muscle, fat and skin tissues – and a clothing layer. A separate series of nodes represents arteries and veins and provides for convective heat transfer between segments and tissue nodes, and the countercurrent heat exchange between the arteries and the veins. The model computes heat transfer between each node using a standard finite-differencing algorithm with variable time-stepping to optimize computational resources while preserving numerical stability.
An initial validation of the model was performed using a test rig designed to simulate a vehicle passenger compartment. A pyranometer was provided to deliver solar-like radiation while a 2.5 liter engine and air conditioning system cooled the compartment. Thermocouples were used to measure air and surface temperatures. The temperatures measured in these experiments closely matched those predicted by FLUENT. The success of the calculations highlighted an important advantage of numerical analysis: that it is able to predict temperature and airflow at any point in the cabin while the data that can be obtained from physical experiments is limited by the number of sensors that can placed within the vehicle.
The physiology and comfort prediction of the model were validated using a thermal mannequin that was also developed at the University of California, Berkeley. In these experiments, air temperature and velocity around 16 body segments, interior surface temperatures, breath level temperatures, air conditioning outlet temperatures and humidity, solar load, and mannequin body heat loss were all measured. The simulated temperatures matched the measured temperatures to within a few degrees for each body segment. The measured and simulated thermal comfort index also showed a near-perfect match. This comfort index is based on the equivalent homogenous temperature developed by M. Bohm that is designed to measure the human sense of thermal comfort.
Based on these successful validation studies, Delphi has begun using these methods in the development of new HVAC systems. The most important advantage of this approach is that it allows engineers to evaluate new design alternatives in a fraction of the time that was required in the past. All they have to do is change the parameters of the CFD model and re-run the analysis in order to determine with a high level of accuracy the effects of the design change they are considering. In the past, evaluating such a change would have required weeks of hardware modifications and physical testing. Using numerical simulation, engineers can now evaluate far more alternatives and determine how proposed designs perform (in terms of meeting critical occupant comfort requirements) even before building physical prototypes. Both the development cost and development time of the HVAC system can be substantially reduced as well. Fewer physical prototypes and experiments are needed because engineers can now eliminate ineffective designs using simulation so that only a few - or possibly even one - highly optimized design need to be tested. In the future, virtual thermal comfort engineering is expected to play a major role in the exploration of highly sophisticated adaptive control devices and algorithms. These advances will make it possible to tune the HVAC system to more closely meet the passengers’ comfort requirements.
Driver skin temperatures during soak and cool down.
CFD & thermal simulation integrated with transient AC and solar load for cabin airflow and thermal analysis.
|
]
