Peter Gullberg

Continued Study of the Error and Consistency of Fan CFD MRF Models

The most common fan model to use in commercial CFD software today is the Multiple Reference Frame (MRF) model. This is at least valid for automotive under hood applications. Within the industry, for this typical application, this model is commonly known to under predict performance. This under prediction has been documented by the authors’ of this paper in SAE paper 2009-01-0178 and VTMS paper 2009-01-3067. Furthermore has this been documented by S.Moreau from Valeo in “Numerical and Experimental Investigation of Rotor-Stator Interaction in Automotive Engine Cooling Fan Systems”, ETC, 7th European Conference on Turbomachinery, 2007. In preceding papers a specific methodology of use has been documented and it has been shown that the MRF model under predicts performance for the airflow in a cooling system commonly with 14% in volumetric flow rate. This is for a system dominated by inertial effects. These 14% was shown to apply across fans of different sizes and designs hence illustrating a consistency of the fan model. In paper 2009-01-3067 a different methodology of use was documented which decreased the under prediction to 8%. In this paper the study has continued and is now presenting a more successful methodology of use which essentially removes the under prediction of the fan MRF model for most of fan operation, taking it for axial flow conditions within reasonable capabilities of reproducing tested geometry in a simulation environment, i.e. a difference of 0-1% between measurements and simulations.

Comparative Studies between CFD and Wind Tunnel Measurements of Cooling Performance and External Aerodynamics for a Heavy Truck

Nowadays, much focus for vehicle manufacturers is directed towards improving the energy efficiency of their products. The aerodynamic drag constitutes one major part of the total driving resistance for a vehicle travelling at higher speeds. In fact, above approximately 80km/h the aerodynamic drag is the dominating resistance acting on a truck. Hence the importance of reducing this resistance is apparent. Cooling drag is one part of the total aerodynamic drag, which arises from air flowing through the heat exchangers, and the irregular under-hood area.When using Computational Fluid Dynamics (CFD) in the development process it is of great importance to ensure that the methods used are accurately capturing the physics of the flow. This paper deals with comparative studies between CFD and wind-tunnel tests.In this paper, two comparative studies are presented. One is a comparison between cooling performance simulations and chassis dynamometer measurements; the other study is a comparison between external aerodynamics simulations and wind-tunnel measurements.The purpose of this study was to evaluate and develop methods and models for determining aerodynamic drag and cooling performance, which ultimately will be used to develop more energy-efficient cooling of heavy trucks.The results from the two comparative studies showed that there was in general good agreement between CFD and wind-tunnel measurements. For the cooling performance simulations, the analysed parameters were very close to the measured values. For the external aerodynamics simulations, the results were not easy to analyse. The overall results were still satisfactory; for the simulated yaw angles, the drag coefficient (CD) values were less than 4.1% different from measured data.

CFD Method and Simulations on a Section of a Detailed Multi-Louvered Fin where the Incoming Air is Directed at 90° and 30° Relative to the Compact Heat-Exchanger

This paper presents results and a Computational Fluid Dynamics (CFD) method for simulation of a detailed louvered fin for a multi-louvered compact heat-exchanger. The airflow was angled at 90°, +30° and -30° relative to the heat-exchanger to evaluate changes in static pressure drop and airflow characteristics. The investigation was based on three heat-exchangers with thicknesses of 52mm and two of 19mm. One period of a detailed louvered fin was simulated for two airflows for each heat-exchanger. The pressure drop data was thereafter compared to experimental data from a full-size heat-exchanger. From the pressure drop and the airflow characteristic results recommendations were made that those kinds of simulations could be defined as steady state, and with the kω-SST turbulence model. For the same heat-exchanger angle the airflow within the core was similar, with a turbulent characteristic behind it. The static pressure drop was reduced significantly for the ±30° cases compared to the 90° angled heat-exchanger to approximately one third, when comparing for the same mass airflow rates. Since the test section area was defined as constant the velocity through the heat-exchanger core varied for the 90° and the 30° cases. When comparing the core velocity it was observed that there were minor losses due to the redirection of the airflow for the 30° angle compared to the 90° case. The results showed that the 30° case, where the inlet airflow was parallel to the louvers, had a higher pressure drop than the other 30° case. It was also observed that even when the inlet airflow angle varied, the outlet airflow angle from the heat-exchanger only varied 4.3-6.4°.

Cooling Airflow System Modeling in CFD Using Assumption of Stationary Flow

Today CFD is an important tool for engineers in the automotive industry who model and simulate fluid flow. For the complex field of Under hood Thermal Management, CFD has become a very important tool to engineer the cooling airflow process in the engine bay of vehicles. To model the cooling airflow process accurately in CFD, it is of utmost importance to model all components in the cooling airflow path accurately. These components are the heat exchangers, fan and engine bay blockage effect. This paper presents CFD simulations together with correlating measurements of a cooling airflow system placed in a test rig. The system contains a heavy duty truck louvered fin radiator core, fan shroud, fan ring and fan. Behind the cooling module and fan a 1D engine silhouette is placed to mimic the blockage done by a truck engine behind the fan. Furthermore a simple hood is mounted over the module to mimic the air guiding done by the hood in an engine bay. The measurements monitor pressure and flow over the system. Supporting this examination is a set of 48 velocity probes in the radiator that measures the local velocity. The simulations of this system are correlated to measurements. Furthermore to support these simulations, specific simulations and measurements are conducted using the radiator only and the fan only. This is done to see how well each separate component is predicted in CFD and correlated back to measurements. This work is the continuation of work presented in [5] and identified in this paper is that one can simulate the cooling airflow system rather well with steady state CFD. However, fan modeling is sensitive and specific care has to be taken in order for these simulations to be accurate.

A Correction Method for Stationary Fan CFD MRF Models

A common fan model to use in automotive under hood simulations is the Multiple Reference Frame (MRF) model and within the industry, for this specific application, this model is well known to under predict performance. In this paper we have examined the possibilities of correcting this deficiency with a simple ?speed correction?. This is done by testing and simulating a production fan in the Volvo Fan Test Rig for two operating speeds, 1200 rpm and 2400 rpm. Pressure rise, fan power and static efficiency are presented as functions of volumetric flow rate. The simulations verify that using the MRF model the common behavior of under predicting pressure rise and performance of the fan occur. In addition, this work shows that; although the MRF is not predicting fan performance correctly it constitutes a reliable fan modeling strategy. In fact, the classical fan laws apply to the MRF model, and this justifies the suggested procedure of simply ?speed correcting? the fan and by doing so a consistent behavior is received. For a standard 750 mm production fan, it is shown that by increasing the fan speed with 14%, throughout the range of operating speeds, one accurately models the pressure rise over mass flow for most of vehicle operating conditions.

CFD simulation and experimental investigation of pressure-drop through 90° and 30° angled compact heat-exchangers relative to the oncoming airflow

This paper presents pressure-drop and airflow characteristics for compact heat-exchangers, where the relative airflow is angled 90° and 30°. The investigation is based on two heat-exchangers with different thicknesses, investigated for a number of airflow rates. The results are obtained from experiments and CFD simulations, where both a part of a detailed heat-exchanger and the complete test set-up have been simulated. The results showed that the thin heat-exchanger at 30° gave 70% of the pressure-drop obtained for the 90° angle, and at the same time resulted in a higher heat-transfer rate.

CFD Simulations of one Period of a Louvered Fin where the Airflow is Inclined Relative to the Heat Exchanger

This article presents Computational Fluid Dynamics (CFD) simulations of one period of a louvered fin, for a crossflow compact finned heat exchanger, where the incoming airflow was inclined relative to its core. Four inclinations were investigated: 90°, which was when the air flowed perpendicular to the heat exchanger, 60°, 30° and 10° angles relative to the vertical plane. The study included three heat exchanger designs, where two of them had symmetrical louvered fins and a thickness of 19mm and 52mm. The third had a thickness of 19mm and had the louvers angled in one direction. All heat exchangers have been simulated when the airflow entered both from above and below relative to the horizontal plane. Simulations have also been carried out when the airflow entered from the side, illustrating the heat exchanger to be angled relative to the vertical axis. Two air speeds have been investigated for each configuration, where the results were compared to experimental data. The results showed that the airflow characteristics were strongly dependent on the inclination angle. A more inclined heat exchanger generated larger separated areas at the entrance of the heat exchanger core. After approximately six louvers the airflow was fairly similar for all inclinations. The pressure drop over the core was not affected by the direction of the entered airflow. It was also seen that the heat exchanger with the louvers in only one direction resulted in approximately the same pressure drop as the one with symmetrical louvered fins. For a constant velocity in the longitudinal direction within the core, the pressure drop did not varied between the inclinations. Depending on the definition of the heat exchanger arrangement the heat transfer rate was affected. Simulated pressure drop followed the same trend as experimental data, even though the values were over-predicted.

Experimental Investigation of Heat Transfer Rate and Pressure Drop through Angled Compact Heat Exchangers Relative to the Incoming Airflow

The investigation showed that a more inclined heat exchanger resulted in lower static pressure drop and at the same time achieved a higher heat transfer rate, for a specific mass airflow rate. This result was obtained for all three heat exchangers. When analysing the parameters at the same core speed it was seen that the static pressure drop was increased for the 10° and the 30° angled heat exchangers, compared to the 90° configuration. For the 60° cases the pressure drop was both increased and decreased compared to the 90° cases, depending on the heat exchanger design. It was also seen that the pressure drop and the heat transfer rate variation were negligible between the downflow and crossflow orientation of the heat exchanger. When defining the static pressure drop to 200Pa either a 19mm thick heat exchanger at 60° or a 52mm heat exchanger at 90° can be used to obtain the same heat transfer rate. This paper presents pressure drops and heat transfer rates for compact heat exchangers, where the heat exchangers are angled 90°, 60°, 30° and 10° relative to the incoming airflow. The investigation is based on three heat exchangers with thicknesses of 19mm and 52mm. Each heat exchanger was mounted in a duct, where it was tested for thermal and isothermal conditions. The inlet temperature of the coolant was defined to two temperatures; ambient temperature and 90°C. For the ambient cases the coolant had the same temperature as the surrounding air, these tests were performed for five airflow rates. When the coolant had a temperature of 90°C a combination of five coolant flow rates and five airflow rates were tested. The test set-up was defined as having a constant cross-section area for 90°, 60° and 30° angles, resulting in a larger core area and a lower airspeed through the core, for a more inclined heat exchanger.

Numerical Investigation of Natural Convection in a Simplified Engine Bay

Presented are results from numerical investigations of buoyancy driven flow in a simplified representation of an engine bay. A main motivation for this study is the necessity for a valid correlation of results from numerical methods and procedures with physical measurements in order to evaluate the accuracy and feasibility of the available numerical tools for prediction of natural convection. This analysis is based on previously performed PIV and temperature measurements in a controlled physical setup, which reproduced thermal soak conditions in the engine compartment as they occur for a vehicle parked in a quiescent ambient after sustaining high thermal loads. Thermal soak is an important phenomenon in the engine bay primarily driven by natural convection and radiation after there had been a high power demand on the engine. With the cooling fan turned off and in quiescent environment, buoyancy driven convection and radiation are the dominating modes of heat transfer. The unsteady and turbulent nature of this complex phenomenon requires high spatial and temporal resolutions and an effective computational strategy. A CFD procedure for modeling buoyancy driven flow in vehicle underhood is demonstrated. Computed temperature and velocity of air under the enclosure are compared with experimental data at a number of different locations in the control volume. The numerical results exhibit satisfactory consistency with measured values.

A 1D Method for Transient Simulations of Cooling Systems with Non-Uniform Temperature and Flow Boundaries Extracted from a 3D CFD Solution

The current work investigates a method in 1D modeling of cooling systems including discretized cooling package with non-uniform boundary conditions. In a stacked cooling package the heat transfer through each heat exchanger depends on the mass flows and temperature fields. These are a result of complex three-dimensional phenomena, which take place in the under-hood and are highly non-uniform. A typical approach in 1D simulations is to assume these to be uniform, which reduces the authenticity of the simulation and calls for additional calibrations, normally done with input from test measurements. The presented work employs 3D CFD simulations of complete vehicle in STAR-CCM+ to perform a comprehensive study of mass-flow and thermal distribution over the inlet of the cooling package of a Volvo FM commercial vehicle in several steady-state operating points. The results from these are correlated with test readings and are imposed on a 1D model of the cooling stack with inlet discretization, which features non-uniform boundary conditions. The 1D model is tested in steady state and transient conditions. Results are correlated with readings from dynamometer tests. No major indications were present to support that the non-uniform approach improves accuracy of simulation. Nevertheless, the results show, that the suggested predictive method successfully captures the thermal effects of recirculation while reducing the necessity for calibrations done by prototype testing.