Thermal investigation of exhaust manifold for a heavy commercial vehicle diesel engine

Nowadays, reducing the fuel consumption, increasing the performance and efficiency of engine are the significant customer demands for the automotive industry. Particularly emission efficiency and fuel economy are greatly dependent to design of the exhaust manifold. Exhaust manifolds are one of the most critical components on the engine assembly, which is assembled on engine cylinder head. It collects the exhaust gas from the cylinders and throws out it to the exhaust after treatment system. It has a significant effect on the performance of the internal combustion engine. Especially, exhaust emission system efficiency and fuel consumption depend on manifold design. Exhaust manifold is connected directly to the exhaust ports of the cylinders and the other end is connected to the turbine if it is a type of turbocharged engine. Each exhaust port is named runner. As the engine does not start, exhaust manifold stays at lower temperature. During the engine start up process the manifold system is exposed to high temperatures due to combustion. Metal temperatures quickly rise from room temperature to high temperatures. Very harsh and crucial thermal loads are occurred due to extreme heating under very high temperatures and cooling under low temperatures. High stresses on the exhaust manifold is caused this high change in temperature. Exhaust manifold displacement is restricted due to constructive reasons. Therefore, it cannot expand or contraction freely and it is subjected to thermo-mechanical loadings. These components must withstand severe cyclic loading that combination of thermal and mechanical throughout their life cycle. These load cycles results in a complex evolution of damage, leading to thermo-mechanical fatigue (TMF) of the material and, after a certain number of loading cycles it will inevitably be a failure of the component at the end. Working conditions at the high temperature affects the components that inside of the exhaust system and/or other part of the engine located to near the exhaust system. To place of parts which are close to hot environment may become overheated and molten components make the exhaust system unable to function. Placement of these components should be considered carefully in design phase. It is important that the calculations need to be done at the design stage. The decrease in the mechanical properties of the materials at high temperature causes the vibration characteristic to change in negative way. Excitation from the motor or real road conditions during the operating will may cause the system to resonate. If excitation frequencies will overlap of the natural frequencies of the exhaust system resonance will take place. The amplitude of oscillations of a system entering resonance will increase significantly, theoretically it will go to infinity in the absence of damping source. Such a growth of amplitudes will inevitably damage the manifold and/or other subgroup components attached to the manifold. In order to produce an optimum and efficient product at the design stage, the thermal assessment of the internal combustion diesel engines is undoubtedly one of the important works to be done. Harsh operating conditions of engine make upfront CAE studies for exhaust manifolds unavoidable. CAE is an effective tool utilized to drive an exhaust manifold design by early assessment within the program development phase, requiring the advanced CAE methodologies. Engine dynamometer test running durations are up to 2500 hours for the heavy duty (HD) exhaust manifold, therefore they are very expensive and time consuming. Either long test time or extra cost effect of physical test makes the CAE studies more and more valuable. To come up with a robust and durable exhaust manifold design, necessitate CAE studies for thermal behavior of exhaust manifold is necessary to predict transient and steady regim of temperature distribution on exhaust manifold accurately and precisely. Then the predicted metal temperature is used to carry out the thermo-mechanical analysis. Critical areas are determined with the CAE at initial design stage and necessary action are taken immediately. Although you have a strong CAE approach even so at the end of the development process, for the design verification purposes, various tests have to be carried out in engine dynamometer cells under crucial operating conditions. In this study, a number of numerical and experimental studies were carried out to determine the metal temperature of the exhaust manifold at the critical operating point which is maximum power at 1800 rpm of the diesel engine of a heavy commercial vehicle. In the first step of experimental studies, nominal test data (temperature, pressure, etc.) were collected for different steady state operating conditions of the diesel engine. Then the motor was operated in the rated power for 10 minutes. Temperature measurements were made with thermal cameras and thermocouples on the exhaust manifold metal skin. Numerical analyses can be collect in three main step during analysis process. As a first step, engine performance analysis were performed in one dimension. Thus, steady state full load measurements from 900 revolution per minute (rpm) to 2000 RPM with 100 rpm increments were taken in the dynamometer test bench. Then, steady state full load correlation studies was performed with engine dynamometer test data. One dimensional engine model of the heavy-duty diesel engine was created and parameters were tuned according to dynamometer test measurement. The averaged values from the measurements were obtained and plus/minus tolerance bands were created for taking the measurement errors (for example due to systematic or random error) into consideration. Then predicted results from the model were plotted on these bands for comparison. Comparison of analysis main engine performance outputs vs. test data is shown brake torque and brake power prediction of the model was within expected correlation margin (2%). Prediction of the peak firing pressure (PFP) in the cylinders of the model was well aligned with the test data at entire engine operating range. Engine air mass flow rate and fuel mass flow rate predictions were also within expected 3% correlation margin. Similar to flow rate prediction, Boost temperature and pressure predictions of the model were also found to the middle error band of correlation margin. Air Fuel Ratio (AFR) was well predicted in CAE. Once the engine performance simulations were correlated against steady state full load dynamometer data then crank angle dependent boundary conditions at 1800 rpm were provided for the 3D CFD co-simulation analysis. Total temperatures and exhaust mass flow rates from each exhaust ports were supplied as inlet boundary conditions while turbine inlet pressure and temperature were supplied as outlet boundary conditions. In the second step, 3D computational fluid mechanics analyses were made by finite volume method. Using co-simulation approach in StarCCM+, simulation was divided into two different domains. Solid and fluid domains were modelled separately. Specified the fluid domain as leading analysis and solid domain as lagging analysis. Maximum power operating condition was simulated with the aim of capturing cyclic time averaged heat transfer coefficient (HTCGAS_Mean) and averaged gas temperature (TGAS_Mean). In this purpose it was possible to account different time scale for each domain. The fluid domain was modelled unsteady whereas solid domain was modelled steady state. Thermal field data is exchanged across the fluid and solid interface. In this technique, time intervals of data exchange on interface are determined by user specified time step. Each simulation waits for the other one until to defined exchange step is over. Co-simulation analysis process was terminated end of the sixth engine cycle. One million total cell approximately was counted for co-simulation model. Solid domain contains of 650k poly mesh. Moreover, the fluid domain occurs in 250k trim mesh. Standard k-ϵ approach is used with high y+ wall treatment for turbulence modelling. Unsteady domain utilized the Reynold-averaged Navier-Stokes (RANS) coupled flow and energy solver with the consideration that the exhaust gas treated as an ideal gas. Energy equation for the solid domains was solved using segregated energy solver. Thermal resistance is not considered in the solid domain during heat transfer at the interfaces. Radiation effect also is ignored during Co-Simulation analysis. The third step was the optimization algorithm with Optimate+ SHERPA algorithm. In the co-simulation analysis external heat transfer coefficient and reference sink temperature for the outer surface of the exhaust manifold modelled as constant since the external air flow was not modelled. In reality it is necessary to apply non-uniform heat transfer coefficients depending on various external flow velocities. For this purpose, Multi-objective based optimization study was applied in order to determine the non-uniform heat transfer distribution without performing external aerodynamics analysis, optimization gives an opportunity to obtain accurate heat transfer distribution. Outer surface of the exhaust manifold was divided into 88 patches and the external heat transfer coefficient for each part was defined as optimization variable to the patches. The main objective of the optimization was determined the proper patches' heat transfer coefficient configuration in order to has good alignment with dynamometer test temperature which is measured thermocouple and thermal camera experiment results. Comparing exhaust manifold metal skin temperature on the control points between CFD and dynamometer test, maximum prediction error is found + %37 in Co-Simulation. Enhanced model which is optimized HTC distribution case improvements were achieved in terms of temperature at the control points. While there is no change in the temperature results of TC-4, successful results were obtained particularly SP03, TC-1 and TC-2 control points. Maximum error prediction is found - %9. The results were improved for each control point except one control point. Consequently, the temperature distribution of exhaust manifold is calculated correctly with acceptable error. In this way, it is possible to estimate strength of the exhaust manifold safely with the correct temperature information. Thus, structural weaknesses are avoided at early design stages and optimum product is obtained with minimum cost.