Search Results

An array of engine bearing analysis methods has been developed at General Motors over the years. All of these analyses consider wedge and squeeze effects, finite-length bearing, variation of load with crank angle, and cavitation effects. The simplest among them utilizes the so-called mobility method for solving the governing Reynolds equation. Others include finite-element solution for bearings with arbitrary geometry and grooving, finite-element solution for elastohydrodynamic lubrication, mass-conserving finite-volume solution, non-Newtonian lubricant analysis, and thermohydrodynamic analysis. This paper reviews these methods, describes when and how these methods are used, compares results and describes some applications.

A mathematical model for tribological analysis of different automotive- valve-train configurations has been developed as a part of the FLARE (Friction and Lubrication Analysis of Reciprocating Engines) package. The model is based on an in-depth kinematic analysis and on a rigid-body dynamic analysis, including dynamic analysis of the valve spring. Lubricant film thickness, contact pressures, and frictional power loss are predicted. A mixed-lubrication model is used to determine the friction force at the cam-follower interface. In addition, lifter rotation is modeled to predict its effect on frictional power loss. Detailed results are presented for a pushrod valve train. Also, this paper compares frictional power loss for five different valve train types. They are: direct-acting overhead cam, pushrod, end-pivoted finger follower, center-pivoted finger follower, and cam-in-head. The valve trains are made equivalent by keeping the valve lift and the no-follow speed the same.

Minimum bearing oil film thickness (MBOFT) was measured in both a main and a connecting-rod bearing of a production V-6 engine using the total capacitance method (TCM). MBOFT was measured at 1500 rpm and at three different engine loads (64, 128, and 192 Nm). The oil sump temperature was controlled at 100°C. Five engine oils were tested (SAE grades 5W-20, 20W-20, 5W-30, 10W-30, and 20W-50) with emphasis given to the SAE 5W-30 and 10W-30 oils. MBOFT was also calculated using a computer code. The absolute minimum of the MBOFT (MBOFTmin, closest approach between the journal and the bearing) increased with increasing values of a Sommerfeld parameter (viscosity/load) and this dependence was similar in both the main and the connecting-rod bearings. But, for the main bearing the MBOFTmin values showed a higher dependence on the Sommerfeld parameter than those for the connecting-rod bearing. Similar results were obtained with the theoretical calculations of the MBOFT_min values.

A new math-based piston noise model is proposed. It is based on a EHD (elastohydrodynamic) piston lubrication and secondary dynamic analysis. The model predicts the force on the bore wall as a function of time, which is then transformed to the frequency domain. The dynamic calculations of the model were validated against a closed form solution for a simple 1-D model. The model was applied to a production V8 engine. Predictions of the model compared well with block acceleration measurements for the same engine both in relative magnitude and frequencies. Effects of changing clearance and oil supply were also investigated using the model.

A comprehensive computer package, FLARE, has been developed for carrying out Friction and Lubrication Analysis of Reciprocating Engines. FLARE considers four major lubricated components in an automotive engine -- piston skirt, piston rings, bearings, and valve train. Hydrodynamic, mixed, and boundary lubrication models are used, as appropriate, to model the lubrication phenomena. All the analytical models are based on solution of governing equations. Three levels of analyses with varying degrees of detail have been developed. Availability of different levels provides the flexibility of matching the complexity and accuracy of the analysis with the objective of the analysis. An empirical engine friction model, which is based on experimental data, is also available. Many user-friendly features are built into the FLARE system to make it easier to use for design engineers. This paper gives a brief overview of all the analysis sub-models incorporated into FLARE.

Comparisons are made between friction predictions of the FLARE (Friction and Lubrication Analysis of Reciprocating Engines) computer code and experimental data for the purpose of validating FLARE. An in-line four-cylinder engine under motoring conditions was selected for doing the experiments. Three major friction producing subassemblies were considered: piston assembly, crankshaft main bearings, and valve train. A Taguchi-type L16 matrix was used for the piston assembly, while an L8 matrix was used for the valve train. A traditional approach (varying one parameter at a time) was used for crankshaft main bearings. The agreement between experimental measurements and FLARE predictions, for all the cases studied, is very good. The match is closest for the valve train, followed by the crankshaft main bearings and piston assembly. In addition, trends and effects of changing design parameters are predicted correctly by FLARE.