Untitled

Abstract

The equations of motion for the major components in an internal combustion engine are developed herein using a recursive formulation. These components include the (rigid) engine block, pistons, connecting rods, (flexible) crankshaft, balance shafts, main bearings, and engine mounts. Relative coordinates are employed that automatically satisfy all constraints and therefore lead to the minimum set of ordinary differential equations of motion. The derivation of the equations of motion is automated through the use of computer algebra as the precursor to automatically generating the computational (C or Fortran) subroutines for numerical integration. The entire automated procedure forms the basis for an engine modeling template that may be used to support the up-front design of engines for noise and vibration targets. This procedure is demonstrated on an example engine under free (idealized) and firing conditions and the predicted engine responses are compared with results from an ADAMS model. Results obtained by using different bearing models, including linear, nonlinear, and hydrodynamic bearing models, are discussed in detail.

Analysis of rigid-flexible coupled torsional vibration and vibration suppression of crankshaft system under multiple working conditions

“…Qin et al 15 introduced the combination of multi-body dynamics and finite element method in the study of crankshaft system torsional vibration response in China, providing insights into considering the inertia of shaft components and other influencing factors in the study of crankshaft torsional vibration. Subsequently, more scholars have conducted analyses on crankshaft modes, 16 stress, [17][18][19][20] and the influence of bearings. Kushwaha M. 21 utilized the multi-body dynamics method to analyze the impact of engine shafting component flexibility on noise and vibration response, verifying the accuracy of the method through experiments, but the support bearings are considered to be a simple thin-shell bearing.…”

Incremental and Decremental SVM for Regression

“…The adiabatic incremental/decremental training algorithm for SVMs (and SVRs) was introduced in [13,14] and it follows from a method proposed by Cauwenberghs and Poggio [3,5]. For each pattern being part of the solution (called support vector ), the associated threshold value is θ i ∈ {− , }.…”

“…It is based on the adiabatic training algorithm. Compared to [13,14], the IDSVMR contains all aspects needed for an implementation: a) We provide the relations that give the exact amount needed for increasing the threshold θ c for the new vector, before any vector migrates; b) We determine the exact threshold variation that leads from one migration to the next. This produces a robust incremental algorithm, that creates a sequence of vector migrations, as part of the learning process; c) We deduct the expressions for the variation of the threshold θ c , from one migration to the next.…”

“…These algorithms solve the quadratic optimization problem using a decomposition method based on Sequential Minimum Optimization [4], or on an incremental learning method. The incremental method considers that input patterns are added to the solution (or removed from the solution), one at a time, without affecting the learning process of the other patterns [13,14].…”

A Novel Coaxial Balance Mechanism for Reciprocating Piston Engines

“…In terms of the engine vibration simulation method, research shows that the use of a flexible body model has a greater impact on the simulation accuracy of the whole engine [26][27][28][29][30]. Because piston engines involve multiple components, it is generally necessary to establish a multi-body dynamic model of the whole engine.…”