Investigation of Gas Blow-by Through Turbocharger Sealing System Using Advanced Computational Simulations ^†

Abstract

Gas blow-by flow through a turbocharger sealing system has a significant effect on lubricant consumption and the formation of particles. This paper presents an experimentally validated computational model to quantify the mass flow of gasses through a sealing system as a function of operating conditions. The computational model is applied to a truck engine turbocharger and the results are verified experimentally. The results show that the gas blow-by is not only influenced by the geometrical parameters of the lubrication system but is controlled by the thermo-mechanical loading of the turbocharger, with a non-negligible influence of the radial dynamics of the rotor.

1. Introduction

The problem of negative emissions from internal combustion engines (ICEs) in the transport sector is serious and needs to be addressed. In addition to reducing fuel consumption and CO₂ emissions, it is important to address the production of particulate matters, which are produced by both diesel [1] and petrol ICEs [2]. Turbochargers (TCs) are used to increase power and reduce ICE fuel consumption [3], but they also cause gas leakages, known as gas blow-by (GBB), which can contaminate the lubricant located in the central housing. The contamination of the lubricant by gasses can lead to speed up wear and the loss of lubricant properties. These gas leaks are a growing problem with the modern trend towards using TCs with increasing pressure ratios [4].

In principle, when higher pressures occur in the central housing, a lubricant can leak back into the compressor or turbine with the gas flow through the sealing systems, which is particularly noticeable at low rotor speeds and therefore low compressor pressures, often when there are problems with lubricant draining from the central housing. This phenomenon, called gas blow-back, transfers lubricant into the compressor or turbine, which is then partially burned, increasing the production of particulate matter [5] and unburned hydrocarbons [6,7]. The piston group is usually the dominant source of lubricant consumption, but lubricant consumption caused by TCs is an inherent and undesirable consequence of operating turbocharged ICEs. A TC’s lubricant consumption can be reduced by limiting the gas blow-back flow, which has a positive effect on the overall lubricant consumption and reliability of the ICE. A key element is then the TC sealing system, which uses flexible sealing rings to control GBB and gas blow-back flows.

The control of GBB by the turbocharger sealing system depends on understanding its mechanisms and predicting them under operating conditions. The aim of the research is therefore to understand these mechanisms by using computational simulations and subsequently enable targeted designs to reduce GBB. These designs have the secondary effect of reducing lubricant consumption and particle transport to the compressor or turbine, where they can be deposited and reduce efficiency or burned and increase harmful emissions.

2. Material and Methods

2.1. Computational Model for Gas Blow-by Solution

It is clear that the GBB depends not only on the pressures in the individual chambers of the TC sealing system, such as the central housing, chambers defined by grooves and rings or the compressor and turbine backplate chamber, but also on the shape of the labyrinth between these chambers.

The computational model must therefore solve a coupled problem involving structural dynamics, fluid dynamics and heat transfer. Such a model can be designed in a variety of ways, from simple (zero-dimensional) 0D approaches to complex three-dimensional (3D) computational approaches. The fact that the given physical phenomena are time-dependent is certainly a complication. The proposed computational model represents a certain balance between the physical depth and physical width of the problem and combines different levels of computational models in terms of dimensionality.

The central module of this computational model is the so-called virtual turbocharger, which was developed and verified by the authors Novotný et al. [8]. This model uses a 3D description of the turbocharger rotor and a fast hydrodynamic model of the floating ring journal bearing. A very important sub-model is the computational model of the thrust bearing, and for a solution of GBBs, the model presented in the work of Novotný et al. [9] is used.

The 3D computational model constructed in this way was extended by a 0D computational model of the gas dynamics in the labyrinths of the sealing system on the compressor side and on the turbine side. This model is based on the assumption that the gas flow through the sealing system can be with some minor simplifications considered a one-dimensional (1D) flow of a fully transient nature. In the simulations of long-time operations of a TC, the model developed in this way proves to be a sufficiently accurate description of the given problem and very computationally efficient.

The computational model is also extended by a 0D thermal model describing the heat transfer through the rotor, taking into account the interactions from central housing through the lubrication film layers of the oil film bearings and the heat transfer from the compressor and turbine.

The computational model shown graphically in Figure 1, i.e., the virtual turbocharger, was assembled in the multibody system ADAMS and extended by user-written functions solving the 3D dynamics of the seal rings, the 1D model of the gas dynamics and the mass concentrated 0D model of heat transfer in solids. The computational model contained 30 rigid bodies with 3D geometry, two 3D bodies based on reduced finite element models, four mass points, 35 constraint joints, 12 differential equations for solutions of a ADAMS system of non-existent physical phenomena, 9 motion generators, four contact forces using 3D geometry, 28 general forces, and 734 solver variables and was controlled by 416 user parameters, and the definitions of the sub-models can be found in [10].

2.2. Experimental Methods for Gas Blow-by Solution

The calculated values were verified by technical experiments on a test bench in the test cell. Basic quantities such as gas temperatures and pressures in front of and behind the impellers, oil film thickness of the thrust bearing and others were measured on the TC used for gas blow-by measurement. The TC was installed the diesel six-cylinder ICE with a power output of 260 kW under 2200 min⁻¹.

The ICE was operated in a slow sped-up regime within 30 s (600–2400 min⁻¹), and during this sped-up regime, the values of gas and oil temperatures, gas pressures and distances were measured using thermocouples, piezoresistive sensors and eddy current distance sensors. The GBB values were measured directly on the test bench for the entire TC using an orifice pressure differential device.

3. Discussion of Results

The computational model is designed to provide a wide range of different results from kinematic quantities and thermodynamic quantities to gas leakage through the sealing system. In the case of this paper, the focus is only on the evaluation of the GBB through the sealing system.

The GBB values are calculated in the time domain as a function of rotor speed. Figure 2 summarizes the GBB values calculated by the computational model, shows the compressor and turbine side components of the sealing system and compares these values with the measured values. The turbine side showed a significantly higher GBB than the compressor side, mainly due to the higher pressures in the turbine and the specific position of the seal rings within the shaft grooves. The GBBs on both sides gradually increased with increasing pressure ratios and reached very similar GBB values at high rotor speeds. At very high speeds, the compressor GBB was limited by regulation.

The results of the computational model and the technical experiment (Figure 2) show some differences, especially at very low speeds. This fact is influenced by the small value of GBB and therefore by the greater influence of other quantities that are not accurately influenced by the computational model.

4. Conclusions

Based on the analysis of the results of both calculations and experiments, it can be defined that the GBB through the TC sealing system is mainly dependent on the geometric configuration of the sealing system and the thermodynamic conditions at the turbine and compressor, respectively.

The geometry of the sealing system is determined not only by the geometry of the individual components (grooves, rings, etc.) but also by the motion of these components. The thermodynamic conditions on the rotor have a significant effect on the axial load on the rotor. The pressure conditions on the impellers cause the rotor to move and this changes the shape of the sealing labyrinth—the individual thin gaps change significantly. GBB through the thin gap between the seal rings and grooves is usually the dominant source of leakage. The axial movement of the rotor due to pressure conditions on the impellers can both reduce and increase GBB. The radial movement of the rotor also has a non-negligible effect on the GBB, which is amplified in the case of strong rotor resonances.

GBB through the ring lock is always present but is often very small due to the flow through the thin gap. The ring lock gas leakage is strongly influenced by the gas pressure gradient and gas temperature. The effect of rotor movement on this mechanism is minimal. There is one case where the influence of this mechanism is dominant, and that is when the thickness of the thin gap is limited when the ring is pressed against the groove wall.

The temperatures of the turbine and compressor also have a significant effect on the temperature distribution in the sealing system and thus on the properties of the flowing gas. Due to the very small geometric dimensions of the labyrinth, the gas almost adopts the temperature of the surrounding walls, changing its properties and resulting in a change in GBB.

The calculated and measured data show non-negligible discrepancies in GBB values, especially at low rotor speeds. These discrepancies are probably due to the imperfection of the thermal model for estimating the turbine wheel temperature due to heat transfer from the flue gas.