Today, there is a strong demand from the industry for cost-effective but realistic semi-quantitative simulation test procedures that can rapidly screen potential lubricants, materials and coatings before expensive and time-consuming component or field testing system-level are performed [3,27,28]. Combination of experimental simulation and computational modelling can be a good alternative to the real-life test because it allows learning about the model more efficiently, finding out quickly which factors are unimportant, building simple proxy models of complicated simulation models and then optimizing the system performance.

The present work is focused on coupling the experimental simula- tion of a ring-on-liner contact at the TDC with computational model- ling of wear and friction as function of various factors that are crucial for tribological performance: (i) type of the lubricating oil, (ii) lubrica- tion starvation/plenty, (iii) type of the surface finishing: fine or plateau honing; and (iv) surface coatings. Considering that ceramic and carbon PVD coatings showed good results in cylinder-on-plane and real engine experiments, in this work two different CrN PVD coatings as well as self-lubricious diamond-like carbon and MoS2 coatings have been studied under simulated experimental conditions at TDC. This selec- tion is underpinned by the author's previous investigations into tribological behaviour of the coated piston rings [11] that determined the coatings with desirable combination of wear resistance, friction reduction, fuel consumption and  durability.

For computational modelling a specific AVL Excite-Power Unit model that linked advanced lubricated surface contact and wear models and contained all the information related to the studied contact, i.e. geometry, profiles, roughness, material and lubricant properties, etc. was used. The obtained volume of systematic experimental data was correlated with the results of modelling for the sake of validation. In addition, the sensitivity of the model regarding various parameters such as coatings/surface treatments, oil properties/additives, surface finishing’s or texturing was studied. Validation of the computational model and its fine-tuning has been an important step towards devel- opment of a numerical tool for simulations of piston group    wear.

2. Experimental study of a piston ring/cylinder liner friction and wear

2.1. Experimental set-up

Reciprocating linear motion tribometer schematically  shown  in Fig. 1 was used in two experiments: (1) standard ball-on-flat test according to DIN 51834-2 for comparison of lubricating properties of two chosen oils and (2) simulated ring-on-liner   test.

A standard ball-on-flat test was done following procedure DIN 51834-2 in order to compare friction and wear for the two fully formulated lubricating oils used in the ring-to-liner tests: 5W30 and 15W40. The experimental conditions of this test were: reciprocating frequency 50 Hz, 1 mm stroke, normal load on the ball 300 N, temperature of the of the oil  180 °C, test  duration 2 h. The ball  and the disc materials are of 100Cr6 bearing steel. Ball-on-flat configura- tion and these conditions and are only related to the results of Fig. 4, while all the rest of the friction tests are performed according to the ring-to-liner configuration and the conditions mentioned   above.

Fig. 1. Schematic view of the experimental test rig: 1 – the fixture; 2 – the receiving block; 3 – the piezoelectric measuring device; 4 – the sample holder; 5 – the electrical resistive heater; 6 – resistive thermometer; 7 – the oscillating drive rod; 8 – the counterpart holder; 9 – the loading rod; 10 – the sample; 11 – the counterpart (reprinted from [12] with permission).

Simulated ring-on-liner test was employed for simulation of TDC friction conditions typical for a two-stroke turbo diesel engine with power 45 kW at 4000 rpm and volume 730 cc. This engine was chosen as reference for experimental simulation and computer modelling, although the results can be applied to other similar engines. The lambda ratio, i.e. half stroke/conrod spacing, of the engine was 0.3. Segments of a real piston and a cylinder liner of the specified engine were cut and used as counterparts in the experimental   simulation.