Experimental characterisation of low-cost regenerators for travelling-wave thermoacoustic devices

This paper presents the construction and testing of three low-cost regenerator designs for travelling wave engines. These include: (i) a regenerator made out of a ceramic substrate of an automotive catalytic converter, with regular square pores, (ii) a steel "scourers" regenerator and (iii) a stainless steel "wool" regenerator. The latter two are examples of materials with random geometries; they are uniformly pressed into a purpose built thin-walled "can" with carefully selected pressure to obtain the required equivalent hydraulic radius. For benchmarking purposes regenerators made out of multiple layers of wire mesh screens are also used as the most representative configuration currently used in travelling-wave thermoacoustic systems. To attempt meaningful comparisons, the materials are selected to ensure comparable hydraulic radii. The baseline set of regenerators was designed around the hydraulic radius of 200 μm. This included the ceramic substrate, steel "scourers", stainless steel "wool" and stacked wire screens (as a benchmark). This set was extended to include steel "scourers" and stacked wire screens (as a benchmark) with hydraulic radius of 120 μm. This resulted in a total of six regenerators. They were initially tested in a steady air flow in order to estimate their pressure drop due to the viscous dissipation. Subsequently, they were installed in a looped-tube travelling-wave thermoacoustic engine to test their relative performance. Testing included the onset temperature difference, i.e. the minimum temperature difference between two regenerator ends required to start the acoustic oscillation, the maximum pressure amplitude and the acoustic power output as a function of mean pressure between 0 and 10 bar above atmospheric. The experimental results indicate that the performance of regenerators made of random materials (scourers and steel wool) is much worse compared to their mesh-screen counterparts of the same hydraulic radius. One of the practical problems is that the fine wire size of the random materials makes it impossible to match the porosity levels of mesh screens when the hydraulic radius is matched. Detailed discussions and performance comparisons are provided.