In this thesis the application of resonant MEMS (micro-electromechanical Systems) cantilevered sensors for condition monitoring of lubricating oil is studied. This work is focused on the real-time detection of particulate matter suspended in the lubricating oil, an important enabler for inline monitoring of lubricant condition for, e.g., predictive maintenance of critical infrastructure such as gas turbine and other heavy industrial machinery. During the course of this thesis two novel sensors for the detection of ferrous wear particles and oxidation-related varnish particles in lubricating oil were developed, fabricated and tested under lab conditions. The novel sensors utilize the resonance characteristics of the roof tile-shaped eigenmodes found in MEMS cantilevers and the unique physical properties of the particles to attract the particles of interest to the sensor where they adhere to the cantilever surface. The adhered particles change the total mass of the oscillating cantilever leading to a measurable change of the cantilever’s resonant characteristics such as the mechanical resonance frequency and Q factor. For the detection of ferrous wear particles, a MEMS cantilever with an integrated planar electromagnetic coil for installation in the oil circuit of lubricated machinery was developed and fabricated. The magnetic field of the coil attracts and accumulates suspended ferrous particles on the surface of the cantilever. The proof of concept was successfully demonstrated by detecting ferrous particles with a hydrodynamic diameter of 250 nm dispersed in deionized water with a concentration of 350 ppm. The detection range spans three magnitudes in particle size, covering the size range needed for detection of abnormal wear in lubricated machinery. The second MEMS sensor developed in this thesis utilizes the polar properties of oxidation-related varnish particles by generating an electric field using a tailored interdigitated electrode structure integrated on the cantilever surface. The varnish particles are accumulated and can be detected by monitoring the resonance characteristics of the microcantilever. The sensor was exposed to artificially aged commercial turbine oil at various stages of oil deterioration. The sensor response in terms of resonance frequency and Q factor was compared to the results of conventional oil analysis techniques and showed a clear correlation with the varnish condition of the oil. To demonstrate the applicability of the MEMS sensor in an industrial environment for real-time monitoring of the oil condition, two test rigs were developed in which the sensor continuously monitored the oil quality. Again, the MEMS sensor performed as expected and the proof of concept was successfully demonstrated.
