Microanemometry: Mo(o)re or Less

Microanemometry Mo(o)re or Less refers to the measurement of wind using microfabricated sensors. In this work these measurements include not only the speed of the wind, but also parameters such as the direction and the humidity of the air. The calorimetric anemometers in this work rely on heat transfer from a heater to sensing elements to measure wind speed, direction and humidity. These parameters all affect the thermal landscape around the sensor in cross-correlated ways, making it difficult to describe them in one complete physical model.

This work covers two types of sensors: MEMS-based (Micro Electro Mechanical Systems) anemometers and CMOS-based (Complementary Metal Oxide Semiconductor) anemometers. The MEMS process allows for material and geometric freedom in sensor design, while the CMOS process allows for industrial scale fabrication with almost perfect yield and integrated electronics. Both sensors were used as calorimetric anemometers and are able to measure wind in the 0.01 m s−1 to 9.7 m s−1 range with low power actuation (20-150 mW).

To interpret the complex sensing outputs, multi-parameter measurement techniques were applied. The simultaneous sensor outputs were used to train machine learning models in order to separate the cross-correlated data. This led to an improved wind speed measurement range and increased accuracy, while allowing for additional parameters such as angle of attack and humidity to be measured.

The industrially fabricated CMOS-based anemometers achieve state-of-the-art wind speed measurements, while determining additional parameters such as the direction of the wind and the humidity of the air, despite being the smallest of their kind (2.2×10−2 mm2). The straightforward MEMS process, and the industrial CMOS process, make both sensor types cost-effective and scalable. In this thesis it is shown that the technology and driving forces behind Moore’s law, can be used in the world of Microanemometry to achieve Mo(o)re with less.