Development of Anemometer Based on Inertial Sensor

Abstract

The current article elucidates a study centered on the development of an anemometer leveraging an inertial sensor for wind speed measurement in the northeast region of Brazil, focusing on renewable energy generation. The study encompassed a series of experiments aimed at calibrating the anemometer, analyzing the noise generated by the inertial sensor, and scrutinizing the data acquired during wind speed measurement. The calibration process unfolded in three stages: initial noise analysis, subsequent inertial data analysis, and the derivation of calibration curves. The first two stages involved experiments conducted at an average sampling rate of 10 Hz. Simultaneously, the third stage incorporated data collected over a 1 h duration while maintaining the same sampling rate. The outcomes underscore the suitability of the anemometer based on an inertial sensor for wind energy systems and diverse applications. While the wind readings from the prototype exhibit considerable fluctuations, a three-length moving average filter is applied to the prototype's output to mitigate these fluctuations. The calibration surface was established using observational data, and the resultant surface is detailed. Data analysis assumes paramount significance in wind speed measurement, and the K-NN algorithm demonstrated superior efficacy in estimating the correspondence between measured and control data.

Keywords: anemometer; calibration surface; inertial sensor; renewable energy generation; wind speed measurement.

Figure 1

Schematic representation (a) of the real-time wind speed measurement system and (b) in a block diagram.

Figure 2

Schematic representation of the block diagram of the components of the inertial sensor, where (I) represents an accelerometer confined in the balancing mass, (II) represents an external case in rigid thermoplastic polymer, and (III) represents a system balancing mass.

Figure 3

Inertial system for inertial measurement. In (a), the system is observed with the control and data processing unit (I), power cables +5 Vcc (II), and the O-ring seal (III) to prevent possible moisture infiltrations in the electronic system. In (b), signal cables (IV) and the 3-axis accelerometer (V) mounted on a rigid thermoplastic polymer structure are observed, completely limiting its movement relative to the structure. In (c), the system’s balacing mass (VI) is observed, consisting of distributed counterweights around the sensor.

Figure 4

Reference axes of the inertial coordinate system of the device.

Figure 5

Flowchart for identification of firmware function blocks.

Figure 6

Physical domain under study.

Figure 7

Details of the numerical mesh used in the simulations.

Figure 8

Behavior of maximum displacement amplitude (mm) (a) and natural frequency (b) as a function of height for a slender rod.

Figure 9

Velocity contour in the XY plane at Z = 0 m.

Figure 10

Pressure contour in the XY plane at Z = 0 m.

Figure 11

Drag force and drag coefficient behavior as a function of fluid velocity.

Figure 12

Inertial anemometer noise in the time domain for (a) x-axis, (b) y-axis, and (c) resultant speed.

Figure 13

Comparative analysis between the inertial anemometer noise distribution for the x-axis (blue line) and y-axis (red line) and with the expected normal distribution (black line).

Figure 14

Behavior of acceleration values for the X-axis (a) before and (b) after calibration.

Figure 15

Behavior of acceleration values for the Y-axis (a) before and (b) after calibration.

Figure 16

Behavior of acceleration values for the Z-axis (a) before and (b) after calibration.

Figure 17

Calibration curve for the inertial sensor after calibration.

Figure 18

Temporal analysis of (a) maximum speed measured by the inertial sensor [m/s] and (b) average speed measured by the inertial sensor [m/s]. The values correspond to the velocity measured by the inertial sensor during the proof-of-concept tests.

Figure 19

Wind rose with the respective wind directions.