Experimental Study on Slosh Dynamics Estimation in a Partially Filled Liquid Container Using a Low-Cost Measurement System

A classic problem in the field of engineering and especially fluid dynamics is the liquid sloshing. Advancements in aerial robotics for agricultural purposes also drew attention towards the stability of pesticide spraying platforms under the influence of sloshing. Several mechanics models of the sloshing phenomenon had already been developed which are capable to predict the slosh within certain limits. This paper aims to identify sloshing parameters using a novel approach that utilizes a low-cost sensor. This research work utilizes Kalman filter to reduce measurement noise which is inherited by the sensor. The experimental setup used in this research consists of a rectangular container placed on a conveyor belt. Ultrasonic sensor was mounted on the top of the container, whose slosh parameters need to be identified. Sloshing data was acquired using an ultrasonic sensor in the presence of the input supplied by the motion of the conveyor. System identification is used to identify the system model. Luenberger observer and Sliding mode observers were utilized to check the accuracy of the predicted model. The RMSE value and the best fit estimation percentages were calculated for comparison of accuracy and error analysis. After several tests with different slosh levels, the recorded data is analyzed. The results support our proof of concept to measure slosh under dynamic conditions using a low cost sensor. The identified model of sloshing can be incorporated on an agriculture pesticide sparing drone for precision spraying resulting in efficient spraying of pesticide.