Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Predicting liquid dairy manure temperature during storage using machine learning and finite element analysis tools

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  2021 ASABE Annual International Virtual Meeting  2100278.(doi:10.13031/aim.202100278)
Authors:   Rana A Genedy, Jactone Ogejo, Matthias Chung
Keywords:   Dairy sustainability, Heat transfer model, Dairy manure, Machine leaning, Manure storage structures

Abstract. There is no standard method for estimating the stored manure temperature. And most commonly, decision support, life cycle assessments, and nutrient cycling tools used to assess sustainability in agricultural production systems use the ambient air temperature as an estimate for the manure temperature. This study explores two different methods to predict the temperature of manure during storage, using a heat transfer-based finite element method and machine learning techniques. The COMSOL Multiphysics software was used to develop the heat transfer model. Four supervised machine learning algorithms included gradient boosted trees, bagged tree ensembles, random forest ensembles, and feedforward neural networks. The data used in this study were collected from two different manure storage structures, a clay pit and a concrete tank. Each storage structure was instrumented with temperature sensors mounted at different depths and locations. A weather station was installed at each farm to collect weather data. In general, the temperature of the stored manure had a sinusoidal pattern similar to but out of phase with the ambient air temperature and solar radiation trends. The average manure temperature was higher than the average ambient air temperature for most of the year. The depth from the bottom of the storage structure had a significant effect on and influenced seasonal trends of the manure temperature. The heat transfer models sufficiently captured the overall patterns of the manure temperature, but the machine learning techniques more accurately predict manure temperatures. We acknowledge that the heat transfer models can be universally used in different locations. At the same time, machine learning approaches work superior when handling different manure management practices and under unusual weather conditions. Each method has its challenges and advantages; hence, regulating the machine learning approach merged with the heat transfer model results may be an appropriate method for estimating the temperature of the manure during storage.

(Download PDF)    (Export to EndNotes)