Optimization of Palm Rachis Biochar Waste Content and Temperature Effects on Predicting Bio-Mortar : ANN and RSM Modelling

Abstract

In the current study, new prediction models were suggested to predict the compressive strength, porosity, and thermal conductivity of bio-mortar samples by replacing cement by weight with pyrolysis of Washingtonia filifera waste biochar (WFWB). Bio-mortars containing different biochar contents were prepared with the addition of 1%, 2%, 3%, 4%, and 5% pyrolyzed biochar at different temperatures 300°C, 400°C, and 500°C. The mortar samples produced were evaluated for compressive strength at 7 and 28 days. Relation between compressive strength porosity and thermal conductivity values (dependent values), and biochar replacement ratios and pyrolysis temperatures (independent values) was predicted by artificial neural network (ANN) machine learning techniques based on the Levenberg Marquardt algorithm. The results revealed compressive strength increase at 28 days of nearly 12%, 3%, and 2% at 1% optimal biochar replacement content for WFWB500, WFWB400, and WFWB300 specimens, respectively. Moreover, these results provide confidence in the manufacturability of bio-mortars with a compressive strength of 63.81 MPa at 28 days using only 1% substitution of WFWB500 and a thermal conductivity coefficient of 0.52 W/m.K. The importance measure of the variables shows that the most influential variables are the percentage of biochar. The statistical and experimental results also revealed satisfactory agreement.

Keywords

• bio-mortar
• cement
• biochar pyrolysis temperature
• compressive strength
• washingtonia filifera waste biochar
• artificial neural networks