Response surface methodology and Box-Behnken design optimize recycled-aggregate concrete mixes to reduce quality variation and cost.
Virtual pipe modeling predicts odorant absorption and desorption in gas pipelines, helping maintain safe concentration while avoiding overdosing.
Cross-fractionation characterization feeds a predictive model to replace costly trial and error when producing polyethylene films with targeted properties.
Segmenting generic training data by material class and processing parameters improves prediction accuracy for specific target properties.
A machine learning algorithm identifies optimal functional monomer and cross-linker ratios for molecularly imprinted polymer nanoparticles.