Quality has become one of the most crucial criteria in an institution’s success and survival as there is nothing more than what an era of globalization and intensity demands. Successful businesses recognize that consumer reliability may have a severe influence on their bottom lines. As a result, several competitive companies are constantly raising their quality requirements. Competitive companies think that improving quality is the best way to recover, and most authors have specified various procedures relevant to their processes. The majority of automobile assembly sectors are looking for high-quality requirements in their manufacturing techniques and are executing a quality system known as total productive maintenance (TPM). The study’s goal is to deploy the TPM program inside the metal forming industry to improve metal industry workstations. The overall equipment effectiveness (OEE) for various workstations such as rolling, bending, cutting, and die punching for the fiscal year 2018–2019 has been evaluated. In addition to the other reasons, inefficient resource utilization is a significant component that diminishes the factory’s OEE. In the financial year 2019–2020, the TPM approach was adopted in the enterprise. As a result, there has been an improvement in overall performance.

This study introduces a robust machine learning framework for predicting hydrochar yield and higher heating value (HHV) using biomass proximate analysis. A curated dataset of 481 samples was assembled, featuring input variables such as fixed carbon, volatile matter, ash content, reaction time, temperature, and water content. Hydrochar yield and HHV served as the target outputs. To enhance data quality, Monte Carlo Outlier Detection (MCOD) was employed to eliminate anomalous entries. Thirteen machine learning algorithms, including convolutional neural networks (CNN), linear regression, decision trees, and advanced ensemble methods (CatBoost, LightGBM, XGBoost) were systematically compared. CatBoost demonstrated superior performance, achieving an R ² of 0.98 and mean squared error (MSE) of 0.05 for HHV prediction, and an R ² of 0.94 with MSE of 0.03 for yield estimation. SHAP analysis identified ash content as the most influential feature for HHV prediction, while temperature, water content, and fixed carbon were key drivers of yield. These results validate the effectiveness of gradient boosting models, particularly CatBoost, in accurately modeling hydrothermal carbonization outcomes and supporting data-driven biomass valorization strategies. Graphical abstract