Maintenance management of stationary combustion engines in the agricultural sector remains largely manual, increasing the risk of unplanned downtime. This study developed a machine learning-based predictive model to anticipate failures within a 60-day horizon, enabling the transition from reactive to proactive maintenance. Following the CRISP-DM (Cross-Industry Standard Process for Data Mining) framework, a sliding-window feature engineering pipeline was built from 2250 historical records spanning 59 engines. Four ensemble learners (Random Forest, LightGBM, XGBoost, and CatBoost) were then compared under two complementary protocols: a strict 60/40 chronological split simulating deployment, and a stratified leave-engine-group-out cross-validation withholding entire engines from training. Nonparametric testing (DeLong test and engine-level cluster bootstrap) showed that the four learners are statistically equivalent, whereas the feature engineering layer contributes a large, significant discrimination gain (ΔAUC $\approx +0.06$, $p\approx {10}^{-27}$) on engines unseen during training. Random Forest, selected as the final model, achieved an AUC of 0.90 with 84.2% recall under the deployment protocol and 0.96 with 90.7% recall under engine-grouped validation. Temporal extrapolation, rather than cross-engine generalization, appeared to be the primary challenge, indicating that rigorously engineered degradation features, more than the choice of ensemble algorithm, drive predictive performance in agricultural maintenance planning.

Maintenance management of stationary combustion engines in the agricultural sector faces critical challenges owing to a reliance on manual methods, which increases the risk of unplanned downtime. This study developed a machine learning-based predictive model to anticipate failures within a 60-day horizon, facilitating the transition from a reactive to a proactive maintenance approach. Following the CRISP-DM methodology and drawing on a historical dataset of 2,250 records from 59 engines, feature engineering techniques were applied to derive 48 predictor variables, while K-Means clustering was employed to identify operational load profiles. The performance of two ensemble algorithms was then evaluated; LightGBM outperformed Random Forest under five-fold cross-validation with a 60/40 temporal split. The proposed model achieved an area under the ROC curve (AUC) of 0.91, a precision of 92.9%, and a recall of 76.1% for detecting actual failures. The findings indicate that gradient boosting techniques are highly effective for optimizing maintenance planning and reducing operating costs in the Ecuadorian agricultural context.