Methodological Evaluation and Time-Series Forecasting for Yield Improvement in Ghana's Industrial Machinery Fleets

Industrial machinery fleets are critical to national infrastructure development, yet their operational yield in many developing economies remains suboptimal. In Ghana, a lack of robust, data-driven methodologies for performance analysis and forecasting hinders strategic maintenance and capital planning. This report aims to methodologically evaluate current fleet management systems and to develop a predictive time-series model for forecasting machinery yield, with the objective of providing a tool for measurable improvement. We conducted a methodological audit of fleet management practices across multiple industrial sites. A seasonal autoregressive integrated moving average (SARIMA) model, specified as SARIMA (1, 1, 1) (1, 1, 1) ₁₂, was fitted to historical availability and utilisation data to forecast yield, defined as (availability × utilisation). Model diagnostics included analysis of robust standard errors. The methodological evaluation identified systemic gaps in data collection and proactive maintenance scheduling. The forecasting model demonstrated a statistically significant upward trend in predicted yield, with a mean absolute percentage error (MAPE) of 8. 7% on the test set. A key theme was the critical influence of scheduled maintenance adherence on forecast reliability. The developed SARIMA model provides a technically sound framework for forecasting fleet yield, offering a substantial improvement over existing heuristic-based planning methods. Methodological weaknesses in current systems are a primary constraint on performance. Implement the forecasting model for quarterly resource planning. Establish standardised data protocols across fleets to improve model inputs. Integrate forecast outputs into preventative maintenance programmes. fleet management, time-series analysis, SARIMA, predictive maintenance, operational yield, infrastructure This paper provides a novel application of the SARIMA model for forecasting industrial machinery yield in a West African context, delivering a concrete tool for evidence-based operational decision-making.