Methodological Evaluation and Time-Series Forecasting for Yield Improvement in Nigerian Industrial Machinery Fleets

{ "background": "The operational efficiency of industrial machinery fleets is a critical determinant of productivity in developing economies. In Nigeria, systemic underperformance and a lack of predictive maintenance frameworks lead to significant yield losses and capital expenditure waste. Current evaluations are often reactive, lacking robust, data-driven methodologies for forecasting and improvement.", "purpose and objectives": "This paper aims to develop and evaluate a methodological framework for the systematic assessment of machinery fleets. Its core objective is to construct a time-series forecasting model to predict yield performance and quantify potential improvements from targeted interventions.", "methodology": "A hybrid modelling approach was employed, integrating Autoregressive Integrated Moving Average (ARIMA) with exogenous variables (ARIMAX) to account for operational and maintenance factors. The model, specified as Yt = \μ + \∑i=1ᵖ\ Yt-i + \∑i=1q\ \εt-i + \∑j=1ᵏ\ Xj,t + \εₜ, was trained on high-frequency sensor data from a fleet of 47 heavy industrial machines. Model diagnostics included Ljung-Box tests and analysis of robust standard errors.", "findings": "The ARIMAX(2,1,1) model demonstrated strong predictive capability, explaining 78% of the variance in yield metrics. A key finding was a forecasted 22% potential yield improvement through optimised preventive maintenance scheduling, with a 95% confidence interval of [18.5%, 25.3%]. The integration of vibration analysis data as an exogenous variable proved statistically significant.", "conclusion": "The proposed methodological framework provides a viable, evidence-based tool for transitioning from reactive to predictive fleet management. The forecasting model offers a quantifiable basis for strategic investment in maintenance, directly linking operational data to productivity outcomes.", "recommendations": "Implement the ARIMAX forecasting model as a core component of fleet management systems. Establish continuous data logging protocols