Low-Cost Predictive Maintenance Framework for SMEs: Illustrative Application in Ghana and Poland

Abstract

Maintenance expenditures constitute a significant portion of operational costs in manufacturing. Small
and medium-sized enterprises (SMEs) in both emerging and developed economies struggle to adopt
high-cost maintenance technologies. This paper proposes a conceptual low-cost predictive
maintenance framework designed for Ghanaian and Polish SMEs. The framework integrates Pythonbased
open-source software and IoT sensor data with machine-learning techniques to forecast
equipment failures and support real-time maintenance decision-making. The framework is validated
through a proof-of-concept case study using a real-world predictive maintenance dataset (AI4I 2020)
from a hydraulic system, serving as an analogue for water bottling production facilities in both
countries. This approach presents a robust, and reproducible validation of the framework’s predictive
capabilities. The framework is validated through a proof-of-concept case study using synthetic data
representing water bottling production facilities in both countries, rather than full-scale industrial
deployment. Grounded in a decision support system (DSS) methodology aligned with Industry 4.0
paradigms, the framework demonstrates potential feasibility and benefits in a simulated environment.
The results indicate that the approach could potentially reduce downtime and maintenance costs in the
simulated scenarios. This study contributes to developing affordable and scalable predictive
maintenance strategies tailored for SMEs across diverse economic contexts, while clarifying that
findings are based on synthetic data for illustrative purposes.