Hospitals frequently manage biomedical equipment from multiple manufacturers and service providers, resulting in heterogeneous maintenance instructions and inconsistent technical records. This study proposes an interpretable software framework for standardizing preventive and corrective maintenance procedures independently of equipment brand. Technical information was extracted from the service manuals of three multiparameter patient monitors and organized into a dataset containing 61 records across eight technical fields. Text descriptions were cleaned, normalized, and represented using term frequency–inverse document frequency. K-means clustering was applied to group semantically related maintenance activities, while an interpretable decision-tree layer organized the resulting information according to equipment, component, maintenance type, frequency, failure, possible cause, and corrective action. The framework was implemented in Python through a Streamlit interface that generated standardized technical checklists with adjustable levels of detail. An Arena discrete-event simulation compared eight-hour maintenance workflows with and without the proposed reporting structure. Under the simulated conditions, the number of processed equipment units increased from 5 to 10, while the average total time per unit decreased from 2.94 to 2.15 h. Simulated technician utilization decreased from 100% in the baseline workflow to 31.92% under standardization. These results support the feasibility of combining text clustering, interpretable decision logic, and digital reporting to reduce documentation variability and improve traceability in maintenance processes. The framework remains a proof-of-concept, limited to multiparameter monitors and simulation-based evaluation.

Keywords

Biomedical equipment maintenance; Process standardization, kmeans clustering: decision-tree layer.