COVER FEATURE
INGENIEUR
COVER FEATURE
INGENIEUR
Ministry of Health Roadmap : Enabling AI and IoT in Biomedical Equipment Maintenance
By Ir . Ts . Dr Mohd Effendi bin Amran Senior Principal Assistant Director Engineering Service Division Ministry of Health Malaysia
It has become more challenging to maintain biomedical equipment in Government hospitals , where it would be required to function under adverse circumstances . In addition to maintaining quality , safety , and environmental protection , biomedical equipment maintenance is required to guarantee maximum efficiency and enhance availability at the lowest feasible cost . As such , with the Fourth Industrial Revolution ( 4IR ) underpinning the modernisation agenda , hospitals must incorporate emerging technologies to minimise functional failures due to the sheer increase in the quantity and complexity of biomedical equipment .
The 4IR ecosystem utilises data from sensors in equipment or devices , along with the monitoring of currents , pressures , temperatures , and other associated profiles , to forecast failures . Microelectromechanical system developments have made it possible to deploy a wide range of inexpensive sensors that can sense , calculate , and wirelessly share data for monitoring equipment and its environment . An alignment with a kind of Predictive Maintenance ( PdM ) that may be carried out even in the absence of data from previous equipment failures can be used to detect operational abnormalities ( or anomaly detection ). Machine learning ( ML ) or deep learning ( DL ) models ( i . e ., the subset of artificial intelligence ( AI ) development ), based on regression or binary classification , are extensively used in this regard to forecast imminent failures so that repairs or replacements may be arranged before the actual failure arises .
CHALLENGES AND STRATEGY
In light of the aforementioned , it is first necessary to strengthen an efficient maintenance strategy mix before integrating AI to ensure the equipment ' s durability and availability while reducing unnecessary maintenance investment . Thus , it is vital to consider the backbone of the maintenance plan to facilitate the best solution given the characteristics and flaws that these devices exhibit . The PdM , often referred to as Condition-Based Maintenance ( CBM ) or prognostics and health management ( PHM ), is the process of estimating an equipment ' s current status and predicting when maintenance should be performed . Rather than performing preventive maintenance ( PM ) regularly , PdM is more costeffective by performing maintenance at the right time . Two key tasks of PdM are prognostics and diagnostics . Prognostics forecasts Remaining Useful Life ( RUL ) in the same way that anomaly prediction does , while diagnostics is related to finding anomalies in that it examines the state of an equipment ’ s health .
Reliable-centred maintenance ( RCM ) is another comprehensive approach for establishing
12 VOL 95 JULY-SEPTEMBER 2023