Smart Ships – Get Smart With Sensemaking

0

Leveraging on the knowledge gained through the many platforms built, ST Marine started developing its very own in-house ship management system (SMS), NERVA in 2015. Striving to make the NERVA SMS smarter, the sensemaking module was developed. The module applies data analytics and machine-learning techniques on data collected by the SMS to provide predictive analysis of the equipment health on platform machineries. Integrating the sensemaking module with the existing NERVA SMS, the SMSis formed.


Figure 1. The sensemaking module applies data analytics and machine-learning techniques on the data collected by the SMS

In the current world of Big Data, there are three distinct layers of analytics, namely the descriptive layer which is addressed by the NERVA SMS, and the predictive and prescriptive layers, addressed by the sensemaking module.

The Descriptive Layer

The descriptive layer uses descriptive analytics to describe and summarise real-time data to provide insights of what and why an event occur. The NERVA SMS performs real-time control and monitoring of the platform systems onboard vessels to achieve this.

The Predictive Layer

The predictive layer uses predictive analytics to forecast possible future events. Through the use of various statistical techniques such as data mining, machine-learning and predictive modeling to analyse the current and historical data captured, the sensemaking module consists of predictive models and algorithms, which will be able to predict possible faults for the equipment health in the near future. This will provide a good overview of the time left till the equipment fail, by assessing the extent of deviation and degradation of the system from its expected normal operating conditions. Other than predictive models, the sensemaking module is also equipped with self-learning capabilities; learning from the data that is being analysed and refining its predictors automatically for improved accuracy.

The Prescriptive Layer

Prescriptive analytics provide advisories and recommended actions for the predicted failure. This is addressed by the decision support engine in the sensemaking module, developed through collective expertise and experience from ST Marine’s Engineering Design Centres (EDC) and Engineering Service Centre (ESC). The decision support engine is able to provide two sets of recommended actions, actions to be performed by the operator during operation and actions to be performed at depot.

Figure 2. Sensemaking module mimic

Figure 1 shows a sample of the sensemaking module visual representation. The health of the equipment is being monitored by various fault modes, which indicates green for normal, yellow for warning and red for critical. When these fault mode icons are clicked on, a dialog box will appear, listing down details relating to the fault. When the dialog box is clicked, the decision support engine can be called upon and recommended actions for the respective fault will appear as a closable sliding panel. This guides the operator in performing a certain set of actions during operation, and if the fault is not dismissed after these actions, the operator can plan for corrective maintenance at the next port of call.

With these features in place, the sensemaking module on the ship management system allows the customer to anticipate machinery faults and do condition based maintenance as well as further refine the maintenance schedule. Therefore, there will be an improvement in equipment performance and reduction in downtime, avoidance of unnecessary maintenance and extension of overhaul intervals, to achieve a reduction in total costs, increasing availability and better assets planning.

Using data from past sailing experiences, ST Marine had placed the sensemaking module to test and achieved encouraging results, showing that the sensemaking module was able to predict a recorded failure correctly. ST Marine will be conducting trials on live vessels and continue to look for new ways to further enhance the sensemaking module.

0