Data Analytics in the Maritime Space

Marine Link | BY Walter Mitchell

Ship owners and operators, machinery OEMs and regulatory entities are embracing much needed technological innovation as demand grows in protecting machinery and communications on maritime assets. Cognitive analytics is a game-changing technology that is now more widely available to the maritime sector.

data analytics

This is the latest evolution of data analytics: from the “days of yore” of logging data in a logbook, to sensing data and connecting to a central console, transmitting data ashore and using artificial intelligence (AI)-enhanced tools to develop a deep understanding of how machines behave.

These platforms have dramatically expanded the toolbox for fleet managers by creating the most in-depth analysis available in the marketplace. It is estimated that 10-12 percent of maritime industry asset owners now use some form of predictive analytics, but only to a limited extent. While many executives understand the potential benefits primarily as cost savings in maintenance and capital cost replacement they often don’t know how to obtain this technology for themselves. 

As data science continues its growth into more industrial applications, tech-savvy and forward-thinking operators are embracing the full reach and potential of AI-enhanced technologies. The progression of analytics from the descriptive (what happened?) to diagnostic (why did it happen?) to predictive and prescriptive analytics (when is it likely to happen again and what can I do to prevent it?) is changing the way industry addresses maintenance and operations.

Predictive analytics

Predictive and prescriptive analytics are the logical next step in analytics, and an important new frontier for the maritime industry. At SparkCognition, we are focused on the foresight presented by cognitive analytics, or analytics that use machine learning. Consider these “what ifs”:

  • What if mechanical anomalies could be detected in real time?
  • What if that detection was so granular that it could categorize those anomalies into minor, intermediate, or serious?
  • What if those anomalies could be shown in 3D, displaying exactly which component of the machine was degrading?
  • What if that detection could be built upon with automated model building so that it could predict when maintenance was actually needed, or when failure might occur?
  • What if gigabytes of sensed data were streamlined from shipboard sensors, aggregated via IIoT functionality from all vessels in the relevant fleet, and transmitted to a central receiving point?

 Cognitive analytic tools are capable of all of these applications. They detect anomalies in machine operations and predict failure with high degrees of confidence. Detection and forewarning greatly assist the operator by allowing for more precise planning of maintenance and capital replacement. The result is that the life of the machine can be extended, yielding considerable cost savings. Ship owners, operators and others in the maritime space are increasingly interested in this new technology for the following reasons:

  • Cognitive analytics can extend its understanding of the difference between traditional diagnostic maintenance and predictive and prescriptive maintenance.
  • Ship owners and operators recognize that there exists the capability to ingest the (potentially) gigabytes of data that have already been generated and can use it to gain new insights into operations.
  • Cognitive analytics allows ship owners and operators to intelligently plan major maintenance periods such as special surveys and drydockings, adjust spare parts and consumables inventories and support seagoing staff in assessing in-voyage and longer-term maintenance needs.
  • Ship owners and operators appreciate that there is a role in developing a deeper understanding of the machine and improving its overall health.

 SparkCognition’s SparkPredict platform is an AI-based cognitive and prognostic system already in use in the aerospace, oil and gas, utility and financial market sectors. It is physics and asset agnostic—it can use data regardless of the platform producing the data, or the format of the data.

It ingests new and historical data (structured or unstructured), installs quickly and easily, and has a low learning curve for operators. Its automated model building feature allows for notification of suboptimal operating conditions before harm occurs to the machinery. Combined with natural language processing (NLP) technology, the platform ingests and presents free-form text from OEM manuals and service guides. 

In one example, a ship operator had attempted to engage NLP across its 105-vessel fleet to digitize more than 35,000 technical manuals, but gave up as it was too large a task. The capacity needed to ingest that volume of information can overwhelm most platforms. This is not the case, however, with SparkPredict, which can easily handle datasets of that size and larger.

Dependency on a new technology to inform machinery maintenance is a new way of thinking about overall machine health. Lloyd’s Register, in its “Global Marine Technology Trends 2030,” estimated a 4,300 percent increase in the annual data generated by ships by 2020, and says that “by 2030, that figure will have increased even further as this is an accelerating trend.” The proper management and analysis of “smart data” will have a major impact on the maritime space. This trend is being driven by the demand for better use of information coming from the ship, and the need to provide the most cutting-edge tools to managers desiring to have stronger control over maintenance and capital equipment replacement budgets.