Predictive maintenance has often been hailed as one of the most immediate and effective uses for machine learning, and big promises have been made regarding its capabilities. However, it’s been slow to take off in practice. Even with falling prices on intelligent sensors allowing manufacturers to collect and transmit the various types of data such as temperature and vibration needed to drive the development of predictive maintenance programs, properly deriving actionable insights from that data without domain experts and on-site data analysts has proven more difficult than initially imagined. Unfortunately, when operators and plant managers can’t properly leverage this value, their industrial internet of things (IIoT) investments may not produce an ideal return on investment (ROI).
|Learn more about predictive maintenance.|
In the hopes of alleviating these issues, cloud-provider Amazon Web Services (AWS) has recently announced the general availability of Amazon Lookout for Equipment, a service that feeds data from end-users’ industrial equipment into the AWS cloud-based machine learning model to assist them in more accurately predicting machine failures. The value-proposition is simple: By training its models on larger quantities of data than any one company has access to, AWS can provide more powerful machine learning applications to end-users that otherwise would not possess the resources or on-site expertise necessary to develop their own. In exchange, those end-users furnish AWS with the data needed to continue developing more advanced machine learning capabilities. Moreover, Amazon Lookout does not charge any upfront fees, and instead bills based on data ingested and compute hours used to train custom models. As a result, the service may help small to medium-sized enterprises (SMEs) begin deploying machine learning driven predictive maintenance models more affordably.
“Many industrial and manufacturing companies have heavily invested in physical sensors and other technology with the aim of improving the maintenance of their equipment. But even with this gear in place, companies are not in a position to deploy machine learning models on top of the reams of data due to a lack of resources and the scarcity of data scientists,” said Swami Sivasubramanian, vice president of Amazon machine learning at AWS. “As a result, they miss out on critical insights and actionable findings that would help them better manage their operations.”
In addition, Amazon Lookout’s machine learning models are more effective than simpler rules-based predictive maintenance modeling procedures based on past performance that are commonly in use, AWS says. By identifying the unique relationships between different sensors and pieces of equipment, anomalies and failures can be addressed more effectively.
Amazon Lookout can be used within a single facility or across multiple locations. The service is available in the U.S., EU, and Asia-Pacific regions, with more regions to be announced in the coming months. Currently, Siemens Energy, Cepsa, Embassy of Things, RoviSys, Seeq, and TensorIoT are among the customers and partners using Amazon Lookout.