With the rise of the Internet of Things (IoT) industries are awash in petabytes of Big Data streaming from an ever increasing array of wired and wireless sensor networks. These interconnected sensors are continuously monitoring and reporting on everything from heart rates, to production flow, to turbine wear, creating opportunities for industries poised to take advantage of this wealth of data to improve product usage, reduce maintenance costs, avoid equipment failures, and drive improvement in their business operations.
Instead of relying on incomplete or unrepresentative laboratory simulations to predict the behavior of devices in service, manufacturers are now able to track the usage and performance of their products in nearly real time. As IoT-enabled products infiltrate the market, embedded software and interconnectivity are blurring lines between the physical and the digital in terms of monitoring device health and device capabilities, and responding to usage patterns.
Some industries using sensor analytics to drive actionable insight and growth from IoT Data are:
- Medical Devices
- Consumer Products
- Oil and Gas
While the potential value derived from sensor analytics appears limitless, the process for unlocking this value can be challenging. Not only is the volume of data changing, but the variety of data sources is increasing as well. All predictive analytics efforts involve some data cleaning, however, the volume and variety of data involved with sensor analytics place an even greater burden on data storage, preparation, and infrastructure design. Process owners must weigh the costs and benefits of analyzing in real-time, versus in batched sessions. IT departments must assess and prepare for the greater cybersecurity risks posed by storing increasing amounts of sensitive customer data and safeguarding it against breaches.
In addition to the usual challenges of model creation and selection, getting from raw sensor data to actionable predictive results requires enormous pre-processing, including careful consideration of sensor data logging frequency, scalable analytics data storage, and rigorous processes that monitor and assess the quality of incoming data for analysis. Analytics teams must decide where their analyses should take place: remotely in the cloud, in a centralized analytics service, or at the site of the device.
Sensor Analytics Applications
Elder Research has extensive experience helping our clients use predictive analytics to filter through the noise of Big Data from sensor networks to reveal actionable insight. Examples of our Sensor Analytics applications include:PREDICTIVE MAINTENANCE AND RELIABILITY ANALYSIS
Equipment downtime can be costly, last minute repairs are often difficult to schedule, and replacement parts may be difficult to find. Elder Research Sensor Analytics can predict when your equipment is likely to need maintenance, allowing you to proactively schedule preventative maintenance and repair to prevent costly equipment downtime.ANOMALY DIAGNOSIS AND DETECTION
Sensors analytics can be the first line of defense in detecting anomalous behavior or readings before a fault degrades user experience or leads to costly device failures. Sensor readings unusual for normal operating conditions can trigger alerts and enable intervention by experienced technicians.PRODUCT USE CASES
Many devices track product usage through log data. As embedded software becomes increasingly commonplace in everyday devices, these software logs act as virtual sensors, creating a rich data source that provides valuable insights into device usage and performance. Elder Research has experience performing Product Usage Analytics and Log Analytics Solutions to provide value for our clients by creating predictive models, helping visualize usage trends, and improving user experience.CUSTOMER BEHAVIOR
Sensor data is an incredibly rich information source that can be used to understand different user behavior, segment populations for targeted marketing, and personalize customer interactions. Elder Research sensor analytics solutions can enhance customer experience, reduce churn, and increase customer lifetime value.
Sensor Analytics SOLUTIONS
With the ubiquity of sensor data from connected devices, software usage logs, and equipment monitors, there are ample opportunities to deliver value from these vast data sets using advanced sensor analytics. Whether you are new to sensor analytics or looking to augment existing capabilities, Elder Research can provide support where you need it most. Examples of our sensor analytics solutions include:
Using Sensor Analytics to Predict Natural Gas Well Freezing
Elder Research used 20 years of detailed sensor readings from hundreds of wells to characterize transient well states. The goal was to predict gas well shut-ins due to hydrate formations four to six months in advance in a North American field, where winter months allowed only very limited access. Our model was far more accurate that the existing heuristic models and allowed operations to know which clusters of well pads should be prioritized for treatment. A second model was built which clearly showed when in the new well life cycle a plunger pump system should be installed. The estimated return on investment for each model was less than one year
Using Product Usage Analytics to Improve Software User Experience
At Elder Research, our analytics consulting expertise enabled us to successfully use over 1TB of log data to build a user segmentation model for a major software client. These segments have proved hugely valuable in helping understand user needs and behavior. Additionally, we were able to create a custom visualization tool which allowed the client to visualize their log data in a way that was previously impossible, helping them to better understand and reach out to their customers.