Data Science and the Future of the IoT: Are You Ready For the Paradigm Shift?
- December 9, 2016
- 2994 views
Put simply, the Internet of Things (IoT) is the current industry term to describe machine-to-machine (M2M) communications and the maturation of that technology. This evolution will make it increasingly attractive for organizations to automate two-way connectivity for sensors and devices into their business processes. As IoT technologies become more widespread, the data gathering and analytics process will shift from passive collection of data to taking a more active role in implementing and operating new sources of data.
Historically, a new analytics project started with defining a business need, uncovering the pre-existing data to support solution development, and using a centralized platform to analyze and operationalize the new insights that are gained from that data. However, the model is changing as organizations incorporate new hardware sensors or distributed intelligence throughout their information network. Two clear IoT focus points for analytics teams looking to unlock business value will be new sensor enablement and decentralized (edge) computing.
New Sensor Enablement: As the cost of new sensors decreases and programming and communications for the sensor platforms become more accessible, teams will be shifting from asking themselves “What data do I have?” to “What new data streams can I enable?”. Concepts that are applied in existing lightweight hardware (i.e. Raspberry Pi, Arduino) and scripting platforms or Integrated Development Environments (IDEs) are quickly transitioning from the hobbyist and academia world into mainstream applications. Cost effective sensor and communication applications for utility and city infrastructure will include but will not be limited to:
- Asset Tracking and Logistics: Prevent supply chain delays with real time tracking of critical materials.
- Electric Infrastructure Monitoring:
- Stray Voltage Detection: Enable remote monitoring and reporting of stray voltage on streetlights and utility structures (safety concern).
- Arc Fault Detection: Detect and report arc faults on the underground distribution system before they develop into a hazardous situation.
- Pole Tilt Monitoring: Streamline pole maintenance, storm hardening, and recovery by deploying sensors to monitor utility pole infrastructure.
- Gas Infrastructure Monitoring: Remotely detect natural gas leaks and voltage at corrosion potential gas piping test points to improve response times and reduce manual testing requirements.
- Water Distribution System Monitoring: Connect sensors to routinely monitor chlorination, leak detection, turbidity, etc. in existing water distribution infrastructure.
- Environmental Monitoring: Monitor radiation, particulates, etc. at facility discharge points.
- Intelligent Lighting, Smart Cities: Extend the sensor network to civic projects such as street lighting, parking, trash collection, signage and advertising, public safety, pedestrian counters, etc.
- Building Sub Metering: Augment energy efficiency service offerings with building sub metering sensors.
Edge Computing: The emerging IoT hardware platforms will allow us to more easily push the processing power closer to the data source. This allows for rapid decision making and reduced dependency on monolithic back office applications. As new insights are gained, teams will go “the last mile” in implementing operational process and deploy logic down to devices or data sources to take action locally. Remote programming capability is now expected to be in the hands of the device owner and operator as opposed to relying on hardware and software manufacturers for all change implementation. “Sensor fusion”, where generation of valuable insights requires data correlation from multiple sensors, will require IoT companies to adopt robust and standardized ways to collect, aggregate and analyze sensory data in order to easily create new interconnectivity models between devices. Utility applications of this technology include (but, again, are not limited to):
- Convergence of Automated Metering (AMI) and Distribution Automation (DA): Communication “teaming” of distribution and AMI devices to share data and act in real time.
- Localized Voltage Regulation and Reduction Schemes: Locally gather intelligence and react to voltage fluctuations on the grid. This could include a feedback loop between AMI and DA devices in the field, or simply programming the localized devices to push real time alerts to the back office based on a specific set of conditions.
- Intelligent Temperature Monitoring: Include logic to monitor AMI meter temperature based on instantaneous demand, time of day, season, or other factors to improve the alerting algorithm.
- Customized Power Quality Statistics: Gather and aggregate detailed power quality stats that are relevant to the particular residential or commercial location to better monitor and forecast grid operations.
- Building Controls Integration: Leverage remote controller/servers with MODBUS, SCADA, or another M2M protocol to integrate with a wide range of endpoints while leveraging a common communications platform.
- Distributed Generation (DG) Monitoring: Gather high granularity data and statistics from AMI and DA devices to analyze and react in real time to grid fluctuations caused by existing or unregistered DG.
- Security Monitoring: Use “sentinel logic” to monitor locally on the device itself for aggregations or sequences of events that are indicative of a security threat. Gather summary statistics for analysis or send real time alarms.
As utilities adopt IoT technology, reacting to problems will become less prevalent. Proactively sensing, predicting, and remotely acting upon operational issues will become the new norm. A cycle of continuous improvement of the business process, algorithms, and sensor and communication technology will be a requirement. IoT companies will rely on the creativity and innovation of business partners, subject matter experts, and open developer communities to execute on this vision.
How ready is your organization for this paradigm shift? Are you moving from a “kneejerk” reactive model to predicting and rapidly addressing business problems by empowering your analytics and IoT capability groups? Revolutions in an operational model do not need to be painful, traumatic experiences, rather, they can be an opportunity for organizational and personal growth. The utility industry is, by nature, full of problem solvers who are ready to embrace this change.