IoT at the Edge: Putting Pressure Where It Belongs
If an incident occurs at a natural gas pipeline, human lives are put in danger and millions of dollars are at stake. The reason: a regular gas pipe serving a private home only operates at an excess pressure of 20 to 50 millibar, whereas a long-distance pipeline is pressurized with up to 200 bar.
The pipeline operator has to keep long-distance pipeline pressure constant across hundreds, even thousands of miles. For this purpose, pipelines are equipped with compressor stations at intervals of 25 to 60 miles. Every two minutes, each of these stations sends information regarding gas pressure, density, and caloric value (temperature) to the control center – 70 to 80 TB of data per station and year.
Analyzing this data is a laborious task: the control center typically receives data in Excel format; an engineer then determines the monitoring results from these Excel files by utilizing macros – a cumbersome and time-consuming process.
However, Hewlett Packard Enterprise and SAP are partnering to make the process more streamlined:
On July 11th/12th, SAP and Hewlett Packard Enterprise (HPE) demonstrated in Frankfurt, Germany, that today this kind of process can be executed easier, faster, and more cost-effectively. At this event, SAP introduced the new industrial IoT platform SAP Leonardo, extending the software vendor’s datacenter portfolio to industrial environments, which tend to be multilayered and complex – the very reason why SAP leverages partnerships with other technology leaders e.g. with HPE, a long-time SAP partner with extensive experience with industrial IoT applications and offers a broad IoT portfolio.
This portfolio includes, for example, services for planning new IoT processes and IoT architectures, for connecting them to the SAP HANA environment, as well as for operational support. Converged IoT systems and gateways specifically designed for rough industrial environments or a micro-datacenter perform on-site data aggregation and analysis at the edge. Big data solutions, based e.g. on Hadoop, allow for storing, managing, and analyzing huge amounts of data, while Aruba switches, WiFi components, and security solutions provide connectivity and security for IoT environments.
Data Analysis on the same place where data is created
Before any trend analysis can be performed in the data center, industrial IoT environments require an initial aggregation and analysis of sensor data at the edge, i.e. as close as possible to where the data originates. Otherwise, in the case of the German utility company mentioned above, PB (Peta-Bytes) of sensor data per year would have to be transferred across the WAN – for a utility network spanning about 7,500 miles, a size comparable to that of Germany’s highway, or “Autobahn”, coverage.
Using edge computing, sensor data can be aggregated on location, resulting in a combined compressor station status: if all sensor values are within the desired range, the station simply has to transmit a “green light” every two minutes; in case of errors, it only has to send the deviation from the target value. This way, aggregation and automation minimize data transfer cost as well as expenditure of time at the control center.
Additionally, this approach enables predictive maintenance: while traditionally, only threshold violations trigger alarms, edge systems can discover correlations early on, e.g. accumulation of minimal deviations from the desired state. Hence, a technician can replace the component before an error occurs. This minimizes operational costs and makes it easier to prevent disruptions and accidents.
Later, in the data center, central storage of preprocessed data in a big data architecture and analyses utilizing SAP HANA provide for optimized planning, quality control, cost structure, and field service management. This way, HPE’s hybrid IT and intelligent edge solutions complement SAP’s Leonardo strategy – making sure that high pressure is limited to the pipeline, not weighing on the operator.