Sensor Aggregation Models

In 1999, six employees of an oil company wrote to government officials about the potential for a major disaster on the pipeline. "It won't be a single gasket, or valve, or wire, or procedure, or person that will cause the catastrophe," wrote the employees. "It will be a combination of small, perhaps seemingly inconsequential events and conditions that will lead to the accident that we're all dreading and powerless to prevent."

In an ideal world, a pipeline monitoring system would ensure leak detection or prevention by monitoring factors such as flow, viscosity, temperature and the surrounding environment. Such systems would also be automated, affordable, and low-powered, since supplying power along a pipeline is difficult, if not impossible. Most importantly, the system must be able to communicate critical information in real-time to allow engineers to minimize the effect of the leak, or prevent disaster. No system to date has been able to meet these challenging requirements. But, researchers at Accenture Technology Labs believe they may have a solution that combines tiny sensors called Smart Dust, and a prototype called Sensor Aggregation Models. The prototype includes a model of a pipeline, and shows how various emerging technologies can impact the oil and forestry industries.

Sensor Revolution
Smart Dust is an emerging technology, being developed at the University of California at Berkeley, which will pack a sensor, power supply, analog circuitry, radio communication capability and a programmable microprocessor all into a package the size of a pinpoint. In its current form, the prototype is the size of a matchbox. But, the Berkeley researchers predict that in less than two years it will be shrunk down to Smart Dust size and will be inexpensive enough to scatter over a forest, field or pipeline. In its final form, Smart Dust will be able to identify, locate, power and organize itself into a communications network with other sensors and transmit information.

Accenture believes that Smart Dust is just one part of a revolution, marked by the proliferation of sensors in a variety of commercial environments where inaccessible, vast or remote assets need to be tracked and managed. In addition to being scattered over vast areas, these sensors can also be embedded into objects and other materials, such as oil in a pipeline, paint on a bridge, or a turbine blade on an aircraft.

Transforming Sensor Data into Knowledge
Accenture Technology Labs developed Sensor Aggregation Models to help manage and aggregate the data coming from potentially millions of sensors. This application receives data wirelessly from the sensors and then transforms individual data points into a cohesive, integrated view of the environment. By summarizing data on the back end, and building visual models with the data on the front end, the user can easily detect or investigate any changes in the object—or the environment—as soon as they occur.

For example, if throughput in an oil pipeline slows, the application reflects that with a change in color over the affected area. Or, if the application notes a change in temperature along an area of the pipeline, the user could zoom into that area to check other factors, such as air quality, which may provide early warning of a possible gas leak. Global positioning capabilities let the user see more, or pull back, adapting the granularity as needed.

A commercial farm with thousands of acres could use this technology to monitor crops in real-time from a remote location—including their growth rate, light exposure, moisture conditions, air quality, nitrogen and phosphorous levels, and pest control. The result? Higher yield, lower cost crops and fewer chemicals introduced into the environment since pesticides can be applied precisely where needed.

In addition to using a graphical representation, Sensor Aggregation Models can feed into desktop applications such as Microsoft Excel, allowing users to easily share critical information. When integrated with Web services, Sensor Aggregation Models can easily be shared among many users, either within an enterprise, or with third parties. For example, Web services would allow an insurer to monitor the conditions and market value of assets under policy in real time, in order to proactively manage risk.