How To Deal With The Flood Of Analog Data

Analog data from a variety of sensors and other devices is a huge problem. Here are three approaches to overcoming the problems that big analog data can cause.

Approach 1: Analyze at the Edge
A lot of data can be collected at the point of capture, but most of it’s uninteresting. You can save and analyze it all or you can take advantage of intelligent embedded software that constantly measures and processes that data, focusing on the information that’s deemed worthy. Using this method, you can:

• Drastically reduce the amount of data you have to manage
• Be confident you’re keeping the important nuggets of information
• Validate collected data
• Analyze data right at the source

For example, imagine a power plant that’s monitoring pump and motor assets to improve uptime and reduce maintenance costs. A single pump/motor setup can generate over 20 GB of data per day. Over an entire plant with 80 assets, that equals more than 1.6 TB of data per day.

NI is currently doing this at its Austin, Texas, headquarters with the NI InsightCM online monitoring demo.

Using the intelligent embedded software of the system, most of the data collected is filtered out so that only the important events detected are kept. Engineers then monitor and analyze this data to gain valuable insights that drive business results.

Brett Burger, Principal Marketing Engineer, Monitoring Solutions, explains the tools used that help keep heavy equipment up and running.

Approach 2: Harness Server Technology to Automate Data Management and Analysis
Sometimes analyzing data at the edge isn’t an option. Like when you don’t know exactly what type of phenomena you’re searching for. In cases like these, you have to collect the entire data set, making it crucial that you have the proper tools to find the important information, validate the data, and analyze it as efficiently as possible. This means automating as much of the process as possible and taking advantage of a server system’s computing power.

Jaguar Land Rover put this into action to help manage the 500 GB of data being produced every day for over 400 engineers in their powertrain calibration and controls department. The company created its own preprocessing tool to automatically validate raw data, and then indexed it with DataFinder Server so the whole team could quickly search through the mountain of collected data. Finally, the engineers used DIAdem to perform standardized or ad-hoc analysis and generate templated reports.

The results? These tools helped them analyze their data 20 times faster than previous methods while also increasing the amount of data they analyzed to 95 percent. They could get even more efficiency gains with new features introduced in the Data Management Software Suite this year.

Figure 2. The Data Management Software Suite provides tools to build a complete enterprise data management solution.

Choose the Right Tools to Wrangle the Big Analog Data Problem
Regardless of the approach you choose, off-the-shelf hardware and software can efficiently and automatically validate, search, and analyze your data so you can make better data-driven decisions.