Why Your Factory Floor Data Needs Purpose

Factory floor data collection has evolved from mainframes and USB data logging to modern Ethernet networks and historians. Most manufacturing data is time-series, and its value depends on a clear structure, hierarchy, and purpose. The key is not just how to collect data, but what to collect, when to collect it, and how to use it to drive measurable results.

The First Rule of Factory Floor Data Collection

With all the fancy hardware and software tools on the market, I have marveled at how easy newer tools have made it for engineers to collect data just for the sake of collecting it. They have no idea what it is for, when to collect it or how to use it. The first rule of data collection should be “Know your reason for collecting data.”

How is Factory Floor Data Typically Structured?

Factory floor data is typically structured as relational data or time-based data.

What is Relational Data?

Relational data is information structured so that relationships between data items are apparent. For example, researchers studying population growth may categorize women of birthing age by race, education and socioeconomic status…etc. That’s relational data. Similarly, a manufacturer might categorize pump data by machine age, manufacturer, or maintenance history to find patterns.

What is Manufacturing Time-Series Data?

Time-based information is data collected over time and stored in a time sequence. Most manufacturing data is time-based. Every manufacturing environment, from food and beverage to automotive to consumer products, generates time-series data. Most of this time-stream data available in a production system is never collected.

Why is Hierarchy Important for Manufacturing Data?

To optimize the use of manufacturing data, it should be hierarchical and aligned with the plant’s systems. For example, Plant 3 has five lines, each with three manufacturing cells, each consisting of a PLC and one or more pumps. The time-series data that you collect for the pumps, for example, should identify that entire hierarchy.

Factory Floor Data Collection Tools: Data Loggers vs. Historians

How to collect data isn’t so much about transferring the bits and bytes into storage, but about how sophisticated we must be? If all you need is to collect electrical data for a short period to identify any spikes, a simple data logger will suffice.

Suppose you need to collect operational data such as pressures, temperatures, weights and the like for quality analysis. In that case, you may want to migrate your data to a sophisticated historian such as RTA’s Allen-Bradley PLC Historian.

This product is more than just a simple historian, but without the advanced capabilities (or expense) of other popular historian solutions. Its key features include logging data directly from A-B PLCs without any programming or logic, local storage extraction via a web interface or USB, and sending data out to MQTT clients through a broker.

For a deeper dive into which setup fits your specific application, check out our detailed breakdown of data historians vs. data loggers.

The History of Manufacturing Data Collection

Engineers began collecting this data in the very early days of automation. The largest corporations led this effort when mainframes were the only available processing systems. In those days, vast sums were devoted to building custom software systems to retrieve data from Programmable Controllers on the factory floor. There was no such concept or implementation of anything resembling IT.

As microprocessor prices fell, the personal computer was introduced, and as more computer-literate engineers entered manufacturing, more and better ways of collecting time series data emerged. Some of the first systems deployed microprocessors on the factory floor that collected time-series data and stored it in USB Sticks. Those sticks were manually collected and transferred to other systems for analysis.

Today, we have much more sophisticated collection mechanisms and very sophisticated systems for processing the data that is collected. We have networks such as EtherNet/IP, PROFINET and Modbus TCP that can access data in devices. We have sensors that can be deployed to capture data not included in control algorithms or not present in a programmable controller. We can even collect data from widely dispersed systems, such as water and wastewater pumping and lift stations located hundreds of miles from the processing facility.

Give Your Factory Floor Data Purpose

If you decide to go through the expense and trouble of collecting manufacturing data, use it to accomplish some goal. Don’t collect it just for the sake of collecting it.