Solving the Manufacturing Data Problem
The Challenges of Automated Data Collection
There are many challenges involved in setting up automated data collection in a manufacturing environment for the first time. The main technical challenges often revolve around connecting to machines that have not been previously connected to a customer’s manufacturing network or communicating with them using a protocol not previously used by the end customer. In most projects, questions concerning what data should be collected, how and where collected data should be stored, as well as the proper method for analyzing collected data also pose challenges.
While there are always a number of such technical topics to be considered when undertaking or leading the implementation of an automated data collection project, perhaps the greatest challenge involves aligning and planning for a future state. While implementing automated data collection on a small scale is an important step forward, understanding how implementation fits into the overall vision of the end customer is of vital importance to ensure that their long-term strategic goals can be met. Understanding how well the solution scales, if it meets the requirements across the entire enterprise and how it fits into an overall long-term architecture are some of the challenges involved in implementing a new digital data collection system.
A Question of Protocol
Any time two or more systems talk with one another, a common language is required in order to facilitate this communication. In technical terms, we call this language a “protocol.” Protocols are everywhere, both within the manufacturing sector and outside of it. In fact, the simple act of accessing this article required the use of common protocols, particularly HTTP. Within the manufacturing context, understanding protocols is of vital importance, particularly when it comes time to interconnect your various machines and applications to one another.
In a perfect world, a single standardized protocol would exist, allowing all machines and software applications to be integrated to one another seamlessly with minimal effort. Unfortunately, the reality is often far more complex, particularly when it comes to integrating software and hardware in a pre-established manufacturing environment. Protocols, protocol versions and communication standards are constantly evolving, and it is often left to the discretion of the machine manufacturer to determine which protocols will be supported, at what version, and how closely the related standard will be followed.
For manufacturers, this means that a long-term strategy for enabling communication between various machines, utilizing various protocols, must be at the forefront of any effort to create a connected environment:
- What standards should be supported, at what versions?
- How can we enable communication between systems that work together, but speak different languages?
- How can we enable our machines, as a whole, to speak to our production and business applications in order to provide better information transparency and enable better decision-making at every level of our organization?
These are just a few of the questions that must be carefully considered from the very beginning of any effort involving plant connectivity.
The Standards for Industry
In a typical discrete or process manufacturing setting, it is common to see the use of the OPC communication protocol. In fact, OPC (UA or DA) is one of the more commonly used protocols for communication to and from PLCs and thus physical machines. While OPC is, of course, not the only standard in manufacturing, it is highly likely that one or more machines in any discrete or process setting will have an interface that supports one of the OPC variants.
The semiconductor industry is a bit different in this regard, preferring its own standard for communication to and from machines used for specific processes in its unique manufacturing environment. Published by SEMI (Semiconductor Equipment and Materials International), the SECS/GEM standard is heavily favored in many organizations related to the manufacturing of semiconductor technology. The SECS/GEM standard is quite different from OPC, as it allows for additional functionalities related to equipment automation and control, measurement data collection, changing of variables, recipe selection and more.
Not All Data is Created Equal
It should come as no surprise that data, particularly data from the shop floor, has an immense value when it comes time to make decisions. From top executives planning long-term strategies at the enterprise level all the way to the shop floor where decisions may be made by the minute or second, collected data is used to inform smarter, more effective choices.
What may come as a surprise to some is the fact that not all data is created equal. In fact, there can be a huge variance in the levels of quality, accessibility and costs associated with data collection, storage and analysis. While data collection is absolutely essential in the world of modern manufacturing, collecting it in the most effective and efficient way should be considered equally important.
Why Automated Data Collection?
Automating data collection, where possible, provides a huge number of strategic benefits over manual collection and dissemination methods. First, and perhaps most importantly, automated collection of data reduces the possibility of human error in the data collection process, thereby improving the overall accuracy of collected data. Moreover, with automated collection, data can be instantly made available, reducing potential delays in information dissemination across the enterprise and increasing the overall level of information transparency throughout the entire organization. Faster, more accurate, more available data enables decision-makers at all levels to make smarter, better and more informed decisions.
While automating data collection does provide immediate benefits, its place from a long-term strategic point of view must also be carefully considered. In addition to the benefits mentioned, automated data collection also begins to pave the way for standardization of reporting and analytical processes. Standardization of storage locations, data formatting, data availability and more make it possible to ensure that key decision-makers have access to the right data, at the right time and in the right way. While the consumers of collected data are often human beings, considering data usage at the enterprise software level is also of vital importance. The automatic collection of data, and the dissemination of that data, is a good first step on the path towards the automation of production and business processes, and an important milestone along any organization's digital transformation journey to Industry 4.0.
