data, blue, code

The Internet of Things (IoT) and Machine Learning are key aspects of Industry 4.0. Both will result in the unprecedented collection and analysis of data to drive new insights and benefits. There is nothing particularly new about wanting to use manufacturing data to drive improvements. What is new is the transition away from the large data preparation effort that was often a large portion of Data Warehouse and even Big Data efforts. Data from disparate systems often went through multiple levels of aggregation and indexing in order to prepare it for answering traditional questions.

What Does this Mean for Manufacturers?

Manufacturers should be planning for all enterprise data sets to be part of the greater data lake. A data lake is a storage repository that holds a vast amount of raw data in its native format, including structured, semi-structured and unstructured data. The data structure and requirements are not defined until the data is needed (unlike a traditional database). The transition to a data lake emphasizes flexible access to analysis tools and is less centered on data preparation. By definition, the data lake will be made up of a variety of data sources and the accessibility requirements and effort will only be defined at the time of the query.

There is information that may be uniquely housed in the traditional ERP data that can be a core part of the Industry 4.0 effort. As an example, flexible manufacturing assets can be used to produce many different SKUs. There is a difference in the wear and tear on the asset while producing the respective SKUs. There can also be a difference between individual operators and how they run the asset. This data on variation between production orders and assigned operators often does reside in the ERP system.

The onslaught of volumes of IoT-based asset-centric process variables must be balanced with traditional business data. ERP-based asset data is an essential element in developing a comprehensive view of asset performance. The form and format of this production data obviously looks completely different than the time-streamed scalar values that represent the feeds and speeds that come directly from the asset and IoT. The difference in data format does not diminish the importance of considering all aspects of production to deliver complete understanding. All of this disparate but relevant data must be considered as manufacturers turn to machine learning technology to drive true predictive maintenance. The sensor data for any given time period can only truly be evaluated in the context of the production order that is being processed during that time. Together, advanced data analytics and machine learning can embrace this “lumpy” data set.

Data is the Foundation

Data continues to be the critical foundation of understanding and potential improvements. Legacy techniques based on batch export and import activities to a centralized and heavily indexed data warehouse need to be reevaluated. Data storage and evaluation tools need to evolve to match the dynamic data collection and query requirements brought on by Industry 4.0.

Industry 4.0 is still in its early stages and will intersect multiple advanced technologies as it grows — data lakes are only one of them. For more information about the role Industry 4.0 will play in manufacturing and ERP, read this recent white paper.