Enterprise Manufacturing Intelligence (EMI) is software that aggregates data from industrial IT systems, enabling them to be viewed, processed, and analyzed to guide and optimize operations.
This concept has now been in use for several years. Some describe it as Business Intelligence (BI) applied to industrial operations. With evolutions in IT, the emergence of Data Science, Machine Learning (AI), and IoT technology, possibilities are increasing to provide a new technological base for Enterprise Manufacturing Intelligence.
Aggregating industrial data from multiple sources
The goal is to combine data from:
- SCADA systems (PLC, DCS, Historian, etc.) – particularly including parameters of control, measurement, operating cycles and line analysis systems.
- Production Execution Information Systems (MES: manufacturing execution system).
- Quality Information Systems (LIMS, ERP, etc.) for quality control, laboratory analysis, etc.
- Financial Information Systems (ERP) with raw material purchase, standard production costs, etc.
- IIoT (Industrial Internet of things): including equipment monitoring sensors and geolocation systems.
Aggregating and combining means collecting data in one place to make it easier to manipulate. This requires storage. The challenge is finding the best method for intended uses while staying flexible. Data Warehouses or Data Lakes provide this service.
Any Enterprise Manufacturing Intelligence solution must provide a Data Lake optimized for storing and manipulating industrial data. For example, the ability to efficiently process time series as well as traceability and relational data. The existence of a pre-established data model adapted to industrial use makes implementing a chosen solution quicker and easier.
Manipulating and analyzing data to make it actionnable information or “uses”
EMI solutions can transform aggregated data into actionable information. This is certainly the area that has evolved the most. Data manipulation is no longer just indicator calculations (specific consumption, yield, productivity, TRS, etc.) and their visualization. New tools mean we can go further using analysis, data science and machine learning. It is now possible to:
- • anticipate processes: detecting errors, anticipating consumable changes, pilot recommendations, etc.
- • conduct predictive maintenance: detecting faults on equipment, anticipating faults, etc.
- • optimize processes: making decisions based on raw materials and energy prices, optimizing operating conditions, etc.
This is where the word “intelligence” in EMI is significant. More than a simple response to a need for knowledge, it provides concrete support for decision making, even automating it, based on business expertise acquired over time.
Finding and using information deep in data
Much of the information that may be useful to guide the industrial tool is unused because it is not available or visible to those could use it. The objective of an EMI is to make this information accessible and intelligible to the right user at the right time. Such solutions streamline the exchange of information in the industrial organization, removing silos that may impede effectiveness.
Solutions for uses and user autonomy
An important aspect of an EMI 2.0 project is the ability to develop skills around data and to continuously improve operations in the field. As a result, focusing on uses is fundamental for EMI digital transformation. Progress made in application developments, such as UX design, ergonomics, and user-friendliness, means using solutions is easier to learn because functions are designed to meet concrete business uses.
EMI 2.0 is finally here
EMI is now reaping the benefits of the last ten years of powerful technological revolution with the arrival of the Cloud, Big Data, Artificial Intelligence and Operational Research, among others. Answers are finally available to the questions asked about optimizing operations in factories every day. As well as providing new possible uses, these revolutions offer access at an even lower cost. As a result, EMI is no longer reserved for big groups. It is now accessible to ETIs, and is gradually becoming more accessible to SMEs.
For big groups, the dilemma is abandoning internal EMI approaches that were once suitable to opt for more mature solutions now on the market. Current market solutions are the only guarantee of staying up to date long term and focus financial resources exclusively on tasks that directly add value to business.
Jean-François Hénon, Mathieu Cura
1 – Enterprise Manufacturing Intelligence according to Wikipedia
2 – La vision de la Manufacturing Intelligence by Capgemini.
3 – Enterprise Manufacturing Intelligence study by Arc Advisory Group