Extending the Horizon with Manufacturing Intelligence
Paving the path towards Industry 4.0 requires building the manufacturing digital capabilities in order to turn the ever-growing data into sound and accurate actions. This is where manufacturing intelligence comes into play…
The digitization of the manufacturing sector accelerated by the advent of emerging and disruptive technologies (such as cyber physical systems, 3D printing, cloud computing, big data, etc.) is gaining in importance and popularity in Europe. This undergoing digital transformation is commonly referred to as the fourth industrial revolution Industry 4.0 also known as Industry 2025 in Switzerland and is considered to be the next era of global growth and innovation.
Inspired from W. Wahlster, DFKI GmbH
In order to capture the potential of Industry 4.0, manufacturing companies need to build the digital capabilities for capturing, tracking, measuring, integrating and analyzing manufacturing related data as a valuable business asset for value creation. Manufacturing Intelligence is a key enabler to build these capabilities.
The goal of Manufacturing Intelligence is to improve manufacturing performance by bridging business and manufacturing systems from the shop floor to the top floor, enabling thus an integrated and smart management of product information across the entire lifecycle.
For manufacturing intelligence to be enabled, three pillars are to be considered. First, solutions for real-time data capturing have to be acquired in order to record production and process data. Second, integration strategy has to be carried out in order to federate data coming from scattered data sources. Third data analytics and deep learning techniques need to be applied in order to make a smart use of collected data.
The key dimensions of Manufacturing Intelligence
Infrastructure for shop floor real time data collection
Different industries require different automation instruments, devices, systems and robots for their manufacturing processes and yield different outputs (temperature, pressure, weights, speed, downtime, etc.). When recorded, these outputs help in monitoring quality, resource consumption, production status and other key indicators and offer a rich ground for value creation. Analyzing the current state of all systems helps in defining to which extent existing installations such as PLC (Programmable Logic Controllers) or DCS (Distributed Control System) cover a set of equipment or the entire facility and when new installations have to be acquired.
Solutions for manufacturing IT-landscape integration
Manufacturers in a variety of industries depend on the capabilities of different information systems. Of central importance for manufacturing intelligence is the interface between shop floor and top floor systems. MES (Manufacturing Execution Systems) are considered to be the backbone for enabling such an interface. They are meant on the one hand to manage and schedule production orders and resources and on the other hand to analyze and report on real-time production performance. MES communicate to PLCs and DCS work instructions and collect from them process and production outputs for analysis. They also communicate to ERP (Enterprise Resource Planning) information on production from the shop floor to update inventory, costs and procurement. While every system has a specific role in the manufacturing IT landscape (MES, ERP or in-house systems), their integration will secure a unified product information flow and strengthen manufacturing performances.
Advanced analytics and deep learning
When analyzing a large amount of data coming from various data sources, it comes as no surprise that some results may be omitted and left unnoticed simply because they were not assumed while setting the metrics or indicators. This is where advanced analytics and deep learning come into play. They aim to understand data behavior over time, analyze correlations and dependencies within data, identify separate clusters, extract hidden patterns and predict future events. Performing advanced analytics requires the application of a set of tools and techniques from statistics and machine learning.
With this integrated and holistic approach along with analytics and deep learning the ever-growing data can be turned into sound actions supporting thus current industries in their transition towards Industry 4.0.