Will the factories of the future be driven by artificial intelligence and advanced analytics?
With sensors everywhere on the production floor generating time-series data, discrete manufacturing and process industries are sitting on a gold mine of data. Industry 4.0 leaders and manufacturers in advanced stages of digital transformation are using this operational data, AI, and machine learning systems to uncover deep operational insights, streamline processes, and enable faster, accurate decision-making. And yet, current manufacturing practices still involve repetitive tasks and manual processes, and face multi-dimensional problems such as optimizing overall equipment effectiveness, which is often beyond the scope of traditional tools.
Predictive analytics and machine-learning-based modeling can address complex manufacturing problems—from reducing product rework to eliminating production line downtime to increasing visibility across the facility. The industry is gradually realizing the potential of AI, and the adoption of predictive applications such as anomaly detection, real-time quality monitoring, and supply chain optimization is gathering momentum. Deloitte recently reported in a survey on manufacturing AI adoption that 93% of companies believe AI will be a crucial technology to drive growth and innovation in the sector
Let’s look at three key technologies that are required for driving AI adoption in manufacturing and realizing smart manufacturing vision.
Augmented data discovery
While AI has the potential to bring dramatic benefits to the manufacturing industry, implementing AI and ML in a manufacturing setting is quite challenging. A traditional AI-powered predictive analytics workflow is not only exhausting but very technical, involving a variety of tools and a background in data science. A typical project has several steps from data preparation, feature engineering, algorithm selection, to model training, testing and validation. Any part of this sequential process can derail the project; it can take anywhere between 3-7 months to get to production-ready models.
Getting clean data in a manufacturing environment is the first big challenge, as production sites can have several CMMS, historians, ERP and manufacturing-execution systems. With disparate systems and the sheer volume of data, cleansing data and aggregating this information manually is time-consuming, expensive, and labor-intensive. Data from diverse systems comes in multiple units, different formats, and is stored in data warehouses and data lakes.
Industrial-operations teams need a scalable, secure and high-performance environment for operational-data management and analysis. That’s where augmented data discovery can help to dramatically accelerate the data-preparation process. Augmented data discovery is the process of using automation for getting relevant data ready for analytics. Unlike manual data discovery, it is scalable for thousands of datasets across millions of tables, robust at an industrial scale, and automatically connects to the right data. Once the data set is ready, it can be stored in data marts ready for AI and ML.
AI automation
Once a use case has been defined and the data aggregated, the next step is to transform the raw data into machine-learning inputs (a.k.a. feature engineering). AI projects require multi-disciplinary teams with expertise in feature engineering and machine learning. Operational-technology experts are not trained in building predictive models. Instead, they rely on data scientists and advanced analytics teams to assist them in using the latest technologies in operations.
How can manufacturers hire AI and ML talent when it is in such short supply? AI automation can reduce the cost to implement AI and also speed up the painfully slow AI deployment. By using automated ML tools, SMEs can focus on day-to-day responsibilities and leverage AI platforms with automated feature engineering and ML to enable them to build predictive models at the click of a button.
End-to-end data-science automation platforms automate up to 100% of the AI/ML development workflow, using an AI-based engine to automatically discover meaningful patterns and build ML-ready feature tables from operational data. Using a supervised ML technique, the AI-automation system should find patterns in production data, identify deviations in quality, and empower SMEs to prevent the recurrence of problems. The ideal AI-automation solution will incorporate a platform designed and built for industrial manufacturing; a system that is usable by the operations team without requiring additional data science resources.
Prediction at the edge
Smart manufacturing requires real-time prediction capabilities for use cases such as inventory stocking for just-in-time production, quality monitoring, and anomaly detection. There are many potential use cases of edge computing in the manufacturing industry. Edge architecture enables manufacturers to process data locally, filter data, and reduce the amount of data sent to a central server, either on-site or in a cloud.
Additionally, a key goal in modern manufacturing is to be able to use data from multiple machines, processes, and systems to adapt the manufacturing process in real-time. This precision monitoring and control of manufacturing assets and processes use large amounts of data; it requires machine learning to determine the best action as a result of the insight from the data, and also requires edge-based computing. The ability to deploy predictive models on the edge devices such as machines, local gateway, or server is critical to enable smart manufacturing applications.
Augmented-data discovery, AI automation, and real-time analytics with prediction capability at the edge will deliver greater agility, deeper insights, and enable faster, more accurate decision-making. Industrial manufacturing needs to embrace the tech culture to move fast, iterate rapidly, and make tech-driven innovation a core capability. Intelligent and flexible automation is a step in the right direction, and it appears that the factories of the future will, indeed, be relying on AI and advanced analytics.
Sachin Andhare is the head of product marketing with dotData.
