The manufacturing industry faces continual disruption from changing markets, new product demands and technological innovations. Whether it’s known as Industry 4.0, Smart Factory or the Fourth Industrial Revolution, the industry is seeing an increase in the adoption of digital platforms, robotics and the wide application of data.
Data science and data analytics are helping manufacturers better understand their operations and processes, and machine learning (ML) is helping improve production efficiency, machine uptime and reduce quality defects and maintenance costs.
50% of companies that embrace Artificial Intelligence (AI) over the next five to seven years have the potential to double their cash flow, according to McKinsey, with manufacturing leading the pack due to its heavy reliance on data. However, several key barriers outlined in the Made Smarter Review are believed to be preventing the necessary Industrial Digitalisation. These include limited access to innovation, perceived cost and limited industry 4.0 knowledge, amongst others.
If this is the case, what kind of issues were previously prevalent within the manufacturing industry that AI and ML set out to solve? As far as common manufacturing issues go, unplanned downtime, machine failure, poor first-pass-yield (FPY) and quality metrics are just a few examples:
Predictive Maintenance for Equipment Failure
Technical issues and equipment failures are often the cause of many disruptions within manufacturing companies. They impede performance, skew delivery schedules as well as a brand’s reputation as a result, impacting profitability. Studies show that unplanned downtime costs manufacturers an estimated $50 billion annually and that asset failure is the cause of 42% of this unplanned downtime.
By using AI and ML-driven analysis, companies can dig deeper into their equipment data to identify mechanical technical issues and trends. This insight then enables them to predict future failures and take a more proactive approach to equipment maintenance.
In this regard, AI and ML play a big part in helping businesses leverage smarter solutions to make production processes more efficient and scalable, which is achieved through monitoring and even predicting performance (TDS).
Here’s an example: An automotive OEM improved their production processes that had seen Overall Equipment Effectiveness (OEE) of the press line reach a low of 65%, which subsequently rose to 85% after the implementation of advanced data analytics. This was achieved by analysing sensor data collected from 15 operating parameters (such as oil pressure, oil temperature, oil viscosity, oil leakage, and air pressure), helping the OEM identify inefficiencies at scale and make informed decisions accordingly (Forbes).
Digital Twins for Asset Performance
Toyota is one of the world’s top car companies in the world, and its lean manufacturing management philosophy and ‘Kaizen’ (continuous improvement) has been replicated by companies worldwide. However, even manufacturers with the best design and engineering facilities can still face issues with post-production quality.
In 2010, Toyota was at the face of one of the costliest and most memorable product recalls of the decade. They were forced to recall 8.1 million vehicles because of the potential for gas pedals getting stuck in floor mats, which resulted in a loss of $3.2 billion for the car giant (Kiplinger).
AI has now made it possible to visualise production and manufacturing tasks by creating identical digital representations of them, called digital twins. These are a virtual representation of both the physical elements and the dynamics of how an IoT-connected product operates and interacts within its environment, throughout its entire life cycle (Deloitte).
Digital twins can predict failures more easily, as they make it possible to virtually see inside any physical asset, helping identify the root cause of production problems, like assembly line faults, factory defects, or supplier errors. By correlating smart factory data with digital twins, defects from manufacturing stages can easily be minimised, and by fully understanding the behaviour of production machinery, business owners can gain a better view of how product quality is being affected.
For many consumers, quality still reigns king. Customers have also come to expect faultless products, pushing manufacturers to up their quality standards in order to maintain their brand reputation in a continually changing consumer landscape.
Because of today’s very short time-to-market deadlines, manufacturers are finding it difficult to keep up with these quality regulations and standards. This is where Quality 4.0 comes into play. Quality 4.0 involves the use of AI and image recognition in quality control, using algorithms to notify manufacturing teams of production faults that are likely to result in quality issues. These faults can range from deviations in recipes, subtle abnormalities in machine behaviour, change in raw materials, and more (CIO).
Quality 4.0 allows manufacturers to collect data relating to the performance of their products, which is very valuable as it allows product development teams to make better, more informed strategic and engineering decisions based around usage-data. This results in a more accurate evaluation of product quality before it is released for purchase. Furthermore, it has been reported that machine learning can increase fault detection rates by up to 90% while slashing the time to pinpoint the root cause of quality issues to minutes from days (Seebo).
Supply chain & logistical complexities
AI and ML technologies are proving to be ground-breaking in the supply chain and logistics industry, with McKinsey expecting businesses to gain between $1.3 trillion and $2 trillion a year in economic value by using AI in their supply chains. This is because running a large supply chain comes with many complexities, especially at scale, and when AI is implemented correctly, companies can make smarter and more agile decisions, all the while anticipating problems that may arise in the future.
AI solutions make it possible for manufacturers to improve efficiency throughout their entire supply chain, which can have a positive impact on customer expectations for on-time and undamaged deliveries. Ultimately, this can result in lower costs and fewer issues across a manufacturers logistics network.
An example of revolutionary supply chain management thanks to AI technologies is the Swedish car manufacturer Volvo. Volvo uses cloud-based services and IoT tech in order to support the logistical side of its supply chain process. More specifically, their implemented solutions help order components from different countries, as well as shipping vehicles to suppliers worldwide. They’re not the only car company doing so either – Nissan have also automated a large part of their supply chain too!
Demand forecasting is a vital part of any manufacturing organisation, as it serves as a fundamental tool for effective planning and general supply chain management. However, manually stock-taking and managing inventory is a long and arduous process.
By implementing AI into supply chain management, inventory managers can more efficiently plan their monthly orders, understand seasonal trends, save time on reordering and reduce stock-outs. Above all, modern-day machine learning technologies have achieved up to 50% increase in accuracy over previous methods such as ARIMA and Exponential Smoothing methods (Medium).
AI can increase the accuracy of demand forecasts and can help companies save money by stopping inventories from lagging behind. This makes it possible for manufacturers to become more proactive, as well as helping them adjust output to match expected demand, leading to lower operational costs in the future.
Ultimately, all manufacturers have a common goal: to produce high-quality products at maximum efficiency and minimum costs. The traditional manufacturing model is evolving into what could be referred to as Factory 4.0, whereby industry 4.0 components such as non-intrusive sensors, wireless connectivity, AI and ML are all adopted in order to mitigate the common manufacturing issues above, all in an effort to improve manufacturing outcomes.