Preventive and Predictive Maintenance in Automotive Manufacturing Plants
The demand for vehicles has increased almost exponentially over the past thirty years placing more emphasis on efficient manufacturing techniques whilst increasing production throughput. Until recently, preventive maintenance has been adopted to ensure minimal downtime and maximum efficiency. However, due to the complexity of modern vehicles, automotive industry is tending towards adopting predictive maintenance within its manufacturing plants to expedite production and minimise downtime. The current market for predictive maintenance within automotive manufacturing is huge and is expected to reach US$2.7 billion by 2027.
Preventive vs. Predictive Maintenance
Preventive maintenance of automotive manufacturing equipment is scheduled on a regular basis. This means the production line must be regularly shutdown therefore reducing vehicle production momentarily. Scheduling of these timely shutdowns is usually based on historical data and duty cycles of the equipment that is used. This proactive approach helps automotive manufacturers avoid equipment failures and unexpected machine downtime all whilst extending the lifetimes of the assets involved in manufacturing.
Predictive maintenance measures equipment performance using condition-monitoring technology and Internet of Things (IoT) networks. This enables connection of electronic devices to mechanical and digital machines so that large amounts of operational data can be collected. Data is gathered over time to track the condition of equipment and create models that can help avert problems and predict when maintenance might be required.
Preventive maintenance tends to result in lower start-up costs as compared to predictive maintenance, but at the expense of increased frequency. Since time and cost are interlinked, lower frequency (but slightly higher initial costs) will tend to result in lower cost over longer timescales for predictive maintenance as compared to preventive maintenance. Hence, predictive maintenance began to be favoured as technology evolved.
Benefits of Predictive Maintenance in Automotive Manufacturing
Predictive maintenance within the automotive manufacturing sector is essential, considering the value it adds to the industry from a time and cost perspective. The main benefits are:
- Failures can be accurately predicted
- Asset condition is continuously monitored
- Continuous data collection is achieved
- Artificial intelligence and complex algorithms are used to determine weak points in the system
Predictive Maintenance within the Automotive Industry
Predictive maintenance in the automobile assembly industry helps to avoid conveyor downtime. The control units of the conveyor systems, for example at the BMW Group Plant Regensburg, Germany, work 24 hours a day and 7 days a week to provide data on electrical currents, temperatures, and positions to the cloud where this data is constantly analysed. Data specialists can then determine the current position, condition, and activities of each conveyor element. Data is further used by predictive AI models to discover abnormalities and pinpoint technological issues. For example, the welding guns at all BMW body shops each perform about 15,000 spotwelds every day. Data from welding guns in all the BMW plants globally is collected by specifically created software to avoid any downtime. It is then sent to the cloud to be collated and analysed with the help of algorithms. All the data is then displayed on a dashboard that can be accessed from anywhere in the world to aid with the maintenance processes.
The cutting-edge Honda manufacturing plant in Alabama, USA uses the Splunk platform (incorporating IoT, machine learning and predictive analytics) for predictive maintenance. It uses the platform to forecast machine wear and tear, safety and effectivity to assist the production of over 350,000 vehicles each year. The Splunk platform is a data platform built for expansive data access, powerful analytics and automation. After implementation at Honda manufacturing facility, unknowns and incidents were almost non-existent. The equipment failures that were previously occurring 2-3 times per week dropped to close to zero after Splunk predictive maintenance was implemented. Furthermore, through smart predictive maintenance using Big Data Analytics (BDA) and machine learning, Honda has reduced energy consumption and allowed employees to focus on higher-level initiatives.
The use of Artificial Intelligence (AI) for predictive maintenance is proving beneficial to Ford’s production operations. AI helps Ford identify when its machines would require maintenance. It also helps with ordering replacement parts. It can determine when a replacement part is required and order it without human intervention. This means Ford can employ a just-in-time strategy rather than holding replacement parts (at a cost) in its inventory. In addition, AI is being applied to vision systems that allow precise inspections of manufacturing equipment during manufacturing processes; for example, detecting scratches during painting. Finally, Ford is experimenting with the use of natural language for voice commands to communicate with the machines on the shop floor and is also deploying AI to assess what is classified as a ‘good’ or a ‘bad’ digital audio signature.
SEAT, a part of the Volkswagen group, has incorporated its ‘smart’ manufacturing techniques into the Volkswagen Group’s Digital Production Platform (DPP) Industrial Cloud which was established in collaboration with Amazon Web Services to connect all the company’s manufacturing plants. Potential incidents can be predicted days in advance using specialist algorithms. This allows the manufacturing plant to better analyse its operations and improve productivity, providing a solid foundation for future innovation.
General Motors (GM) is using monitoring and data analytics in production to improve a variety of processes, such as evaluating the quality of vehicle paint in near real-time to discover problems before vehicles reach the final inspection phase. In addition, the company is collaborating with FANUC, a provider of industrial automation solutions, to employ data analytics to predict robot failures in manufacturing plants, lowering or eliminating assembly line downtime. Furthermore, GM Brazil had its Information Technology (IT) teams create a complete warehouse management system for the company’s assembly factory in Sao Caetano, Brazil. Benefits of this predictive maintenance system included near real-time data on goods in stock, rules-based computation engines to maximise labour and equipment productivity, and error-proof material receiving, picking, and delivery. This type of system was later deployed in other GM facilities in Europe, followed by North America.
Mercedes-Benz has implemented IoT-based predictive maintenance in its Factory 56 in Sindelfingen, Germany. The foundation for this is a high-performance wireless and mobile network in which 5G technology was deployed. Furthermore, in Factory 56, carts are employed to supply the assembly stations across the board. In a so-called ‘pick zone’, these are loaded with the necessary materials for assembly, utilising intelligent picking systems. In total, 300 self-driving vehicles are in use in the factory. Big data technologies along with these vehicles are utilised to collect and analyse the resulting data to predict bottlenecks and failures in manufacturing.
Predictive maintenance has overtaken preventive maintenance in the automotive manufacturing industry. Although typically more expensive to implement in the first place, predictive maintenance results in less frequent maintenance and reduced overall operating costs due to smart technologies such as AI, big data and connected equipment, as compared to preventive maintenance techniques.
In the near future, the use of ‘digital twins’ will improve predictive maintenance further. Digital twins can be tested over a limitless number of duty cycles at faster than real-time to investigate and determine the types of failure mechanisms and when in the equipment’s lifetime they might occur. By creating digital twins, automotive manufacturing lines might never breakdown.