By: Sundar Ganapathi, Chief Technology Officer for Automotive, Tata Elxsi
Today’s vehicles are no longer just mechanical machine with few sensors and electronics control units. Over the last ten years or so, developments in vehicle connectivity, software, and data analytics have gradually transformed vehicles to software intensive intelligent devices capable of continuously assessing their own condition and performance. With the growing software content of vehicles, the volume of vehicle operation-related data has grown substantially, opening new doors to improving vehicle reliability in the transportation sector.
This transformation has altered the approach of the transportation sector toward vehicle maintenance and design validation. Traditionally, vehicle manufacturers have counted on reactive approaches to vehicle reliability and maintenance. However, with the help of artificial intelligence and digital simulation technology, vehicle manufacturers are now able to use predictive analysis to identify vehicle complications and optimize vehicle performance.
Predictive Maintenance Models: Anticipating Failures Before They Occur
Modern automobiles use a vast network of sensors to monitor performance and environmental factors. In many circumstances, a single car may have 70 to 100 sensors that continually collect data on engine performance, temperature changes, braking patterns, vibration levels, battery health, and exterior driving conditions.
Meanwhile, car links have exploded. More than 400 million automobiles around the world now have some form of connectivity and more than 60% of the vehicles sold today are linked. The growing web of connected cars is generating huge amounts of data on how vehicles actually perform in the real world.
The influx of data has arrived alongside major advancements in artificial intelligence and machine learning technology. Sophisticated AI algorithms can now sift through large amounts of data and identify patterns that traditional analytic methods might miss.
Predictive maintenance models use this skill to monitor vehicle performance data and operating signals. Over time, they learn to recognize early warning signs of component wear or abnormal behavior. For instance, slight changes in vibration, heat distribution, electrical load or energy consumption may suggest that a component is close to failure.
By recognizing these signals early, predictive systems can predict the remaining service life of certain components. Rather than waiting for an engine to fail, maintenance crews or vehicle owners can pre-empt possible failures before they happen. The shift to predictive, rather than reactive service, improves reliability and saves on costly and wasteful maintenance.
Digital Twin Analytics: Closing the Loop Between Design and Operations
In addition to predictive maintenance, the creation of digital twins is transforming the way that automobiles, and their parts, are designed and tested. The digital twin is a virtual representation of a physical part or system that an engineer can use to assess how it behaves in various operating environments. Simulation techniques have long been used to determine mechanical component structural strength, endurance, and stress levels. These simulations help engineers to understand how components may behave in real-life scenarios.
Recent developments in simulation technology have expanded the range of digital twins. Now, engineers can develop accurate digital models of complex components, such as electric motors, battery systems, integrated circuits, and electronic control modules. These models can simulate electrical, thermal, mechanical, and even chemical properties.
However, simulations cannot capture every real-life scenario. Driving conditions differ dramatically depending on how often a vehicle is used, the environment it is operating in, and the driving patterns of its occupants. Real-life data gathered from connected vehicles is extremely valuable in these conditions.
Engineers can use operational data and digital twin models to improve and increase the precision of simulations. Machine learning algorithms can analyze data and predict how components will behave over time.
This creates an ongoing feedback loop between design and real-life operations. Engineers can assess whether components behave as intended once vehicles are in operation and make better-informed decisions in subsequent iterations of the design. In some cases, this data can also be used to make changes while the vehicle is in operation.
Operational Efficiencies: Minimizing Downtime Through Intelligent Diagnostics
Greater availability of real-time vehicle data increases diagnosis and maintenance management.
Predictive diagnostics in passenger vehicles can enhance the whole ownership experience. Drivers can receive alerts of probable trouble in advance, instead of experiencing sudden failures. This allows them to plan for repairs in advance and avoid unnecessary interruptions on a road trip.
The benefits of predictive diagnostics are more evident when it comes to commercial transportation. Fleet depend heavily on vehicle availability and performance. Unplanned downtime will upset logistics, delay shipping and lead to loss of money.
Connected vehicle data and predictive analytics enable fleet operators to continually monitor their truck’s health. That means if a truck part is wearing down, they can plan maintenance to happen at the most convenient time for them – during normal logistics planning or a scheduled break.
This proactive planning reduces unscheduled repairs, boosts fleet availability and enhances operational efficiency. The prospect of being able to forecast maintenance across large transportation networks could bring about huge gains in reliability and cost savings.
Sustainability Outcomes: Optimizing Performance Across the Lifecycle
Aiming for larger sustainability goals, predictive vehicle health solutions are moving beyond operational benefits.
An increasing trend is the use of predictive information for optimizing software-assisted automotive systems. With the number of software-assisted features in automotive increasing, performance parameters can be altered based on the operating data. Such modifications can help to increase energy efficiency, reduce wear-out of components, and improve overall system performance.
Digital twin technologies can also provide more accurate estimates of the remaining useful life of vehicle components, with components being replaced according to maintenance schedules designed to avoid unexpected failures.
Predictive analytics can support maintenance decision-making based on the condition of components rather than calendar time. This means that components can often be used for longer durations without compromising safety or performance. Therefore, fewer components are prematurely rejected, which results in less material waste and more efficient use of resources.
This functionality is especially critical in the era of electrification of vehicles. High-value components such as battery systems and power electronics are large capital expenditures in both material and energy investment. Knowledge of the degradation of these systems over time can enhance their longevity and lead to vehicle operation that is more sustainable.
Toward a More Intelligent Mobility Ecosystem
The transportation industry is rapidly moving toward a future in which cars are constantly monitoring their own performance. Combining sensor data, connectivity, artificial intelligence, and digital twin technology is the key to unlock smarter vehicle health management.
Predictive maintenance models that predict likely failure modes years in advance, digital twins that better explain component performance, and advanced diagnostics that minimize downtime in both passenger and commercial vehicles, improving the sustainability initiatives of OEMs.
The shift to more networked and software-based vehicles will require a greater ability to understand and apply operational data. Predictive vehicle health is a critical component of a more dependable, efficient, effective and sustainable transportation system. With Sustainability as a strong focus, artificial intelligence becomes the anchor to achieve this goal and enhance the transportation lifecycle.