Directly after a car has been bought, an additional set of ongoing costs kick in. Owning a car will face the car owner and fleet manager with significant additional costs. Year after year. AAA research for example they have been publishing on the annual cost of car ownership since 1950. But one can find more of them – sometimes showing local/national differences due to taxes, insurance costs and fuel prices. However, the categories and their relative sizes are mostly very similar.
Typically, these following classes of ongoing costs are defined:
1) Repair & Maintenance
In this series of blogs, we will zoom into the first four cost categories one by one.
At the end of the day, these costs can be expressed in a “cost-per-mile”. “Cost-per-mile” is relevant to the individual car owner, but even more to the owner of a fleet who must manage these costs in order to remain profitable in their business.
In this first blog on “cost-per-mile”, we will zoom into the Repair & Maintenance costs. As the technical quality of cars has been improved significantly over the last decades, the costs for repair and maintenance is still significant: 5 cent / mile.
With new technologies that give us the ability to “listen to the car”, data scientists and automotive experts are able to create new solutions to become more efficient. These newly enabled efficiencies allow to reduce the cost-per-mile cost.
What is Predictive Maintenance?
What are these new technologies? Most of them evolve around data from cars. For quite some time cars have been packed with hundreds of sensors measuring velocity, steering wheel angle, engine temperature, RPM, vibrations, etc. This progress made it is possible to: 1) extract the data from the car; and 2) interpret that data and draw conclusions from it. As an example: if engine data of thousands of cars has been collected over a long period and this data is mapped with engine failure or damage (reported by owner or garage), data scientists are able to distil specific patterns that preceded that failure or damage. By knowing these specific, prior patterns engine failure can be predicted. When one can predict engine failure one can also prevent it.
The latter is an example of what falls under “Predictive Maintenance”.
With the ability to efficiently collect data from cars, the potential of Predictive Maintenance grows rapidly. Collecting data requires processing, storing and transmitting. Which - on its turn - requires more hardware (to compute and store) and higher mobile network costs (to transmit). In the abundance of car data coming from hundreds of sensors - some of them producing high frequent signals - it becomes essential to process data in a highly efficient manner. The faster and cheaper car-data is processed, the more data can be analysed at lower costs. The higher the quality of data captured in the car, i.e. identifying the potential event leading to the breakdown, the better the predictions will become. The better the predictions, the bigger the positive impact of “Predictive Maintenance”.
Exactly this is what Teraki enables. The Intelligent Edge Processing software of Teraki makes it possible to extract 10X more data from the car without increasing the costs. Moreover, Teraki’s Intelligent Edge Processing does this without loss of information as required by predictive maintenance machine learning models. An additional characteristic is that customers keep the ability to relabel the data if required. This overcomes the key challenges to extract the required data in an efficient and accurate manner. The first challenge evolves around the costs of doing this at scale. The second challenge evolves around the quality of the prediction, the so-called accuracy rate of the algorithms that make up “Predictive Maintenance”.
Summarized: Teraki makes it cost efficient to collect data for Predictive Maintenance and – more importantly – increases the accuracy of the predictions, i.e. lowering the percentage of times that the predictions are not correct. The actualization rate of the models is also optimised. With that we mean the ability to update the models by collecting data from 10x less miles driven while maintaining the predictive power of the trained algorithms for predictive maintenance. Normally data collection campaigns are expensive and require several weeks of data collection over a full fleet. With Teraki this can be done in 1/10th of the time.
In this blog we will first zoom in on the predictive maintenance of trucks and how it will save costs. Although the first example is geared around trucks its principles can easily be transposed to cars. We end with the maintenance costs for a regular car owner and how much these can be reduced.
What can Teraki-enabled predictive maintenance offer?
Correct maintenance predictions lead to better planning and preparation. A larger number of correct predictive maintenance alerts will therefore lead to a higher utilisation degree of the fleet (i.e. less idle time) as well save on maintenance costs; and prevent customer disappointment & lost revenues. By its very efficient way of collecting the required data but particularly of its high prediction accuracy, Teraki’s technology delivers better Predictive Maintenance results and therefore a better business.
With Teraki’s Intelligent Edge Processing, better planning and utilisation of the truck for the fleet owners and service providers are achieved. The idle time and down-time of the trucks can be easily reduced with timely alerts on required repairs for the trucks in the fleet. This reduces the service cost by roughly 50%1. With real time data the fleet owner can make advanced decisions to understand which trucks need maintenance at what point of time, hence increasing the overall productivity. With Teraki’s data extraction technology the user will benefit with improved control over maintenance time and costs.
This leads to improvements of 35% in predictive accuracy enables to save up to $ 5002. Another interesting insight is that an out-of-service truck costs the fleet owner roughly $ 850- $ 1.0002 per day.
Teraki’s tech delivers higher accuracy of predicting repairs or maintenance in time. These concrete benefits of less idle time and lower repair costs are then easily quantified. Next to these benefits, there are benefits that are harder to quantify but that still are very relevant: customer satisfaction, higher NPS, reducing cascading costs and on-time performance (SLA’s).
Predictive Maintenance for private cars.
The above table shows the savings in maintenance cost per year. Roughly $ 300 can be saved by using telematics. This is a clear case of monetizing the data and helping customers save money on their private vehicles. As the automobile world gets more connected and more sensor data is generated the direct beneficiary of this is going to be the customer.
As discussed above maintenance can be a quick tire change/oil change/break repairs such maintenance can be speeded up with real time data generated from the cars and hence predictive maintenance can play a big part in less down and cost savings. Not to forget the role AI and ML will play in predicting maintenance.
In this first part of this series of blogs on “cost-per-mile” we talked about the annual maintenance cost. Technology enabling predictive maintenance can directly lower these costs.
For now, we hope we have made a start in explaining that data is going to change the way cars are operated and managed. Using data can also lead to lower insurance premiums and costs. In the next blog of this series we will relate crash detection and driver behaviour to reduce insurance costs and visit the insurance part of the car’s annual cost and moreover actively help to improve the driver’s driving style.
If you want to stay updated with the latest tech news and research on the connected vehicle and autonomous driving, then feel free to sign up to our fortnightly newsletter by clicking here. Sign Up now