Automotive drive shaft design with virtual load estimation

Assure that a drive shaft will not break during the vehicle lifetime

What is an automotive drive shaft (or propeller shaft)?

In any ground motor vehicle on 4 or more wheels, the energy generated by the thermal or electric motor is transferred to the wheels via the driveline components which may include the transmission with several reduction stages, a drive shaft, a differential, and side shafts. The drive shaft – also called propeller shaft – is the link between the transmission out and the differential in. The side shafts are transferring the power from both end of the differential to the wheels. Drive shaft and side shafts are key elements of the complete drive train. Drive and side shafts are subjected to strong mechanical loads, especially when there is a change in load: for example, when driving downhill and then uphill, and each time you let off the gas again after accelerating.

A drive or side shaft is constituted of two or more sub-shafts connected to each other with a joint like universal joints (U-joints), constant velocity joint (CV-joints) or Tripod joints. The joints allow relative movements between the sub-shafts, so that drive and side shafts can move up and down by suspension deflection or while steering. Thus, power can be transmitted when the drive or side shaft is not in a straight line between his two end points. Rear-wheel-drive vehicles have usually universal joints at both ends of the drive shaft (so-called cardan shaft), because the movement of the sub-shafts is very limited. Side shafts have have usually a CV joint at one end and a Tripod joint at the other end, in order to allow bigger working angles and also compensate for movements in axial direction like during steering.

Universal joints are usually made of solid forged steel rather than stainless steel. The sub-shafts themselves can be made of steel or of composite materials, like carbon fiber. For instance, drive and side shafts manufactured with composite materials are widely used in racing applications. For high volume applications, composite materials are still too expensive.

The durability question: How long does this drive shaft last?

In the automotive industry and beyond, everyone is wondering: “Will the drive shaft break?” If you turn this question the other way round, automotive manufacturers and TIER 1 suppliers need to know, how they can assure that a drive shaft will not break during the vehicle lifetime.

Ten years ago, a vehicle development cycle in the automotive industry typically had a duration of about 6 -7 years, whatever the manufacturer. Within this time period, the corresponding R&D department would get in touch with suppliers for drive and side shafts and send them a requirement specification based on the planned future vehicle data and an internal calculation formula to estimate the approximate loads on the given shafts. For drive and side shafts, loads are the combination of the transmitted power (torque and rotating speed), where the torque has a preponderant role, and the working angle(s) between the different sub-shafts. Assuming that the lubrication of the joints is correct, the fatigue and thus the durability of drive shafts, and in particular of the joints, is directly and mainly related to the number of times (so-called load cycles) that a given load is applied (i.e. give torque at a given rotating and by a given working angle).

At this point in time, 5 to 6 years before start of production (SOP-6 or SOP-5), there is of course no vehicle available, not even a draft of a prototype. The TIER 1 supplier would then design a propeller shaft or side shaft, produce the first parts and test them according to the manufacturer’s specifications. Everybody is then crossing the fingers that the specifications will really reflect the real loading conditions of the drive or side shaft in the future vehicle.

A couple of months to one year after the requirements are specified, the manufacturer typically receives the first real pieces of drive or side shafts. They are then mounted in very early prototypes, for which endurance tests are started over tens of thousands or hundreds of thousands of kilometers at certain driving conditions like speed and acceleration behaviors, on test tracks with or without potholes. These tests can take another year, depending on the targeted mileage.

It can happen that the estimations at the beginning are inaccurate, and that one of the shafts in the prototype vehicles brakes or is too much worn. Why does it break? – Difficult to tell because there is usually no consistent and continuous data recording related to the loading conditions of the shaft and joints. In this case, another R&D cycle starts with updated specifications, if possible without much additional weight, and again the fingers crossed that this time the load estimations are closer to reality.

As long and frustrating this process may be, it was still acceptable as long as the vehicle development as a whole took 6 years and more. In the current decades however, this process has been drastically shortened by all major manufacturers in the world, by trying to use more digitalization and simulations in many areas of the development process. The overall timeline has shortened, while the complexity of modern powertrains is ever-increasing. However, traditional load estimations for vehicles with a combustion engine and a 5-gear transmission can not necessarily be transposed to drive and side shafts and joints for hybrid and electric vehicles. For a lot of OEMs, the above described traditional process is still the current state-of-the-art, which is a source of stress due to uncertainties and ever-increasing time pressure.

Today, an unplanned R&D cycle of 1-2 additional years because of an incorrect specification would jeopardize the overall timeline. This is the reason why different players in the automotive industry, manufacturers as well as TIER 1 suppliers, tell us that it is so important to know the load estimation of the drive and side shaft and of the related joints at an early stage of the vehicle development process. The necessity for reliable load estimations is clearly shared by both OEMs and TIER 1 suppliers.

If you wish to dive deeper into our methodology, check out our collection of publications.

Read the customer case study below:

Automotive drive shaft design with virtual load estimation | COMPREDICT

Case study: virtual load estimation of a side shaft and its related joints

One of our customers, a major European car manufacturer, has contacted us regarding this use case: load estimation of side shafts and their related joints.

Step 1 – vehicle simulation

The first step in our project with them was to simulate the vehicle before it physically exists. The simulation is based on standard parameters that were already available during the early stages of development, like the vehicle mass, the engine performance curve (max. torque along engine speed), the reduction ratios etc.

Simulating the loads of a drive shaft requires a deep dive into mechanics and data science: understanding the interactions between the shaft, the joints, the transmission and the road connection of a vehicle, analyzing the vibration phenomena and the natural frequency (or frequencies), and understanding which effect is contained in which data.

With simulation, we can determine the working angles and the torques of the shaft, the CV joints, cardan joints and or Tripod joints. The major contributor to the loads is the torque. Regarding the working angles, they have also a great impact on the fatigue and the durability (the higher the angle, the lower the lifetime). At the front axle, the steering behavior is the main contributor to the working angle, whereas at the rear axle, the influence of the suspension is dominant. To cover all the influence factors, we have created virtual sensors for transmitted torque and joint working angles incl. suspension displacement and wheel angles.

Step 2 – quality check of the simulation

In parallel to the theoretical data, we have received load measurement data of other vehicles of this manufacturer, for example vehicles of a previous model generation but also different car models, equipped with hardware sensors during endurance tests on a track. For example, the vehicles have been equipped with suspension displacement sensors and strain gauges on the side shafts to measure the transmitted torque.

The data from the test track allowed us to compare the output of our simulation environment with the measurements from the real hardware sensors. We could check the quality of our simulation models, and demonstrate their reliability to our customer in various configurations.

Step 3 – virtual load estimations with a high level of confidence

After this quality check, our virtual sensors have been used for a precise load estimation of side shafts and joints for a vehicle that does not exist yet. Our customer has integrated the data resulting out of this project in his requirement specifications for different component suppliers. Together, we could provide load estimations with a high level of confidence, for a complex powertrain architecture.

The first prototype vehicles are now ready, with side shafts and joints that have been built according to the precise load estimations of this project.

What’s next on the road to digitalization? | Connecting endurance vehicles

To tackle the next step of digitalization, we are now going into the direction of an end-to-end approach from semi-physical simulation to data-driven simulation, going through continuous endurance test monitoring. For that, we now deploy our solution on a fleet of endurance test vehicles:

At first, one or more vehicles are equipped with hardware sensors for the suspension displacement and side shaft torques, from which we are also gathering CAN signals (Controller Area Network signals). While these vehicles are running a few hundred kilometres of endurance tests, we are training our virtual sensors based on CAN-bus signals so that they reflect the measurements of the hardware sensors.

When this training phase is completed, the virtual sensors are deployed in the cloud so that they can be applied to a larger test fleet for the whole duration of the endurance tests, at least a hundred thousand kilometres per vehicle, without the necessity to equip the vehicles with hardware sensors.

Our customer will gain extensive endurance data over a long time and for several vehicles, and thus enhance the statistical basis of their tests.

We will use this data in parallel for continuous and data-driven improvement of the the simulation model for shaft and joint load estimations of pre-development vehicles. Our initial model will get ever more precise, for more and more different vehicle models

And that’s it? | AI-based simulation

With this end-to-end approach, we are now creating the foundations to enable the training and deployment a purely data-driven simulation approach… We are currently investing R&D on this topic: Stay tuned, more to come in the next months!

Related use cases

What is the durability of a truck drive or side shaft, of a composite drive shaft, or of the front or rear differential on a 4×4 truck?

Drive and side shaft design based on virtual load estimations can easily be applied to heavy vehicles like 4×4 trucks, buses or coaches, heavy trucks, mining vehicles, construction vehicles etc. As well, the approach can be applied to other component and systems like differential transmission.

De4LoRa project: virtual load estimation for an 800V hybrid powertrain

We are providing similar load estimations within the De4LoRa (Double E-drive For Long Range), a project for an innovative powertrain concept conducted by Vitesco Technologies as coordinator, TU Darmstadt and other industrial partners.

Let’s analyze together how your shaft is spinning!


+49 (0)6151 3844614


Rheinstra├če 40-42

64283 Darmstadt, Germany