(Cancelled) Development of an Upper Control Arm Plate Using Generative Design

REYNOLD ELSEN: Good morning, everyone. My name is Reynold. I am from Vellore Institute of Technology. And in this session, I am going to discuss on development of an upper control arm plate using generative design. This use of procedures followed to develop high performance component of a formula student car. The implementation of the generative design methodology estimates unnecessary weight addition and provides a considerable enhancement in part performance by virtue of significant weight reduction.

The introduction. This particular work we have selected double wishbone suspension as it provides high stability and ensures consistent wheel alignment in any vehicle. A double wishbone suspension consists of upper control arm, lower control arm, upper ball joint, lower ball joint, push/pull rod, tie rod, and upright. Of these, an upper control arm plate is an important component in suspension assembly and has a significant effect on the unsprung weight, which directly influences vehicle handling. High unsprung mass can have adverse effect on kinematics and ride comfort of the car while simultaneously affects driver ergonomics.

The control arm plate is optimized by implementing generative design. And the results of the simulation studies conducted prior to the optimization. This particular upper control arm plate weighs 118.7 grams, which is the conventional upper control arm. And we are trying to optimize this.

What is the significance of lightweighting in today's industry? Lightweighting can improve productivity by reducing the amount of material that needs to be processed and assembled. This can lead to shorter manufacturing lead time and lower cost. Lightweighting can also help to improve efficiency by reducing weight of the vehicles and machinery. This can lead to improved fuel economy, reduce emissions, and increase payload capacity. Also it helps to optimize cost by reducing the amount of material used and by improving fuel efficiency. This can lead to lower manufacturing cost, lower operating costs, and lower environmental impacts.

Lightweighting a product can also improve its overall performance and durability. Lighter products, for example, may be able to move faster and with greater agility. Lighter products may be less susceptible to fatigue and wear, resulting in longer lifespan. It also helps to improve customer satisfaction by providing parts that are more fuel efficient, have longer range, and are easy to handle.

And of the most, lightweighting can also help to improve safety, which is really very important. It can lead to improved handling and braking performance and reduce risk to rollover accidents. And of the most, environmental impact also plays a major role. It also brings about new opportunity where one can open up with new products, developments, and innovations. For example, lightweighting components can be used to create new products that are possible before or to improve the performance of existing products.

Lightweighting techniques in designs nowadays has evolved to a greater extent and we have the various following techniques. First one is topology optimization. Topology optimization is a computational method that uses mathematical algorithms to design structures that are both strong and lightweight. It works by removing unnecessary materials for a design while still ensuring that it meets all of the required strength and stiffness constraints.

And next most recent improvement is lattice structures. They are lightweight strong and stiff structures that are made up of repeating patterns of struts and nodes. They can be designed to have wide range of properties and are suitable for lightweighting applications. And finally, we come to generative design. It is a new approach to design that uses artificial intelligence and cloud computations to generate a wide range of possible solutions to our given design problems. Generative design can be used to design lightweight structures that meets specific performance requirements while also taking into account other factors such as manufacturing and cost.

So we go on to our particular work, generative design of control arm plate. As told earlier, generative design can be used as a design exploration process. And we have used the same. It enables an engineer to create thousands of design options by specifying their design problems using AI and cloud computations. We used this technique to provide an alternate design solution to long periods spent in design space because of its ability to generate several possible outcomes in a fraction of a second.

The implementation of generative design methodology in our work eliminated unnecessary weight additions and provide considerable enhancements in part performance by virtue of significant weight reduction. In this particular generative design methodology, initially we create the 3D modeling of the component and we define the geometries.

Here we have the preserve geometry and obstacle geometry. This preserve geometry creates a component and obstacle geometry will remove the material. Finally, we go for applying the boundary conditions and loading conditions, which is the prime of any engineering problem. And then we define goals, which will be discussed in the coming slides, and solving. And finally, we will go for selecting the designs.

The constraints and objectives are defined when we go for preserve geometry and obstacle geometry. We provide the materials here. We have used titanium-6 here, 4-vanadium, and the minimum factor of safety here is 1.725. And we have used RT manufacturing for our work. We are going to use this particular component for our racing vehicle whose life cycle is very small so that gave us the privilege of going for lesser factor of safety.

Next we have the following design goals. The designers or engineers have to input design goals in generative design along with parameters such as spatial requirements. Materials, size of the component, the target weight, the strength required, the manufacturing method that is to be done, like additive manufacturing, casting, machining, and the cost constraints can be given as the design goals.

Here the problem definition is discussed. The mass of the car is classified into two types, sprung and unsprung mass. Sprung mass is the mass that is supported by the spring damper system-- that is the chassis-- and the unsprung mass is the unsupported mass of the vehicle. It is utmost important to keep the weight as low as possible. Also it is equally crucial to maintain adequate stiffness and strength. To achieve this design optimization for making the unsprung mass of the vehicle lighter can be a viable option. The suspension assembly is taken into consideration, particularly the control arm plate of the double wishbone system.

Prior to the load calculation, a coordinate system, which would be followed consistently throughout the design and analysis phase is defined. The force calculation methodology begins with analyzing the tire data. Tire decision matrix comparing three different tires on the basis of peak, lateral, and longitudinal forces, variation of pneumatic trials, load sensitivity, drop in lateral force of the peak slip angle, drop in longitudinal force after peak slip [INAUDIBLE], and packing are considered. These data is used to determine the force and momentum generated on the contact path under different loading conditions. The forces and moment values are then translated to the wheel center, which provides the resultant force on the suspension linkages.

We have considered braking force to be 1.5G, wheel cornering to be 1.5G, braking width full load transfer, cornering with full load transfer, braking cornering to be 1.4G, and 5G bump. As these force calculation does not take into account the transient state loading condition, the 5.9 bump loads cases is taken into account for immediate change in load surface height and prolonged loading scenario which provides a sufficient factor of safety to prevent the buckling of suspension linkages under compression.

The GD of the control arm plate is created by using the previous procedures. And before going for the final solution step, the previewer is used to understand the shape that the final organic body would take. Considering the location of the preserve and obstacle regions, it is a step used to ensure that no obstacle of the precise geometries are omitted. The diagram in the left hand side shows you the previewer. This helps a good understanding of how the component will be designed. The design shows an organic body connecting the preserved bodies and avoiding all the obstacle regions.

So the particular component, which is revolving on your right hand side, is the actual GD component. Initially, in the previous slides, I have given the actual weight of the upper control arm, which is somewhere around 118.7 grams. After subjecting this particular component to generative design, we are able to reduce the weight to 41.89 grams so we are able to achieve a weight reduction of 51.6%. However, it is necessary to ensure that the simulation objective of the maximum deformation and the generated [INAUDIBLE] are still met.

In order to go for further analysis, we are giving different names for the parts. So initially, we have the bearing slot, post insert, fore push rod tab, aft push rod tab, aft insert, and the organic body in the GD controlled arm plate. And on the right hand side bottom, you can see the assembled GD control arm plate.

So in order to do the further analysis, we have opted for finite element analysis. Based on various dynamic conditions, it is necessary to minimize deflection. We have chose a deflection of less than 0.5 to ensure minimal compliance to the corresponding vehicle dynamic system. It is also important to certify that the observed stresses are well under the yield strength of the material of the respective parts.

A global mesh is developed to analyze the localized stress and the deformation. The meshes are ensured to have uniform distribution with minimal deformations to ensure accuracy in the research. Based on the geometry and the features, fine tetrahedral meshes are chosen in an attempt to achieve less distorted mesh with low computational requirement. The proximity and the curvature function is chosen to ensure that the mesh at the curved and fillet regions are not distorted.

The fixed support of the plate is determined to be inner surface of the bearing slot which would be in rigid contact with the upper control joint. The loads are applied on the force and fore and aft insert of both the lower and upper plate and the push rod tab slot for the upper plate. The direction of each component of the load is carefully assigned based on the coordinate system from which the loads were initially derived. The maximum deformation and the stress generated in this control arm plates are found to be well below the allowable limits.

Greater deformation is observed on the upper plate due to additional set of forces acting on it from the push rod. Considering the bearing slot to be the fixed support, the resultant of the force on the push rod tab brings about stress concentration [INAUDIBLE] near the slots. The stresses generated are well within the yield strength of the materials and are uniformly distributed.

Now we go on to the development of the control arm plate. So we have gone for additive manufacturing of the component and we have used powder bed fusion techniques where powder particles are melted and it is continuously done one over the other for a particular product development. Initially, we create the component and then we convert that into an STL format and then it is created into gcode using a slicing software. And finally, it is imported into the selective laser melting machine and the component is fabricated.

So we have different SLM parameters for this particular component. The [INAUDIBLE] deposits a layer of 60 micron thickness powders by employing bi-directional powder coating technology. The entire printing process is conducted in an inert atmosphere filled with argon gas, two 400 watt IPG fiber laser scans the layer at a build rate of 113 centimeter cubed per hour with a process parameter followed with the printing operations. So in this particular work, SLM 280 machine was used.

After optimizing the process parameters, we went for optimizing the build orientations. We selected four orientations. Of them, we did various post-processing analysis. The results for varying the orientation and the different support structures is given in the table. Overall orientation can lead to increased post-processing requirement and can also damage the features due to poor tool passages while removing the supports.

In this particular process, we simulated the thermomechanical simulation of the building parameters and for different orientations as well. You can see how the reflections are calculated. And we were able to finalize that the orientation force allowed for reduction in bulky supports, and let several sections to be supported, and prevented the generated support from getting wedged out in narrow regions. In spite of increased build height, this orientation for is chosen as it gives quality a greater priority than the printing cost.

So this is the picture which gives you the idea of how the particular material is printed and how it gets removed. And finally, the left hand side image shows you the material, which is still attached to the plate and then, finally, the right hand side image shows you how this particular component is post-processed and it is allowed and given for final utilization.

Finally, we go for the dimensional analysis. This is the quality check phase where we try to understand how precisely this particular component is manufactured. Since this particular component has a lot of thick and slender structures, it always has an issue of deformation as this particular process, there is a lot of heating and cooling process is occurring. So in order to do this particular analysis, we opted for CT scanning, which is an image technique popularly implemented in biomedical applications for diagnostics.

Owing to its accuracy and the ability to scan complex features, CT scanning is implemented over coordinate measuring devices to develop a 3D model of printed parts for this particular dimensional analysis, also due to the presence of internal features such as bearing slots and the inner surface of the inserts which cannot be registered using coordinate machines.

In this particular work, we have used Siemens Sensation 62 scanners which scans the component with 0.65m slicing intervals at a precision of 0.02 plus or minus 0.09 [INAUDIBLE]. The entire components is divided into 752 slices, which are stored in dicom format. And it is converted into the 3D model using open source packages.

Finally, there is a comparison is done using Bohm software between the actual design and the parts printed. It is investigated by overlapping the designs with the CT scan model of the part. The dimensional analysis provides greater insights into the accuracy of the manufactured parts. The linear dimensional analysis concludes the parts to be within allowable tolerance limits with no impact on assembly. Three surface profiles are observed to be consistent with the predicted deformation, keeping aside minor distortion caused by post-processing.

In comparison with EM simulated results, the CT scan model shows a close relation in terms of the material distributions. The surface finishing operation performed after printing and support removal performed to have impact on the surface profile and deviation from the original deviations were observed in the organic regions, which is not a very big-- no deformations or deviations from the expected results.

Finally, it's really time to thank people who have really supported me. I have to thank my organization, Vellore Institute of Technology, Vellore, for providing me with the resource and facilities; the Autodesk education team, especially Mr. Dipankar, Mr. Anand Pujari and Badri for their continuous support and efforts to help me to work in this particular project; also I have to thank Mr. Aayush and Mr. Daniel Abishai for working in this project and helping me and Pravega Racing Team VIT Vellore for providing us the important information for calculating the load constraints.

We were successfully able to publish two journals in this particular domain on this particular work. And you can really go through this work, which will really help you to understand better of this particular work.

And these are the references. I really recommend you to go through these references which will really help you to understand the various additive manufacturing terminologies and concepts. I really thank everyone who has really visited this particular video for your time and hope this was an informative session. Thank you all.