Factors Influencing Silica’s Effectiveness in a Passenger Car Radial Tire Tread
Precipitated silica increases tire wet traction while decreasing rolling resistance when used as a direct replacement for carbon black in the tread. In the lab, performance is predicted from the temperature-dependence of the tangent delta curve: 0oC for wet traction and 60oC for rolling resistance. Lab abrasion is not sufficiently accurate to be a predictor of in-service tread life. For abrasion resistance, silica needs to be dispersed to a similar extent as is carbon black during compound mixing, but also needs to chemically react with the bifunctional organosilane, called hydrophobation. Coupling of the treated-silica to the elastomer backbone needs to occur during curing.
Results of four lab studies of PCR tread compounds are reviewed: 1. Comparing performance of Eecosil 350MG to commercial highly dispersible and easily dispersible silicas mixed in 3-pass, 4-pass and 5-pass sequences; 2. Statistical design comparing Eecosil 350MG to the EDS in SSBR vs OE-SSBR formulations along with mixing variations; 3. Adjusting silica mixing to optimize dispersion and compound performance; and 4. Measuring solid-state 29Si NMR to quantify silica surface silanols silicas to correlate to compound performance. Mixing efficiency is based on silica dispersion from SEM/ImageJ analysis to determine reinforcing aggregates vs agglomerates. The effectiveness of the hydrophobation/coupling reactions are based on maximizing the reinforcement index.
SPEAKER: Walter Waddell, Sr. Technology Consultant, Oriental Silicas Corporation
Herzlich Award Medal and Technical Presentation
Deep Learning for Visual Inspection and Classification of Tire Defects
In this paper, we propose an economical method to use rules-based and machine learning image classification techniques to automatically classify tires with visual imperfections. Using images provided by the existing tire geometry systems, tire manufacturers can prevent tires with visual imperfections from being sent to the customer. Artificial intelligence is transforming every industry, from social media to self-driving vehicles. Visual identification tasks that were once the sole domains of human inspectors are increasingly achievable by intelligent computer vision systems. These vision systems, propelled by the availability of enormous training data sets and advances in deep learning algorithms, can attain accuracy rates as high as its human counterparts Rule-based software algorithms can provide good results for identifying many imperfections, but these have an accuracy rate only high enough to be used to assist the human classifier. By adding a simple step to capture the classifier's labeled tire geometry images as they perform their normal inspection workflow, a large dataset can be created for training a deep learning algorithm to detect and label imperfections on new images. Given enough data, deep convolutional neural networks have been shown to perform extremely well on object detection tasks in images. Such a model could provide higher-quality assistance to a human classifier, or even replace the most mundane aspects of their work, freeing them up to process many more tires.
SPEAKER: Troy Anenson, President, CTI
Robotic Automation in the Tire Industry
Improved process precision, Increased throughput, and removing employees from hazardous or boring repetitive work are long cited benefits of automation. In this paper we will share advances in Robotics, vision acuity, dispense control and safety systems as they apply to tire manufacturers.
SPEAKER: Peter Shepler, Account Manager, Pioneer Industrial Systems