Skip to Main Content
 

Global Search Box

 
 
 
 

ETD Abstract Container

Abstract Header

Techniques for Real-Time Tire Health Assessment and Prognostics under Dynamic Operating Conditions

Abstract Details

2011, MS, University of Cincinnati, Engineering and Applied Science: Mechanical Engineering.
Since maintaining normal tire inflation would dramatically reduce the tire maintenance cost, which is the largest one for fleets, many researchers have invested in means of monitoring and controlling tire inflation. One of the systems is Tire Pressure Monitoring System (TPMS) which usually has the functionalities of storing the temperature and pressure data and setting up an alarm when either type of data achieves a preset threshold. This system achieves a simple condition based monitoring (CBM) by monitoring the pressure on a real-time basis. However, all tires work under dynamic operating conditions causing pressure data to fluctuate with operating conditions in a big range, so the single pressure threshold faces a tradeoff between low sensitivity and high false alarm rate. To conquer this dilemma and achieve a higher-level CBM, either a dynamic threshold or a true tire condition extracted from its superficial operating condition is needed. In this research, prognostic health management (PHM) techniques were employed to develop a systematic real-time methodology for tire health assessment and prediction. The proposed methodology could extract the current real tire condition out of its operating condition and predict its future trend based on the historical tire condition. The raw TPMS data which includes pressure and temperature was taken as the input to the model. The ideal gas law was applied on the data to reduce the temperature effect in pressure on all the tires. Then health assessment which includes individual tire self-comparison and peer based comparison was conducted. Three models (Logistic Regression (LR), Self-Organizing Map – Mean Quantization Error (SOM-MQE), Distance Based Assessment (DBA)) were employed and compared during self-tire comparison to eliminate the data variance between tires from different positions; and peer based comparison was applied to eliminate the fluctuation within each tire. Meanwhile another anomaly detection algorithm was developed to identify system changes (tire inflation, tire fixation), after which health assessment models are retrained to obtain robust outputs from health assessment models. When an initial drop was seen from the current tire health, a prediction would be triggered to infer the future health trend. Different prediction algorithms were compared in this step and no single model was accurate enough. Thus a model fusion was employed to combine single prediction models together to achieved better prediction results. Real fleet TPMS data was used in the case study to validate the proposed methodology. Two conclusions were drawn: 1. The proposed research methodology is able to evaluate the actual tire health condition on a real-time basis. 2. The tire health generated from this model was improved greatly over the outputs from previous methodologies; and the improvement could be seen from both shortened leakage detection time and higher prediction accuracy.
Jay Lee, PhD (Committee Chair)
Hongdao Huang, PhD (Committee Member)
Manish Kumar, PhD (Committee Member)
89 p.

Recommended Citations

Citations

  • Xu, S. (2011). Techniques for Real-Time Tire Health Assessment and Prognostics under Dynamic Operating Conditions [Master's thesis, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1318607916

    APA Style (7th edition)

  • Xu, Su. Techniques for Real-Time Tire Health Assessment and Prognostics under Dynamic Operating Conditions. 2011. University of Cincinnati, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1318607916.

    MLA Style (8th edition)

  • Xu, Su. "Techniques for Real-Time Tire Health Assessment and Prognostics under Dynamic Operating Conditions." Master's thesis, University of Cincinnati, 2011. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1318607916

    Chicago Manual of Style (17th edition)