Wheel deformity warning system

1. A deformity warning system for a wheel of a rail vehicle, the system comprising:

a first sensor located adjacent to a first side of a rail to provide real-time data associated with a wheel of a rail vehicle passing over the first side of the rail;

a second sensor located adjacent to the first side of the rail to provide real-time data associated with the wheel of the rail vehicle passing over the first side of the rail; and

a controller in communication with the first sensor and the second sensor to receive the real-time data from the first sensor and the second sensor,

wherein the controller is configured to:

receive historical failure data associated with a deformity of the wheel, wherein the historical failure data occurring within a predetermined time period is labeled;

determine a model based on the labeled historical failure data;

determine a plurality of features associated with the real-time data from the first sensor and the second sensor;

calculate a score by comparing the plurality of features with the model, and

determine whether the score breaches a threshold, and if so, transmits an alert to an output device.

2. The system of claim 1 , wherein the first sensor and the second sensor indicate a level of pressure asserted by the rail vehicle while traveling on the rail.

3. The system of claim 2 , wherein the first sensor comprises a bank of gauges placed at different locations along the first side of the rail to measure vertical and horizontal forces exerted by the wheel.

4. The system of claim 2 , wherein the level of pressure asserted by the rail vehicle is indicated in pound-force per square inch.

5. The system of claim 1 , wherein the deformity comprises a wheel flat.

6. The system of claim 1 , wherein the historical failure data is combined with the real-time data from the first sensor and the data from the second sensor to form combined data.

7. The system of claim 1 , wherein the plurality of different features comprises one of a look behind period, a vertical force ratio, a speed corrected force, and a haulage corrected value.

8. The system of claim 7 , wherein the look behind period is determined by the following formula:

recentmin ( i, m )=min( w 1 avg,m , w 2 avg,m , , , w i avg,m )

where,

i refers to a time horizon that will be calculated;

m is the wheel or axle number;

w 1 avg,m is an average force for a current time period;

w 2 avg,m is an average force for a first previous time period; and

w i avg,m is an average force for an i th previous time period.

9. The system of claim 7 , wherein the vertical force ratio is determined by the following formula:

vertfcratio( m )= w 1 peak,m /w 1 avg,m

where,

m is the wheel or axle number;

w 1 avg,m is a peak force value; and

w i avg,m is an average force value.

10. The system of claim 7 , wherein the speed corrected force is determined by the following formula:

SpeedCorrectedLoad( m )= w 1 avg,m −β*speed

where,

m is the wheel or axle number;

w i avg,m is an average force value; and

β is a constant that is associated with speed.

11. The system of claim 7 , wherein the haulage corrected load is determined by the following formula:

HaulageCorrectedLoad( m )= w 1 avg,m −γ*haulage weight

where,

m is the wheel or axle number;

w i avg,m is an average force value;and

γ is a constant that is associated with haulage.

12. The system of claim 1 , wherein the model is created by a classifier.

13. The system of claim 12 , wherein the classifier is associated with logistic regression.

14. The system of claim 13 , wherein the logistic regression models a probability distribution of a binary dependent variable.

15. The system of claim 14 , wherein the classifier comprises a support vector machine.