The phrase ballasted railroad track refers to railroad track that consists of steel rails mounted on top of support beams called ties that are perpendicular to the rails and that in turn rest on and are surrounded by a bed of small rocks called ballast. A bed of track ballast tends to change shape with time due to its lack of rigidity, due to settlement of the earth on which it rests, due to strains within the rails, and due to forces applied to the rails by passing railroad vehicles. As a result the shape of the guidance path provided by the rails tends to change and degrade with time, and maintenance needs to be performed from time to time to restore the shape of the rails. Such maintenance work is now normally performed using a large semi-automated tamping machine that is able under computer control to adjust the location of a few ties at a time (and the portions of rail that are attached to those ties) while agitating the ballast so that the ties are easier to move and so that the ballast tends to fill in around the adjusted tie locations. This invention has to do with the way that instructions for controlling the operation of such a tamper are computed.
Prior art relating to tamping ballasted railroad track is extensive. Here attention is limited to methods for calculating the lateral displacements between current and desired locations of points along a track. Such displacements are sometimes referred to as track throws.
Prior calculation of corrections to the horizontal geometry of a section of track was based on calculation of a desired location along the track for the start of each transition from tangent or circular arc to spiral and from spiral to another circular arc or back to tangent. Typically a computer program that an engineer or tamper operator uses for calculation of tamping instructions will analyze recently made measurements of track shape to estimate the locations for those transition points, may allow the engineer or operator to adjust them, and will then calculate tamping instructions aimed at bringing the track shape closer to a form with transition points as specified. See, for example, U.S. Pat. No. 5,012,413, entitled “Railroad track curve lining apparatus and method”.
The method of this invention calculates a target track shape composed of tangents and arcs plus spiral transitions that fit together with both compass bearing and curvature continuous with distance and that together constitute a shape whose lateral displacements from points on the existing track are as small as possible with respect to a selected measure such as the root-mean-square.
The method of this invention also includes calculations to construct the existing shape of the track from so-called chord offset measurements when measurements of the existing track are in terms of such offsets rather than in terms of surveyed coordinates of points along the track with respect to a fixed coordinate system.
1. Obtaining an Estimate of an Existing Track Shape.
For an application in which the existing shape of a section of track is known in terms of measured offsets rather than in terms of surveyed coordinates at points on the track it is necessary to obtain estimates of coordinates of points on the track by processing the offset data. The following paragraphs outline how this estimation can be performed when the distance between track locations at which successive offset measurements are made is close to one sixth the length of the chord used for measuring the offsets. In this case the measured offset values can be separated logically into two sets that are relatively independent of one another. If the distance between track locations at which successive offset measurements are made is close to some other fraction of the length of the chord, then the number of relatively independent sets into which the offset values separate will generally be different, and the details for averaging the results obtained from extrapolations of values of the independent sets are adjusted accordingly.
Referring to
Then, extrapolation is initialized by placing point a at an arbitrary point in the xy plane and placing point y so that a main circular arc through points a, x, & y has an arbitrary orientation. Point c is located along that main arc based on arc length from point a along another initializing circular arc through points a, b, & c whose curvature is varied slightly with respect to the curvature of the arc through points a, x, & y and which serves to locate point b in a manner that will be explained. Point b is located midway between points a & c on this other initializing arc. Point d is tentatively located on the main arc an arc distance chord_length/3 beyond point c.
As an illustration of how the extrapolation progresses, the front of the chord is then moved to point b and the chord is rotated so that a new main arc with curvature corresponding to the second offset passes through point y. The location of the back of the chord is recorded, point d is adjusted to be the mean of its tentative location and the point an arc length 2*chord_length/3 beyond point b on the new main arc, and point e is tentatively placed an arc length chord_length past point b on the new main arc.
The above steps are placed inside a loop in which the curvature of the initializing arc governing points b & c is varied in a Newton search procedure to find the positioning of point b such that the disorder of the extrapolated line is minimized. This is necessary because the measured offsets do not give direct information about the placement of point b comparable to the good information that they give about the relative locations of other points within the set.
The above steps are carried out independently for the odd numbered offsets and for the even numbered offsets with a small curvature increment added to all the odd numbered offsets and subtracted from all of the even numbered offsets to further help prevent discord between the two independent sets of point coordinates extrapolated from the even and odd numbered offsets. Then the above steps are repeated inside an iterative search loop that varies both the amount of curvature shifted between sets and the positioning of the points extrapolated from the even numbered offsets relative to the points extrapolated from the odd numbered offsets in order to find the arrangement that minimizes the disorder of all the points taken together.
2. Fitting Target Geometry to an Existing Track Shape
Once the existing track shape is known either via survey or via estimation from measured offsets as exemplified above, the way that this invention determines the target geometry can be illustrated for the case of a single arc bounded by neighboring spirals bounded in turn by sections of tangent track as follows. The target geometry can include either traditional linear spirals or any of the several types of spiral that have been proposed as improvements. The target geometry is then specified by two parameters each for locating the two tangents (for example the compass bearing and depending on the value of the bearing either the x or y intercept) plus three parameters for locating the curve (two to locate the center and one for the radius) so that there are a total of seven parameters to be determined. (A spiral of a specified form is determined by the curvature of the curve and the offset between the curve and the tangent so that a traditional spiral or a simple improved spiral does not offer an additional parameter.) The parameters of the target geometry can be found by a simple Newton type iterative search process in which each parameter is varied to find the configuration that minimizes a measure of the lateral track throws needed to move the track from the existing to the target shape. The foregoing procedure can be extended to cover progressive and reverse curves that include multiple arcs connected by spirals.
Once the target geometry has been calculated the lateral track throw from each extrapolated point on the existing alignment to the corresponding point on the target alignment is obtained by a simple trigonometry and can then be used for calculating instructions to a track lining machine.
It will be understood that various changes in the details which have been herein described and illustrated in order to explain the nature of this invention may be made by those skilled in the art within the principle and scope of the invention as expressed in the following claims.
This application claims the benefit of U.S. Provisional Application No. 60/565,666, filed Apr. 28, 2004.
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/US2005/014749 | 4/28/2005 | WO | 00 | 10/23/2006 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2005/104789 | 11/10/2005 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
3732827 | Anderson | May 1973 | A |
3905568 | Watanabe et al. | Sep 1975 | A |
3939777 | Moran | Feb 1976 | A |
4323013 | Theurer | Apr 1982 | A |
4693183 | Pötzsch | Sep 1987 | A |
4860666 | Smith | Aug 1989 | A |
4915504 | Thurston | Apr 1990 | A |
5012413 | Sroka et al. | Apr 1991 | A |
5791254 | Mares et al. | Aug 1998 | A |
5988519 | Jordie | Nov 1999 | A |
6257494 | Tokuoka et al. | Jul 2001 | B1 |
6347265 | Bidaud | Feb 2002 | B1 |
7027966 | Klauder | Apr 2006 | B2 |
20010010197 | Kassab | Aug 2001 | A1 |
Number | Date | Country |
---|---|---|
WO01 98938 | Dec 2001 | WO |
WO03083214 | Oct 2003 | WO |
WO2004097114 | Nov 2004 | WO |
Number | Date | Country | |
---|---|---|---|
20070225877 A1 | Sep 2007 | US |
Number | Date | Country | |
---|---|---|---|
60565666 | Apr 2004 | US |