General Characteristics

The shrinkage always takes place locally around the traces along which heating/welding occurs. However, due to the complexity of the task to compensate for shrinkage, it is inevitable that statistical and approximate methods must be used to some extent. AVEVA Marine has two different options for evaluation and application of shrinkage compensation, one is entirely statistical, the other supporting local application of the shrinkage to the extent possible (that means, compensation for the shrinkage locally along the trace where the heat is applied). It is also possible to combine features from the two options in certain cases.

Below, and throughout this document, the following terms will be used:

• The shrinkage compensation is applied in certain directions, therefore it is important to evaluate these directions (normally the direction is defined by the orientation of the largest number of weld traces). There may be one or two trace directions, identified as the primary direction and the secondary direction.

• The shrinkage in the direction perpendicular to the shrinkage direction (for example, perpendicular to stiffeners) is called perpendicular shrinkage.

• The shrinkage along the direction (for example, along stiffeners) is called longitudinal shrinkage. Longitudinal shrinkage compensation is always applied as a scaling factor.

The figure below shows a case where these entities are explained on panel level. This figure is especially applicable for the statistical method of evaluating the shrinkage compensation as explained below. The panel in the figure has only one shrinkage direction. Independent partitions (pitches) have been calculated between stiffeners (d) and seams (D) in the shrinkage direction.

Figure 1:1. Shrinkage definitions.

The main characteristics of the two shrinkage evaluation methods are listed below in a situation when they are operating in a fully automatic mode, that means, when the user does not take over any control.

1. The statistical method operates in the following way:

   1. The most frequent stiffener direction (considering all the stiffeners in the panel) is chosen as the shrinkage direction.

2. The most frequent partition (pitch) between the stiffeners in this direction is evaluated and so is the partition between seams in this direction.

3. For each plate the following is done:

   • Based on the plate thickness the amount of shrinkage for fillet welding is fetched from a user defined shrinkage table. By dividing this with the pitch a scaling factor is calculated by which the plate should be magnified perpendicularly to the stiffeners.

   • The same process is applied for the shrinkage caused by the butt welding along the seams.

   • These factors are combined to one which is used for scaling.

   • Based on plate thickness and the pitch between stiffeners a shrinkage factor is calculated by which the part will be scaled in the (longitudinal) direction of the stiffeners.

This type of shrinkage compensation cannot be applied to shell plates.

Plane panel profiles are multiplied by the same scaling factor that has been used for the plate (longitudinal scaling).

The method for local application of shrinkage has the following characteristics:

4. The total evaluation is on part level.

5. The evaluation of the shrinkage direction is done in a similar way as described above, however all weld traces influence the evaluation, not only stiffener traces. Moreover, depending on the trace pattern there may be two independent sets of shrinkage directions. The corresponding compensations are applied independently of each other.

6. Shrinkage compensation for butt welding is applied locally along plate edges which are butt welded.

7. Shrinkage compensation for fillet welding is applied locally along the weld trace (the outer geometry of the part is actually split, translated and then combined again). However, before doing so an analysis is made of the length of the trace compared to the size of the part. Shrinkage is applied only if certain conditions are fulfilled. These conditions can be set-up and modified by the customer.

8. When the plate part has several different plate thicknesses (which may happen in "assembly parts" for panel lines) the shrinkage compensation will consider the thickness where the stiffener is located on the part.

9. The longitudinal shrinkage compensation is applied as a scaling factor as described under 1) above. However, all weld traces are considered in the evaluation of the pitch and of the scaling factor.

10. The shrinkage compensation is applied according to the same principles for both plane panel plates and shell plates.

11. The longitudinal shrinkage in plane panel stiffeners is applied so that the change in length of the profile is identical to that of its marking trace in the plate part(s), even if the trace should pass several plate parts.

12. A special local shrinkage compensation may be applied at the ends of butt welded profiles (including shell stiffeners).

The automatic evaluation as described above may be partly overridden by manual input. Both methods use currently the same shrinkage table.

The most important advantages of the second method are:

• It makes a much more detailed analysis of the weld traces and, therefore, the result should be more reliable.

• The customer can control the conditions for application of shrinkage to a larger extent.

• The shrinkage is applied locally whenever possible which is important, especially in combination the use of the GSD option for alignment of parts for butt welding.

• The shrinkage compensation may be applied also for shell plates in the same way as for plane panel plates.

• Using the statistical method also areas with no weld traces might be subject to shrinkage compensation (since in this method it is always applied as a uniform scaling). This will not occur with the second method, at least not to the same extent.

• The shape of parts (for example, if they are (close to) rectangular or triangular) will affect the way the shrinkage is evaluated.

A disadvantage may be that the second method requires much more process power. However, since the evaluation is executed in a background process that is not so critical.