Prompted by practical engineering needs, Foster Wheeler Upstream developed a design tool to facilitate the design of subsea plate mudmats. Dr Jinming Xu, Paul Schank and Ray Young discuss how Mudmatician, through automation and built-in intelligence, significantly improves engineering efficiency, reduces the potential for human error in mudmat design and offers other advanced capabilities.
Subsea structures are typically custom designed to suit specific criteria of each site. As a consequence, it is difficult to achieve a generic design. The design process, however, is usually similar and often follows the same design procedures. By automating the design procedure, added engineering efficiency can be realized.
This is especially the case for plate mudmat design that incorporates a significant number of plate elements that include the top plate, subdivided into subplates, and the longitudinal and transverse stiffening plates. Due to variation in installation requirements, soil properties and the size, weight and applied loading of the structure it is supporting, the mudmat is custom designed from site to site. The design process does, however, consistently follow the same industry engineering codes, such as API Bulletin 2V, that require a great amount of meticulous stress calculation and structural checking on each of the mudmat plates. Without automation, the design process is very lengthy and cumbersome; the overwhelming details involved during the design can also easily become a source of human errors.
The automation of this process has been developed by Foster Wheeler Upstream in the form of Mudmatician, the subject of this article.
Background
The design process for a mudmat can be divided into three steps:
I. Structural design and model building
At this step, the design layout, dimensions and panel thickness of the mudmat are selected, and a finite element (FE) model is built based on the design.
II. FE stress analysis
During this step, the FE analysis is performed to evaluate the stress distribution in the mudmat. This FE analysis is normally conducted with commercial software such as Abaqus.
III. Structural integrity check
Once the FE analysis is finished, the structural integrity can be assessed. The procedure of structural check typically follows the process as described in API Bulletin 2V.
Each of the aforementioned steps involves a large amount of engineering work, and at step III, the structural check has to be performed on each of the subdivided top plates, and on each of the longitudinal and transverse stiffeners (Figure 1). In other words, the engineer has to trace every plate in the mudmat and perform the necessary stress checks for each. Considering the irregularity and large quantity of the plates, this is very time-consuming. To help retrieve and identify the plates at step III, a consistent naming scheme for plates has to be developed. In turn, this requires that the mesh nodes and elements be numbered orderly in step I.
Commercial FE packages usually provide an interaction for the user through its graphical user interface (GUI). But unfortunately, the GUI does not allow custom numbering of plates, elements, and nodes. The GUI hides the numbering of nodes and elements from the user, rendering the user no control over the meshing process. Without a design tool like Mudmatician, all the numbering has to be implemented by hand without graphic visualization. Considering the large quantity of nodes, elements and even plates, this is very cumbersome, as one has to carefully track all the nodes, elements and plates by writing an input file to build the FE model in the FE program. With this approach, even greater care has to be exercised later in extracting stresses from each plate and performing the structural adequacy verification.
What Mudmatician does
Mudmatician aids mudmat design by providing (1) easier structural design and model building at step I, and (2) an automated structural integrity check at step III. In addition, Mudmatician offers advanced features which are often practically out of the reach of manual designers. Among these are the capability of designing a mudmat with varying depth and shape complexity, and the capability of displaying checked structural strength graphically as contours.
Mudmatician consists of two modules: the preprocessor and the structural checker (Figure 2). They provide design assistance at steps I and III respectively.
1. Pre-processor
The pre-processor builds a FE model based on the users design, written in text file *.py. The created model, written in file *.inp, will have all the nodes, elements and plates properly numbered or named, according to designated naming/ numbering schemes. Figure 3 shows the possible plate layout of a portion of mudmat top. The controlled numbering and naming prepare the mudmat for the structural check to be conducted later on at step III.
In building the mudmat model, Mudmatician employs a top-down approach, freeing the user from low-level details such as the creation of nodes and elements. These details would otherwise require undue effort by the mudmat designer. With this top-down approach, the user is only required to instruct Mudmatician as to what kind of mudmat model is to be built. This is achieved by assigning values to a handful of predefined parameters, such as the overall size of the mudmat, layout of longitudinal and transverse stiffeners, etc. Then Mudmatician is used to build the FE model with nodes, elements and plates properly numbered.
2. Structural checker
The structural checker performs the structural adequacy check on each of the mudmat plates. First, the information of each plate, such as dimensions, thickness and boundary conditions, is retrieved from the FE results file (*.odb). Second, normal and shear stresses are retrieved from nine locations on each plate based on the retrieved plate dimensions. Finally, the extracted stresses, along with the other plate information, are plugged into the appropriate equations for structural integrity verification.
How it works
Mudmatician is written in C++ with object oriented architecture. While running, Mudmatician works together with Python and Abaqus to accomplish the mudmat design. Python serves as the parser for the user design file, while Abaqus performs the actual FE analysis.
Illustrated in Figure 2 is the data flow with Mudmatician. During the process of design with Mudmatician, the designer performs the following tasks:
1. Writing the design file (*.py)
The user writes the mudmat design in a Python file. Specifically, the user assigns values to a handful of predefined parameters, such as the size and plate thickness of the mudmat.
2. Generating Abaqus input file (*.inp)
Mudmatician pre-processor reads the design file, and generates an Abaqus input file (*.inp) for FE modeling. As the mudmat design is written in Python, Mudmatician interacts with Python interpreter to parse the Python file.
3. Performing FE analysis
Abaqus takes the input file (*.inp) and runs the FE analysis, generating an output database (*.odb) upon completion. As the input file .inp is a text file, the user can easily add the loading conditions and additional auxiliary components into the FE model by editing the text file. Should the engineer prefer graphics, the user can import the input file into Abaqus CAE, and then modify it in graphic mode.
4. Performing structural integrity check
Mudmatician structural checker extracts the stresses from the Abaqus output database and performs structural integrity checks based on API Bulletin 2V. For each plate on the mudmat, Mudmatician extracts normal and shear stresses at nine key locations, and perform the stress calculations with appropriate equations.
Once the structural check is complete, Mudmatician generates a report on the structural integrity of each plate and concludes the initial design cycle.
Based on the structural report at step 4, the designer now needs to decide whether to modify the mudmat design or not. If yes, the designer now needs to go back to step 1 for another design cycle.
Other advantages/features
As with most engineering practice, the design process for mudmats is iterative. Once structure inadequacy or excessive conservatism is identified at step III, one has to go back to step I and repeat the design cycle. To reach an optimal design with the right balance between structural efficiency and adequacy, multiple passes are often needed.
Without Mudmatician, excessive conservatism may have to be tolerated in order to minimize the engineering effort. With Mudmatician, the mudmat can be easily modified and even redesigned. This improves the engineering quality, as well as efficiency.
When structural inadequacy is identified in the mudmat, stiffening has to be implemented to enhance the strength. To reduce the engineering effort, a manual designer may have to confine the design change to those demanding least efforts. With Mudmatician, however, the designer can easily add and/or thicken the plates; it is also a trivial thing to redesign the spacing between the longitudinal and transverse stiffeners of the mudmat. Figure 3, for example, shows multiple stiffeners (in solid red lines) are added to the original design. With the built-in support to engineering from Mudmatician, the initial design or redesign of a mudmat is a much more efficient process.
Mudmatician also offers advanced features which assist the designers. One of these features is the built-in module for designing a mudmat with varying depth stiffeners, such as the one shown in Figure 1 that is deeper in the middle than the ends.
Equipped with another feature, Mudmatician supports graphic display of the checked structural integrity by writing back the results to Abaqus output database. This way, Mudmatician is able to utilize the full graphic capability of Abaqus. Illustrated in Figure 4 is the contour plot of buckling utilization factors (field B), which are shown at the checked elements on each plate. OE
About the Authors
Jinming Xu is an offshore engineer with Foster Wheeler Upstream in Houston, and has five years’ experience in the offshore oil and gas industry. He received his PhD degree in mechanical engineering from Texas A&M University and is a registered professional engineer in Texas.
Paul Schank, a senior principal structural engineer with Foster Wheeler Upstream, has over 17 years offshore oil and gas industry experience. Also a registered professional engineer in Texas, he received his BS and MS degrees in civil engineering from Texas A&M University.
Ray Young, VP of operations with Foster Wheeler Upstream, is a chartered engineer with over 35 years experience associated with the offshore oil and gas industry.