Computers have always been good at storing large amounts of data and sorting and searching through that data quickly. This is why database applications are so suitable for computerisation. To date, though, it has been difficult to make software 'clever', because commercial development tools have not made it easy to incorporate expertise.
The problem with building expertise into software is that it is necessary to have a deep understanding of both software development, and the technology being computerised. In our industry, this includes a knowledge of many subjects, including metallurgy, engineering, production, quality control and standards. Standards are particularly important as many aspects of fabrication are specified via national and international standards, such as ASME IX, AWS D1.1, EN 287/288 and ISO 9000.
Software houses with no depth of welding expertise or engineers with no depth of software development skills both find it difficult to develop expert welding systems. It may be possible for individual engineers to develop software, but long term support is at best difficult, and in most cases impossible. For storage of large amounts of information, where considerable time is invested in entering the data, long-term support is critical.
In addition, most existing software systems in the fabrication industry are tools for individuals, not large parts of organisations, because until recently, most organisations have simply not had the infrastructure to allow information to be distributed electronically. The widespread use of email has been one of the main driving forces behind the move of most fabricators to use local and wide area networks. It is now possible to share welding procedures or welder approvals across a company via a multi-user software system.
Even so, welding procedures and welder qualifications are usually managed by one or two key personnel. By using local and wide area networks, it is now possible to computerise, and integrate database systems. This paper describes the benefits of computerising welding procedures and welder qualifications, as well as the integration of these with systems of wider scope to manage all aspects of fabrication. Practical examples are used to illustrate how such software can be used in practice.
2. Welding procedures and welder qualifications
The management of welding procedures and welder qualifications is one of the most time-consuming jobs of a welding engineer. Creating, verifying and approving new procedures and checking, adapting and approving existing procedures takes a long time. Consequently, this was one of the first welding engineering tasks to be computerised. TWI's Weldspec and Computer Engineering's Welding Pro-Write are examples of such software.
Much of this job is administrative. That is, the management of paper welding procedures involved a lot of time with a typewriter and liquid paper. Creating or modifying procedures was very cumbersome. Searching for existing procedures for new production welds was a very time-consuming task and required expert skills.
The first welding procedure database management systems were simply electronic filing cabinets. They used the speed of data sorting that computers could offer to make searching for existing procedures much quicker. The ease with which documents could be copied and edited to create new documents meant that this task could be undertaken much faster. What they could not do very easily, however, was to give the welding engineer much help when creating new procedures for new applications.
The problem was that the sources of such information are wide and disparate. They comprise standards (welding and application), consumable and parent material handbooks, technical literature and, most difficult of all to computerise, experience. To build all this into a computer program would be, at best, very time-consuming, and impossible without a wide knowledge of the sources available.
More recent software, such as TWI's Weldspec 4, has been designed to give the welding engineer help in writing and drafting new welding procedures while still giving the benefits of speed and editing of existing procedures. The development of such software has required expertise in a large number of factors, including :
- World-wide welding and application standards, from such organisations as ASME, AWS, European Standards, etc.
- Industry practice in developing, qualifying and using welding procedures.
- Typical interactions between customer, fabricator and inspector.
- Welding engineering and metallurgy.
- Software development and knowledge representation techniques.
Management of welder performance qualifications is very similar to welding procedures, in that their administration is defined by standards. Thus, the variables that must be recorded, the extent of approval given by a test and the destructive and non-destructive test regimes are specified in National and International standards.
However, unlike welding procedures, welder qualifications are only valid for a specified time without practice or additional testing. Certificates expire, so the date of expiry becomes another key variable and the fast sorting capability of computers is even more beneficial.
3. Production welding and quality control
The management of welding procedures and performance qualifications, as described above, can give great benefits in saving time and reducing errors. However, if this is integrated with software to manage production welding and quality control, the benefits can be multiplied.
3.1 The problems
Most fabricators work in a compartmentalised way. The functions of design, engineering, production and quality control are discrete and communication is often difficult. This can give a number of problems :
- It is accepted that any error in the design stage of a fabrication can be very expensive to rectify once a weld has been completed. The concept of simultaneous engineering, where all personnel can contribute at all stages of manufacture, is particularly relevant to fabrication.
- It is difficult to monitor fabrication, as paperwork can take a long time to circulate. This means that problems can become serious before they are identified.
- The sheer weight of paperwork can make it very difficult to update welder performance qualifications based on satisfactory production welds. Many fabricators retest welders unnecessarily. Also, it can be very difficult to identify those welders whose qualifications are close to expiring.
- Similarly, it can be very difficult and extremely time consuming to collate project data books, simply because of the amount of paper involved.
3.2 The solution
An integrated software system such as TWI's Welding Co-ordinator can overcome all the problems listed above. Such software is designed to be used live to manage fabrication as it is progressing. It is usually based around an electronic 'weld map', or 'weld data sheet' or 'weld schedule', into which data are entered as welds are designed, engineered, welded and tested. The weld map would also usually have some space for approval, either weld by weld, or once a project or structure has been completed. Figure 1
shows an example of part of a weld map suitable for the power generation industry and Figure 2
shows an example of the system as used in the chemical process industry.
Data are usually entered into the system from four functions :
- At the design stage, where information such as the weld ID number and other design parameters (material type, thickness, joint type, etc) are entered.
- At welding engineering, where a welding procedure and possibly class of welder is assigned.
- At production, where welders are assigned and the completion of a weld is recorded (usually by entering the date).
- At quality control, where acceptance of the weld is recorded by entering test report numbers into the system, or via links to other software systems, such as TWI's NDTspec.
It should be stressed that these data should be entered live into the system, as fabrication progresses.
As well as solving the problems identified above, the system can give a number of other benefits, including the following :
- Instant progress reporting. Anyone with access to the system can see how fabrication is progressing. This may be simply by looking at the weld data sheet on screen, or by explicitly-programmed progress reports. These can identify bottlenecks (e.g. by comparing the number of welds completed with the number of welds radiographed), or help to produce reports for stage payments in large projects.
- Automatic assignment of welding procedures and welders. If enough information is supplied at the design stage, the system has everything it needs to automatically search through Weldspec's database for suitable WPSs. In addition, having chosen a suitable WPS, the system can search through Welderqual's database of WPQs to identify suitably qualified welders. If necessary, the system can list welders in order of expiry date of their certificates, so maximum benefit can be made of extending their qualifications.
- Automatic production of test requisitions. The system illustrated in Figure 2 automatically produces NDT requisitions, based on the NDT requirements for the project. These can be selected either manually, or the system could 'randomly' select welds for testing.
- Automatic generation of test percentages and repair rates. The system can produce reports on repair rates per welder (to identify training requirements) by procedure (to highlight defect-prone procedures) or any other measure, providing the relevant data are recorded.
- Automatic update of WPQs. This is a very time consuming task on paper and may be impossible due to the amount of paperwork. However, it is ideally suited for computerisation because it is a simple data sort and test function. The integration between Welding Co-ordinator and Welderqual makes it possible for the system to update all WPQs in a project, based on satisfactory production welds.
- Automatic generation of document packs on completion of a project. This can also be a very time-consuming task manually, but again is ideally suited for computerisation. On the click of a button, the system can print the weld datasheets for a project, along with all the WPSs (with backup PQRs if necessary), WPQs and NDT reports. Alternatively, this information can be archived on CD.
- Instant traceability between production welds information and the back-up information. Thus, if the inspector wants to see a WPS that was used on a weld, or wants to see proof that the welder was suitably qualified, this can be done on the click of a button. This facility can also be useful after a number of years of service. If a defect is found in a structure, it is possible to access the original WPS (for repair purposes), or the NDT report (to see if evidence of the defect was present at testing).
3.3 Problems and pitfalls
If all the above seems too good to be true, there are pitfalls, as described below :
- For any customised software, it is critical to get the requirements of all personnel who will use the system identified at an early stage, and approved by all concerned. In practice, the software developer and the fabricator will collaborate in producing a list of requirements (WHAT the system will do) and both parties will sign this document.
The production of a requirements document can be a time-consuming process, and the timescale and cost of the system can only be estimated after this has been generated, so it is an important stage in development.
- A system such as Welding Co-ordinator affects many people, as it is, by its nature, a multi-user program. Any form of change is often resisted, so it is important to identify all the potential users of the system and involve them in the requirements analysis stage.
- It is important to find a champion of the system within the fabricator's organisation. This person can help to promote the system in the absence of the developer.
- It is very tempting to develop the system to be too prescriptive. For example, it is possible to develop the Welding Co-ordinator system to only allow the selection of welders who are registered in Welderqual as being qualified. In practice, however, the qualification test may have been taken, but the NDT results may not have been received. Prohibiting selection of that welder would not be correct in such circumstances.
Arguments often arise between QA/QC personnel and engineering staff on how prescriptive to make the system. Experience shows that it is preferable to err on the non-prescriptive side, at least in the early stages of system implementation.