2.3 PA30 -- Process Plant 30

PA30 is an abbreviation for "Prosessanlegg 30" in Norwegian. In English, this means "Process plant 30". The number "30" refers to the overall objective of the project: To reduce operating costs by 30%.

2.3.1 Background and objectives

The Kårstø gas treatment plant is operated by Statoil and located at the west coast of Norway. They receive natural gas and condensate from offshore production sites and produce sales gas and liquefied petroleum gas (LPG) to markets in continental Europe. Kårstø is a part of the Statpipe gas trunkline (Statoil, 1997d).

In summer 1994, they initiated a restructuring project with the overall objective of reducing operating costs by 30% within four years. None of the employees were to loose their jobs due to the restructuring, with the consequence that operating costs not related to personnel had to be cut by about 60%.

The restructuring project had three main phases:

The main purpose of developing the "as-is" enterprise model was to understand more of the current work processes of the process plant before attempting to change them. In addition, benefits like ownership and involvement from the plant workers were expected.

The discussion of the project is delimited to the mapping phase and to some degree the analysis phase, as this is where the enterprise modeling activities were carried out. Mapping and analysis ran from August 1994 to January 1995.

The investigation of the project is mainly based on access to project documentation (an experience report on process mapping and two final reports from the mapping phase and the analysis phase), two interviews with project participants (the first was with two actors -- denoted actors A -- from the project team, the second with a representative from the workers union, denoted actor B), and informal discussion with a third actor from the project team, actor C. The interview guide is included in appendix B. The presentation is also based on experiences reported in (Christensen et al. , 1995).

2.3.2 Enterprise models

The main deliverables from the mapping phase were a report, a statement database and enterprise models representing core processes at the Kårstø plant. The report summarized the findings and experiences from the mapping process. The statement database was populated by a collection of claims made by workers and suppliers that were interviewed about the gas treatment plant. The enterprise models were high-level diagrams representing the core work processes.
Figure 2.5 Modeling language used in the PA30 project

The main concepts and their symbolic expressions are provided in figure 2.5. Process, product and flow are similar to the concepts used in the VPT project (this is no coincidence, as one of the participants was central both to VPT and PA30). The additional elements are opportunity (something that works today, but can be improved or exploited), problem (something that does not work properly today), and role (the position held by an actor).

Physical flow was at the core of the models. Actors A claimed that modeling of physical artifacts was much easier than modeling of ad hoc, creative processes, and the project team used the physical process as a backbone of their models. Opportunities and problems were attached directly to the processes where they occurred.

According to actors A, the models were explicitly designed to include few details. As such, they had to be combined with the statement database in order to provide in-depth accounts of aspects of the processing plant. Both subjective claims made by individuals and more widely agreed upon facts were represented in the database.

Actors A also claimed that model simplicity was important to the project team: In order to make the models understandable to modeling illiterate workers, the symbols had to be simple and intuitive.

As for the VPT project, a simple computerized diagramming tool (ABC Flowcharter) was used to draw the models. They were printed on large A0 sheets of paper and hung on the walls on two locations at the plant.

2.3.3 The project organization

The mapping phase of the project involved the following groups of actors: The project team did most of the work in terms of interviewing and modeling. They were responsible for the mapping and analysis, and reported to the steering committee. Of the six members, only the two facilitators were experienced in process mapping. The remaining four actors had experience from working in different parts of the plant (both staff and line).

2.3.4 The enterprise modeling process

The planned modeling activities were the following (Christensen et al., 1995:1177): In practice, they followed a much more ad-hoc and adaptive approach. One of the facilitators recalls the modeling approach as follows (ibid.:1177):
"The project team members from the plant got quite a shock when they realized that the two Statoil facilitators did not come with solutions to all problems, nicely wrapped in a fancy PC program. Instead, they had to face -- and find solutions to -- all the problems on their own. After recovering from the shock, the project team proved that a simple method, pen, paper, a spreadsheet, a simple PC drawing tool, enthusiasm and common sense can be quite a powerful toolbox."
Hence, the modeling process was not designed in detail in advance, but rather adapted as the project went along, contingent on the situations at hand.

A fundamental decision was to interview all employees in addition to some customers and suppliers as well. Interviews were conducted in groups, and people who worked together were interviewed together. This ensured that all workers had the opportunity to present their views on problems and opportunities, in addition to ensuring a more correct picture of the work processes, according to actors A, B and C.

During interviews, statements were recorded and later inserted into the statement database. The enterprise models were built based on the statement database, they were not developed during interviews. There were several reasons: Not all of the project team members had confidence in the effectiveness of maps, and workers felt more comfortable when interviewed (instead of actively drawing boxes and arrows). In addition, drawing the models was very time consuming and required much of the resources. All members of the project team worked part time on the project, and resources were scarce (in terms of available time to carry out all intended activities).

However, the belief in the diagrams as a means to communicate increased during the project, and in the end, the project team would have liked to have some initial sketches of the process just to initiate and structure communication during interviews. Their occasional experiences with using the diagrams indicated that they were pedagogical and that actors got engaged in the discussion, according to the experience report.

The communication between the project team and the interviewed workers concerning the need for the restructuring project was perceived to be highly successful and appreciated by the workers, according to actors A and B. The interviewed workers tended to wander off a bit from the predefined themes, but this was experienced as fruitful by the project team. To many of the workers, this was the first time they had the chance to present their views of problems and opportunities related to their job, other than in informal discussions with fellow workers and the customary proposition box. The workers stated that they appreciated to have a forum and the opportunity to discuss matters related to their job.

One day per week were set aside for internal discussion in the project team. Initially, they intended to use the day to plan activities and discuss contents of the interviews. However, they turned out to spend much of the time reflecting, questioning and arguing over why, what and how to carry out the modeling process. This "outlet of frustration" was recognized by the project team members as necessary.

The enterprise models that were hung on the walls were intended to be commented upon directly by workers passing by. The comments were imagined to provide corrections and feedback. However, few comments were added to the models.

Initially, a bulletin with information about the project was created and distributed to the plant workers. The intention was to distribute the bulletin frequently, updating the plant workers on how the project progressed. This was not done, and after the mapping phase, the project group received feedback that more information during the process was considered preferable.

2.3.5 Problems and particularities of PA30

Actors A, B and C all expressed much enthusiasm and reported on few problems about the enterprise modeling part of the project. The expectations were high to the proceeding implementation phase. The plant workers were also claimed to be enthusiastic about the project.

One observation was the failure in using enterprise models directly in the interviewing sessions. Initially, the project team wanted the workers to draw "boxes and arrows" themselves, directly on the whiteboard. After a few unsuccessful attempts, they continued with traditional interviews instead. Still, ad-hoc symbols were used in the interviews but not the modeling language that had been developed.

Actors A stated that in retrospect, the project team wished they had run some pilot interviews in order to gain more certainty regarding choice of method and how to use the enterprise models. The work load on the project team became high, as more resources than expected was required to do the interviews and drawing models. Pilot interviews might have revealed some of the problems they faced early in the process.

An illustrating example of the effect of enterprise modeling was the discussion concerning the product of the processing plant. Some participants claimed that sales gas and LPG were their products. Others meant it was processing capacity. The "truth" is processing capacity: The Kårstø plant does not own the natural gas and condensate that they receive from the offshore reservoirs. They process it on behalf of a customer. Hence, the optimum for Kårstø is to produce the required quality at a lowest possible cost. They are not paid for producing higher quality than agreed upon in contracts. This insight is valuable to both the workers and the management: They sell processing capacity and their customer is the consortium that owns the gas.

According to actor B, both the workers union and management at Kårstø stressed that trust was important when mapping. If workers suspected that management would misuse the information, they would not be as open during interviews as they turned out to be. And commitment to the mapping process was of equal importance: Both workers and management had to show that they took the effort seriously and wanted to act upon the results of the mapping. Actor B stated that the restructuring process so far was one of the best he had participated in due to openness and trust.

All workers were told that they would keep their jobs independently of the restructuring project. Hence, they could be more frank regarding what the problems were in the production process, and stop worrying about making their own job look important, according to actor B.

The PA30 project was run by plant people. Hence, they were the ones who controlled it (not external consultants). It would be far more difficult for involved parties to be disappointed and reluctant to changes in work processes afterwards, as they all had several chances to express their points of view and influence the project.