Integrative Approach to the Plant Commissioning Process

Commissioning is essential in plant-modi�cation projects, yet tends to be ad hoc. e issue is not so much ignorance as lack of systematic approaches. is paper presents a structured model wherein commissioning is systematically integrated with risk management, project management, and production engineering. ree strategies for commissioning emerge, identi�ed as direct, advanced, and parallel. Direct commissioning is the traditional approach of stopping the plant to insert the new unit. Advanced commissioning is the commissioning of the new unit prior to installation. Parallel commissioning is the commissioning of the new unit in its operating position, while the old unit is still operational. Results are reported for two plant case studies, showing that advanced and parallel commissioning can signi�cantly reduce risk. e model presents a novel and more structured way of thinking about commissioning, allowing for a more critical examination of how to approach a particular project.


Introduction
1.1.Background.Plant modi�cations are an ongoing process throughout the life of any process plant.Reasons for mod-i�cation include efforts to improve reliability, production capacity, quality, or productivity.Seamless incorporation is the key concern associated with the installation of any new equipment in an operating plant due to the high cost of process downtime.Several steps can be taken to minimise the risk associated with the installation of new equipment such as hazard and operability studies, project management, development of redundancy plans, and commissioning of the new equipment.
Of these, commissioning is an essential activity in many plant-modi�cation projects and has signi�cant implications for project success.Yet paradoxically it tends to be approached in an ad hoc manner.It is oen included in project plans, so it is not that people are ignorant of commissioning.Rather, the problem is that there is a lack of systematic approaches to commissioning, so it is frequently le to tradespeople and plant operators to manage in whatever way they see �t. is is an undesirable situation since it results in unpredictable outcomes.In some cases it can even cause serious problems.An extreme example would be the catastrophic failure of the Chernobyl nuclear power plant (1986), which was caused by operators attempting an ad hoc test of the efficacy of a modi�ed emergency cooling system.is paper presents a structured conceptual model for the commissioning process, and two cases studies showing application to operating plant.

Existing Models of Commissioning
2.1.Literature.Many authors have highlighted the value of commissioning from a range of different perspectives but they all agree that commissioning and the integration of a new project is critical to the success of any project [1][2][3][4][5][6][7][8][9][10].However commissioning is poorly de�ned and is interpreted ambiguously [6,11], which leads to inefficient utilisation within industry.In this paper �commissioning� is de�ned as the disciplined activity involving careful testing, calibration, and proving of all systems, soware, and networks within the project boundary [5].F 1: ree common strategies for commissioning, broadly re�ecting the approaches described in the literature.If commissioning is perceived to be simply a routine set of tasks, which is a common assumption, then this tends to preclude any more thoughtful approach to the problem.

Current Models of Commissioning
(i) Type of project.us situational variables are important; that is, the factors that resulted in a successful (or failed) commissioning outcome in one case are not necessarily transferrable to a different situation.(ii) Who is in charge of the phase.Commissioning can be completed by a range of different groups depending on the project.It can be the equipment manufacturers, operation team, or a separate commissioning team depending on scale and requirements of the project [12].e relationships between these people are also important (social dimension) [6,13], hence also contractual obligations (see (iv) below).(iii) Number/type of phases.Commissioning can also be broken down in several sections such as planning, precommissioning, testing, integration, monitoring, documenting, and handover depending on the level of complexity of the project.is requires careful project planning (see (iv) below).(iv) Project planning and contractual sufficiency.It is widely recognised in the literature that commissioning requires deliberate planning, as opposed to ad hoc treatment.us it needs appropriate consideration in the work breakdown structure and project planning [14], allocation of resources, transferral of those costs into the initial contract [9,[15][16][17][18], and creation of speci�c operating procedures (especially important for safety-critical plant like boilers [19]).is corresponds to the "integration" tasks in the project management approach [6].
e commissioning process has been examined for a wide range of different projects [2-5, 8, 20, 21].e predominate approach can be described as task speci�c; the literature tends to identify speci�c tasks that should be completed as part of commissioning.us the focus has been on completing multiple checks on a system to ensure it will operate as expected.us there are many reports in the literature, too numerous to mention, about commissioning experiences in speci�c case studies.ese are undoubtedly helpful, especially for lessons learned and application to comparable situations.ey are also systematic, in a way, especially in the provision of templates and checklists to guide practitioners.
However there is a lack of holistic or integrative models.ere is much less literature at the next higher level of abstraction, which is the commissioning process in general.At this level we are interested not so much in case-speci�c experiences but in the fundamental principles and the methodology.What exists at this level is primarily in the area of instrumentation and control; some examples are [5,22,23].
us the existing commissioning strategies in the literature can be categorised into three types, see Figure 1.ese are (a) ad hoc, which is action-orientated problem solving; and (b) template, which involves using a checklist, or operating procedure that worked before or in another situation.Both (a) and (b) are premised on the assumption that commissioning is a routine set of tasks.e third strategy challenges that premise and calls for a deliberately thoughtful approach.us the third category in the literature is (c) methodological, which involves analysis of the situational needs and deliberate selection of the most relevant of several possible commissioning methods.

Issues and Problem Areas.
A clear refrain in the literature is that commissioning (i) needs deliberate project management, but (ii) is too oen not given the attention it deserves.One of the issues with commissioning, which contributes to problem (ii), is that the value thereof is hard to quantify.Justifying the value of commissioning may be completed using qualitative analysis similar to quanti�cation of risk in a project [24].is is based on the consequence and probability of the system failing to operate as anticipated.In other cases there is no attempt at justi�cation at all, so the value is not appreciated.
Another issue is the tendency to underresource the commissioning in the project planning, which is issue (i) above.Underresourcing is due to several factors such as its omission in the project management.ere is oen a high level of variability as a result of the case-speci�c nature making it difficult to �t into the established planning structure.Existing project management frameworks, such as the PMBOK [9], are general approaches.While they acknowledge the commissioning stage they do not, and cannot reasonably be expected to, provide case-speci�c guidance on commissioning.ey treat commissioning very lightly and rely on the practitioner to identify whether or not commissioning is an important part of the project.e literature suggests that practitioners too oen fail to realise the importance and therefore fail to plan sufficiently.Alternatively, project managers may simply be overly optimistic about the risks associated with the installation of a new system.Whatever the reason, the result can be insufficient resources being allocated, with the consequence of poor completion.Incorporation of a broad conceptual model of commissioning into the project management practices would be the �rst logical step.Commissioning draws from several project knowledge areas such as integration, communication, and risk management.e logical approach is to incorporate into the project life cycle between the execution and closing phases [4,6,8].

Pro��em De�nition.
Current models of commissioning tend to be simplistic, and relevant only to speci�c areas.ey are focused on the process and consequently tend to produce a somewhat prescriptive list of tasks that need to be performed.A higher-level reconceptualisation of the commissioning process, with the development of a more general theory, could be valuable.e purpose of this work is to develop a more holistic and integrative theory of commissioning.e speci�c emphasis is on reducing process downtime, without compromising plant reliability.is is worth attempting as it has the potential to provide a general framework in which the other more process-speci�c models can be placed.

Approach
We start by reconceptualising commissioning in broad terms.We categorise the commissioning strategies according to the operational risk.is results in three categories: direct, parallel, and advanced.We then apply a system modelling method to embed these within the broader manufacturing context.Finally we apply the new framework to two case studies to demonstrate the applicability.

Categorisation of Commissioning Projects
4.1.1.Starting Premise.We start with the premise that the value of commissioning is essentially one of systematic risk reduction, that is, used to minimise the risk associated with the installation of a new piece of equipment.More speci�cally the application of commissioning for the installation of new equipment into the process industry reduces the risk of equipment damage, environmental health and safety, and process downtime.
us commissioning is a strategy for treating risk [24].is has the further important implication that the treatment, hence type of commissioning, can be aligned with the degree of technical risk that the organisation can accept.us we speci�cally link commissioning, as a treatment strategy, to the concept of "tolerable risk" within the risk management literature, and to the concept of strategic risk for the organisation as a whole [25].is also has contractual implications in project-setup phase, where there is a need to differentiate between the commissioning risk elements and proportion them between the equipment manufacturer, project management organisation, and plant owner [26].
From this starting assumption we identify three categories of commissioning, as strategies in response to organisational risk-tolerance.ese are direct commissioning, advanced commissioning, and parallel commissioning.Each has strengths and weaknesses.ey can be deployed individually or together.

Direct Commissioning.
Direct commissioning is the classical approach to commissioning where the new equipment is installed and the system must remain offline as commissioning is completed.Direct commissioning is the most straightforward approach as no additional equipment or simulation is required.e new equipment is installed into its operational position and the process cannot restart until the system has been commissioned and is running correctly.ere is a high level of downtime in this process as the whole system cannot be operated until the new unit is electrically, mechanically, and operationally tested.ere is also the risk of having to reinstall the old unit if there are signi�cant complications at any phase of the commissioning process.Direct commissioning is oen reserved for well-established unit operations such as new pumps and �ow meters.Direct commissioning is most effective when it is used on wellestablished system and ones that are not a key requirement of the process.

Advanced Commissioning.
Advanced commissioning is the process of operating the new unit in advance of installation and in isolation of the main process operation.Advanced commissioning requires the simulation of all proprietary systems that interact with the new unit.Simulation can be extremely complicated or simple depending on the level of interaction between the process and the new unit.(In this context "simulation" can refer to the arti�cial provision of physical inputs to the new machine or unit, smaller scale models, and mathematical modelling of the functional behaviour of the unit.)Advanced commissioning allows for the electrical, mechanical, and part of the operational testing to be completed.e full functionality of the unit cannot be proven as the system is being simulated by external means, which will always be an approximation of reality.Advanced process is extremely valuable for the development of new technology as it allows for the veri�cation of novel processes at low risk.e most common type of advanced commissioning is the development of model systems which both simulate the operation of the system and the new unit.Advanced commissioning can also include computer simulation of new process which provides a cost effective method of developing concepts in the early stages of design.Advanced commissioning is valuable at proving conceptual designs of new technology.e main drawback of advanced commissioning is that the process is only simulated so there is still the potential that the unit can fail when installed into its operational environment.

Parallel Commissioning.
Parallel commissioning is the testing of the new system in parallel to the operating system.Parallel commissioning is the most rigorous form of physical and operational commissioning.It allows for the new unit to be tested under full operational conditions, with low risk of signi�cant process interruptions due to the added redundancy of the old system present in an operational capacity.However it also has the highest cost as it requires the duplicate hardware systems and additional structural space.e only risk associated with parallel commissioning is the integration between the two systems.Oen there is some type of switching or merging component in these systems which may require minor process stoppage for installation.Parallel commissioning is oen completed when it is critical that the process must not stop for any extended period of time.It oen lends itself to processes with few interactions between new unit operation and the rest of the process.Parallel commissioning is seldom utilised due to the requirement of a process that can accommodate both the new and old unit.

Conceptual Model.
Having identi�ed three types of commissioning, we next seek to set those within a conceptual framework.is is worth doing as it has the potential to identify the situational variables relevant to each type of commissioning.is in turn can be used to further build a theoretical foundation, and provide guidance to practitioners.

4.2.1.
Approach.e modelling method uses a structured, deductive process to decompose the process being analysed into multiple subactivities (functions) and for each deduce the initiating events, the controls that determine the extent of the outputs, the inputs required, the process mechanisms that are presumed to support the action, and the outputs.e model was then inductively reconciled with elements of the existing body of knowledge on this topic, and successively re�ned.e end result is a graphical model that describes the relationships between variables, thereby providing a synthesis of what is known and surmised about the topic.e model is expressed as a series of �owcharts using the integration de�nition �ero (IDEF0) notation [27,28].With IDEF0 the object types are inputs, controls, outputs, and mechanisms (ICOM) and are distinguished by placement relative to the box, with inputs always entering on the le, controls above, outputs on the right, and mechanisms below.

Develop Production Capability (Prd-1)
. e broader context is that commissioning occurs as part of the development of production capability, and our model starts at this level.(is is already the second level into the model; the top level, which is not shown here, includes product design, operation of the plant, control of production �ow, quality, distribution to market, packaging, health and safety, lean/JIT, among other activities.However the present paper focuses on the commissioning activities.)See Figure 2. Commissioning is included as element 5 and occurs towards the end of the plant-development process.Other important activities are the following.
(ii) Design of the production plant, which also includes the plant layout, material handling, plant control and automation, and (for manufacturing) the development of production tooling and �ow control, for example, just-in-time (JIT).Analysis of technology risk ( 9) is another activity associated with the design phase.
(iii) Building the production system (4), and the associated project management activities.
(v) Project management (8).We note the importance of project management methods in supporting many of the activities of commissioning.ere are several models of project management that might be inserted here, including [9,29], but these are not speci�c to commissioning and therefore not detailed further at this point.
We do not deal with these other activities here, but instead move the focus to the test and commission activities.Before doing so, we draw attention to some hollow arrows, which represent errors, in particular design and construction errors, at (2) and ( 4), respectively, and the possibility for unintended plant behaviours at (5): low productivity, safety issues, staff usability problems, product quality defects, and so forth.is point is important because the commissioning model that follows speci�cally seeks to address these risks.

Test and Commission Production System (Prd-1-5)
. e model for commissioning a new piece of plant equipment is shown in Figure 3 (Prd-1-5).e conventional commissioning process is included here, as are the new concepts for commissioning approach.One of the conventional activities is to verify instrumentation and control systems (1), which involves the systematic checking of installed hardware against plant schematics.e checks are progressively done for connectivity, cold operation, hot operation, and process control.We do not detail those processes here and instead refer the reader to source material [5] which has information that is useful to practitioners.e �nal objective of commissioning  is also well known, to deliver an operational production system (6) that is ready for the client to use.
Where our model differs is the inclusion of a deliberate stage of deciding which of three commissioning approaches to use in the situation (2): direct, advanced, or parallel.We also note in passing that the quality and lean imperative for continuous improvement will generally mean that there will be ongoing adjustments to increase productivity and quality (7) aer the machine has been commissioned.us commissioning the machine and closing the contract with the client may be the end of the involvement of the machine builders, but are not the end of the life cycle for the machine itself.is again has contractual implications in the form of service and warranty support from the vendors, and  reliability centred maintenance by the plant operators.ere is also the decommissioning to consider, which can be a project in itself.(In extreme cases, e.g., nuclear power plant, the cost of decommissioning is comparable to the initial construction cost.If there has been a catastrophic failure of the plant then the decommissioning cost can vastly exceed the construction cost.)

Select Commission Approach (Prd-1-5-2)
. e decision involves a choice of direct, advanced, or parallel commissioning.ese are not mutually exclusive.Instead some of them may be done sequentially, as shown in Figure 4.For example, advanced commission may precede either of the other two.
e various factors relevant to this commissioning decision are anticipated and clustered in groups: ability to recover from a failed installation, assessed or perceived technology risk, desired operational continuity, and timing considerations.e detailed model and the factors within each cluster are shown in Figure 5.At this stage the model is primarily logical and qualitative and is intended as a debiasing tool and a guide to action rather than a decision algorithm.It is also a framework for further research in that it proposes subjective relationships of causality that can subsequently be tested and developed as appropriate.(It may even be that in certain areas it could be possible to develop a mathematical model to support the decision, particularly in well-de�ned areas.Speci�cally, the model incorporates risk assessment and it is not impossible that there could be well-de�ned situations where the variables can be determined with sufficient precision that a quantitative risk assessment coupled with (say) a Boolean consideration of the other factors might make for a sufficient mathematical model.However further research would be required to take it to this level of detail.)us the model proposes that the following decision factors are relevant.
(i) Advanced commissioning is appropriate where technology risk is high, operational continuity is required, and timing constraints are tight.
(ii) Parallel commissioning is appropriate where operational continuity is required and timing constraints are tight.
(iii) Direct commissioning is appropriate where technology risk is low, operational continuity can be disrupted, and timing constraints are loose.
Finally, to complete this part of the conceptual framework, a model is provided for the testing activities of commissioning; see Figure 6.

Case Studies
Two cases studies were completed to determine the relevance of this commissioning model in the process industry.First was the development of a novel vertical screw system in the fertilizer industry which used the advanced approach to commissioning.Second was the installation of a ship unloader in the aluminium industry which used a parallel approach to commissioning.

Vertical Screw Project.
Ballance Agri-Nutrients single superphosphate plant at Awarua (New Zealand) had recently designed and installed a new phosphate rock feed system.A new vertical screw system was developed to replace the old gravity feed vertical chute which was prone to blockages in the highly reactive and humid environment present in the reaction chamber.e vertical screw was designed to increase the reliability of the process by forcing the rock into the reaction chamber, hence reducing the number of blockages.An advanced approach to commissioning was completed for this new project due to the risk associated with the installation of a complex and untested unit into a critical position in the production line.
e advanced approach to commissioning allowed for rigorous testing to be completed in a controlled environment with low risk to production capacity.Commissioning was completed in two stages with the development of a model system and full scale commissioning of the new system before installation.
Select commission approach (Prd-1-5-2)  Development of the model system allowed the basic concept of the system to be tested.e model was constructed out of crude materials and was tested under a range of conditions to determine the optimal operating parameters.e full scale vertical screw was designed and constructed based on the results obtained from the model system.
Commissioning was completed by operating the system under a set of conditions established to simulate normal operation.e simulation was completed by assembling the new system directly adjacent to its intended future position in the operational plant.It was wired into the system using all of the �nal wiring components but was not installed into the process.� feed hopper was �tted to the inlet of the screw to simulate the priority feed system and water and various other components were used to simulate the environment of the reaction chamber.e operation of the vertical screw during the advanced commissioning phase can be seen in Figure 7. e process of running the system under simulated operating conditions allowed the full commissioning of the mechanical and electrical systems.It also allowed for the partial commissioning of the operation capacity.During this process it was found that several components did not operate as expected.Changes were made to the system and were recommissioned without any negative effect to the production capacity of the plant.
e process of commissioning prior to the installation of the vertical screw into the system was extremely successful.Full mechanical and electrical commissioning was completed as the full electrical system was used to drive the advanced commissioning process.e operational testing identi�ed several design �aws in the vertical screw.It also reduced the uncertainty associated with the operation of this new technology.
Advanced commissioning has one signi�cant drawback; it only reduced risk associated with the installation of the new vertical screw.e simulated process cannot represent the real process exactly.ere are several factors such as continuous operation in the highly reactive environment that cannot be tested until the system is fully installed.Nonetheless the advanced approach was effective in eliminating latent defects and thereby reduced the overall risk.

Tiwai Point Ship
Unloader.New Zealand Aluminium Smelters (NZAS) in Bluff installed a new ship unloader on the Tiwai wharf for the unloading of alumina and coke from incoming ships.e new unloader produced by Alesa Engineering Ltd. replaced the forty-year-old unloader that was installed when the smelter was �rst constructed in 1�71.e new unloader has a signi�cantly increased capacity capable of discharging at 1,000 tonnes per hour (TPH) of alumina and 600 TPH of coke compared to the old unloader that was only capable of discharging 235 TPH of alumina and 250 TPH of coke [30].e new ship unloader was installed with the aim of reducing the time required for a ship to spend unloading.Less time spent unloading will mean a more efficient use of shipping resources and the reduction of costs associated with slow turn around (demurrage).
Installation and commissioning was completed by Alesa under the guidance of specialist project engineering company Bechtel who work onsite at the Tiwai point aluminium smelter.is process had to be completed under tight constraints as it was critical that there were no process interruptions.e smelter is a continuous process which cannot be shut down or restarted without high associated cost.
It was decided to take a parallel approach to commissioning; the old unloader must remain in a fully operational state until the new system has been thoroughly commissioned and proven capable of carrying the full operational load.
Keeping the old ship unloader in an operational state signi�cantly reduced the risk of supply interruptions, but introduced additional concerns relating to the integration of the two systems.Integration of both unloaders on the same wharf was completed by limiting the operation of the old unloader to the north half of the wharf, while the new unloader was installed and commissioned on the south half, as seen in Figure 8. Limiting the old unloader to half of the wharf increased the unloading time as the ship had to be manoeuvred around to allow access to all of the holds.is was taken as a minimal sacri�ce to ensure consistent supply.
e main modi�cation that was required for the integration of the new unloader into the existing infrastructure was the installation of a new conveyor system to replace the southern half of the existing wharf conveyor, clearly this has cost implications.is was completed before the new unloader was installed.Both the new and old conveyor systems operated as one continuous conveyor that serviced the new and old unloaders simultaneously.e upgrade of the conveyor acted to integrate the two unloaders into the overall process.e new conveyor and ship unloader were constructed and assembled off site.e units were then transported and lied into place.e use of pre-constructed assembled units signi�cantly reduced installation time, therefore allowing installation to be completed in the short window between scheduled shipping movements.
Several complications emerged during the commissioning, and these extended the project duration.However the parallel approach meant that these had no impact on the overall production capacity of the smelter.e reduction in unloading capacity caused by the limitation of the old unloader to the north half of the wharf was quickly offset by the high capacity of the new unloader even when it was operating at a reduced output.Parallel commission proved to be a successful method of commissioning as there were relatively minor additional costs and the risk of process downtime was completely mitigated.e old unloader was decommissioned once the new unit was fully operational [31].

Discussion
is paper makes several novel contributions.First, it provides a novel conceptual framework for the commissioning process.e model represents the decision making within the process, the broader context in which plant commissioning occurs, as well as making provision for the �ner details. e novelty is creating a structured representation of the commission process, where models are otherwise sparse.Commissioning is generally an ad hoc process, and the value of this new framework is that it provides a structured theoretical foundation for this important activity.
A second contribution is the categorisation of commissioning into three main types: advanced, parallel, and direct.is exposes the variability of strategies within the commissioning process, so it becomes apparent that there is not merely one universal approach to commissioning.Achieving this adds choice to the project planning.It makes it clear that there are choices that practitioners can make, and stating these choices encourages a thoughtful consideration of the planning and resource implications thereof.is categorisation thus adds richness to the conceptual model and makes the decision points more explicit, without being prescriptive.
A third contribution is the development of a model for use by practitioners.e model captures and represents the proposed situational variables (contingency factors) involved in the process.is is valuable for informing the decision making of practitioners.e applicability of the model has been demonstrated by case studies.
A fourth contribution is the integration of commissioning into other management models.e model provides integration with the "risk management" and "project management" disciplines.is is valuable because it shows practitioners how commissioning may be approached in a more holistic manner.e commissioning model is also integrated into a wider model for the development of production plant, and thereby into "manufacturing engineering" including quality and lean manufacturing.Space does not permit full description of this integration, but the point is that the work shows that this integration is indeed feasible.e model is represented in IDEF0 notation, which is a production engineering notation, meaning that it is readily comprehendible in that context.
Overall what has been achieved is to replace the otherwise ad hoc process of commissioning with a systematic process complete with proposed decision factors and internal models of causality.ere are implications for practitioners in the model, in the form of �owcharts identifying the critical success/failure factors for commissioning.us tentative recommendations can be made for the best commissioning approach for a given situation.
ere are also implications for further research.e model is at least partly conjectural, and further research could be directed at establishing the validity of the proposed causal relationships.Another strand of research could be directed at further re�nement of the model, and its extension deeper into speci�c cases, that is, further investigation of the situational variables.

Conclusion
Commissioning is extremely valuable to all projects but is poorly de�ned in the project management body of knowledge.e existing literature on commissioning is focussed on speci�c tasks, and holistic perspectives are lacking.is work has reconceptualised commissioning and shown that it is possible to identify three main types of commissioning (direct, parallel, and advanced) and construct a generalised conceptual framework around them.is approach to commissioning has been demonstrated by application to case studies.
e value of this work is that it presents a different and more structured way of thinking about commissioning.is allows for a more critical examination of how to complete the commission for a particular project, and ultimately the potential for a better commissioning outcome for practitioners.For theorists the bene�t is that a generalised model has been developed, thus a foundation for future advancement of the subject.We have shown that the commissioning activities can be integrated into the risk management, project management, and production engineering bodies of knowledge.
Factors that are known to affect the commissioning process include the following.

F 3 :
Model for the test and commissioning activities.

F 4 :
Relationship between the three commissioning approaches: advanced, parallel, and direct.

F 6 :
Model for the test activities.

F 7 :
Advanced commissioning of vertical feed screw during plant operation.Image shows plant being fed with material via a temporary arrangement while being commissioned.(Photograph by K Lawry and used by permission of Ballance Agri-Nutrients.)

F 8 :
Parallel operation of old (le) and new (right) ship unloaders on the Tiwai wharf.(Photograph by K Lawry.)