This article was printed in CONTROL's December 2009 edition.
By Willem D. Hazenberg
The choice of a distributed control system (DCS) for a process company is one of the most important ones it can make. First, it is expensive. Average cost for a DCS is $1.5 million. Expenses for the selection process alone can run from €100,000 (approximately $150,000) for one project to €3 million (approximately $4.4 million) for a corporate-wide implementation. Investments in large-scale projects can run as high as €30 million ($45 million). The service costs over the lifetime of the DCS can run into multiples of the initial investment. Vendors typically will spend upwards of €35,000 ($50,000) to make a sale. Second, the selection process takes a long time and involves a lot of personnel. Selection alone can take nine month to three years, and the implementation can take another 12 to 24 months.
Then, the decision involves a long-term investment. The average life expectancy of a DCS is 17 years. Also, the choice of a DCS will have profound effects on the efficiency, productivity and profitability of an operation. It's a decision that companies can't afford to get wrong.
At the same time, local employees tasked with making the decision about the firm's next DCS are often at a disadvantage. Because this investment only takes place every 10 to 17 years, most selection team members don't have the knowledge needed to perform this analysis. Selecting a control system also is uncertain due to the subjective judgment of decision makers. The main problems are: 1) the exact criteria for selection are not known; 2) the method for deciding is unknown; 3) there are multiple actors, each with their own biases and preferences for particular suppliers; and 4) internal politics.
So, the choice of a DCS is not always univocal, and the relationship between the business case, chosen solution and selection process isn't always there. This disconnect will have growing consequences in the future. In 2006, the value of the DCS installed base more than 20-years-old was more than $65 billion, making them replacement candidates soon. As a result of my research, I've concluded that process users need to choose DCSs in a more univocal way, and so make their decision process more transparent.
Within the framework of my MBA thesis, I conducted a literature review and interviews alongside a survey. Research was done in 39 countries with 166 people about the selection criteria and processes for the purchase/migration/expansion or replacement of a DCS. The respondents work for end users, DCS suppliers, system integrators and engineering companies.
I investigated what kind of value proposition end users wanted from their DCS vendors, and asked DCS vendors the same (Table 1). For end users, it was not the superior product, but the best product for the best price that was the most important value proposition.
When I looked at end users' expectations about the technology, I discovered that "cutting edge" was not necessarily what most of them wanted. More then 70% of all users call themselves industry followers, and 51% said they will wait one year after the first product is released before considering it. Twenty-four percent said they would wait more than a year to install newly released products.
Multiple Phases of Selection
There are multiple phases in the DCS selection process (Figure 1). The first is the trigger point. A company decides it needs a new system for any number of reasons. The primary business case and trigger point to start the process of buying a DCS for replacement and migration projects is external—the DCS supplier will no longer maintain the current system. The vendor ceases support, and replacement parts become unavailable.
A second reason for many customers is the need to reduce equipment maintenance and related expenses, or the desire for process improvements and increased production. So, it's reasonable to say that migrations and replacements are more driven by fear of an obsolete system than by potentially improving capacity, gaining better control or other functional possibilities of a new system.
However, when considering extending an existing system or choosing a DCS for a greenfield project, the main drivers are improving automation, availability of improved algorithms, ability to get business information to the workplace through real-time data to enable faster decision-making, and automatic start-up and shutdown routines. In these cases, an internal desire to improve automation is the driver.
Once this trigger point is reached, a funnel effect begins to take place. The choices, determined by a variable group of people and a variable set of criteria, continuously narrow the options until one vendor remains. The first set of narrowing variables is corporate guidelines about preferred vendors, usually defined by the purchasing manager and an engineering consultant. The field of possible DCS vendors is narrowed to between one and five vendors—often only one or two. Government or related organizations mostly follow government procurement rules, which often require choosing the lowest-cost or, in the best case, the most economically advantageous tender.
Next, the selection process begins again, and the long list is narrowed even more, using more specific criteria, until finally the decision makers arrive at one choice.
Before a company starts the prioritization process, it should first define its business needs (function requirements) and its strategic outlook for how it wants to operate the plant in five to 10 years. Is it in the plan to outsource the system maintenance or not? (This has an influence on the priority for training or needed services). What about integration with other units? When these questions are answered, selection can begin.
Then we go to the long, short and final list phases, the thing to keep in mind about selection is that in a selection phase, there are actors with influence or power, and there are criteria with a certain weight factor or priorities. The results of this combination will be a shorter list of vendors in the next phase, where this process starts again, sometimes with other actors and a new set of criteria. The goal of our research was to define the core selection criteria and their priorities for the purchase of a DCS in the chemical industry, and to design a decision-making model, so the decision-making process for new systems was more balanced, more transparent, more consequent and faster.
Actors and Criteria
The main actors during DCS-selection phases are the control engineer, purchasing manager, project manager, consultant from headquarters, plant owner and maintenance manager. Other actors at the short list phase are engineering firms. The role of operators—the ultimate end users—is minor in the selection process.
An in-depth study inside and outside the academic world, inside and outside the DCS world, and company evaluation models provided 24 models. From these, I extracted 12 main criteria:
- Supplier vision
- Supplier's ability to execute
- Supplier's guarantee of the business case
- Service and support
- User experience and costs, including the initial costs, on-going expenses and exit costs.
The four most important criteria for end users are functionality (11.5%); technology (10.73%); service and support (10.65%); and business case guarantee (8.97%). On average, the highest priority is given to initial costs when buying a system, and the lowest to exit costs. Exit costs (switching costs) don't play a significant role in changing a DCS system. Lifecycle costs don't play a major role; the longer the period, the lower the given priority by the respondents.
At the long phase of the prioritization process, the business case guarantee (9.34%), interoperability (6.88%) ability to execute (5.05%) and exit cost (2.99%) are the most important criteria. Using a mathematical model, I calculated the influence from every actor for every phase, in which zero is no influence and five is a veto on every aspect. The control engineer (2.31), consultant from headquarters (1.68), project manager (1.49) purchasing manager (1.31) and plant owner (1.31) are the most influential actors.
At the short phase, the functionality (11.50%), technology (10.42%), implementation process (8.08%), user experience (6.66%) and vendor's vision about automation (5.60%) are the most important criteria. User experience is the feedback the actors will get from reference visits to other end user sites. The control engineer (2.34), consultant from headquarters (1.88), purchasing manager (1.82), plant owner (1.75) and engineering firm (1.58) are the most influential actors.
At the final phase, service and support (10.57%), initial cost (9.57%), training (4.86%); documentation (4.39%) and on-going cost (4.13%) are the most important criteria. The purchasing manager (2.41), plant owner (2.40), the control engineer (2.25), project manager (2.17) and plant manager 2.03) are the most influential actors.
It's not that certain criteria are important only in a certain phase. But, vendors who lack scoring on these criteria will not proceed to the next stage in the process (Figure 2).
Between the various industry sectors is a remarkably wide variance in the weight factors for the initial costs. For greenfield projects, over 90% of all respondents find that they are only interested in the pure system price or the initial installation price when they select a DCS. Life-cycle costing (LCC) considerations play no role. The initial costs get the highest weight priority from the group's headquarters consultant (often an internal engineering group) and an engineering company/engineering procurement contractor (EPC) (Figure 3).
Improving the Method
To improve the selection method, I propose using a multi-criteria analysis, analytical hierarchy process (AHP) method (Figure 4), and the presented selection criteria provide a reliable method for an objective system and partner selection. The application of an AHP model improves the decision making, and the AHP's systematic approach reduces the time needed to select a supplier.
Purchasing a DCS system is a balance between costs, returns and risks of migration and replacement projects. The goal is to choose a supplier that offers the lowest risk in the long run and best added-value services during the DCS's life cycle. By combining the lowest risk to the highest (satisfaction) score on the selection criteria, and by keeping the costs as low as possible, it's likely to result in a recommendation for that particular supplier. The AHP method provides an objective, systematic approach to achieving this goal.
Willem D. Hazenberg, MBA, MIM, EUR, ING RI, is a senior process control consultant, Stork Industry Services, the Netherlands.
For more information on this subject, including a PowerPoint summary of Mr. Hazenberg's thesis, go to www.controlglobal.com/0912_DCS.html. See also the following LinkedIn group: www.linkedin.com/groupRegistration?gid=142172b