Desalination Megaprojects: Standardizing P&ID Development

Compared to conventional P&ID the information capacity of a multi-layered one discussed previously is at least 10X bigger and its scope goes far beyond "piping" and "instrumentation".

Such a P&ID consists of 2 distinct parts - an image showing interconnected P&ID symbols and a database linked to the symbols.

The P&ID image is generic - it is intended to describe a group of identical systems unlike its conventional counterpart, which is always mapped to a specific case. So generalization is not simplification,

Normally, the plant engineering produces between 10 and 20 P&ID images. To make this process faster and less error-prone Crenger.com came up with a set of guidelines covering the P&ID content selection and graphical preferences.

In conventional engineering the content selection follows a simple rule - fill up the A4 format drawing with symbols till it shows signs of cluttering.

Digital P&ID does not have a notion of the AX format. So creating a single P&ID image for the whole plant is quite possible. Such a P&ID features zero generality.

So, we may say that the P&ID image count is an implicit indicator of the P&ID generality and, hence, reusability. Interestingly, the plant customization does not change the existing images - it adds new ones.

Five rules govern a division of the plant on P&ID images.

1. Don't make P&ID smaller than a Control Module (CM)

CM is a group of P&ID items sufficient to produce some useful effect like to pump, to filter, etc. CM behavior is described by a set of standard states and events; it is a brick of PLC programming. As it traditionally concludes the project engineering, the rule in question may be re-phrased as "Start with the end in mind".

This rule is a serious obstacle as process engineers do not think in PLC terms.

2. Classify systems on design experience

The company's experience in engineering is not evenly distributed among the plant systems. Some of them are associated with the company's strength, others - with the weakness.

Crenger.com introduces the following levels of the design experience classification:

Except for the "copy" level, all others introduce some risk of project failure. Implicitly it may be linked to a number of updates performed on the system during engineering.

P&ID should not mix systems with different levels of design experience.

3. Separate procured systems from DIY ones

The company's decision to procure a system instead of assembling it (do-it-yourself) may be driven by insufficient design experience or the availability of a product on the market that fully meets the company's technical and commercial requirements.

In the latter case, the P&ID becomes part of the negotiation package, and it should be clean of any irrelevant information.

This requirement is critical for eProcurement where the seller gets online access to the package normally auto-generated on-the-fly. Errors in its scope may lead to sensitive information leaks.

Another reason is that the P&ID image is likely to change to absorb the manufacturer's superior experience. Here we default to a general rule - what is likely to change shall be torn off from what is stable.

4. Critical processes first

Two strategies are applied to the P&ID set development - once-through and iterative.

The former considers the plant as a queue of P&IDs, and the development of the next image starts when the previous one is fully formed. This "process-first" strategy dominates in the desalination.

The second "go-with-the-flow" strategy considers P&ID as a source of data to initiate the project schedule activities like procurement. It goes from long-lead items to mid- and short-lead ones. By analogy, the iterations shall sequentially address main continuous, auxiliary continuous, and auxiliary batch processes.

A less obvious advantage of the iterative strategy is that it decreases the amount of human errors as iterations happen to go from the highest risk impact to the lowest one.

5. Follow connections

A collection of P&IDs is not a plant if relationships among them are not known. P&IDs shall be sequenced and connected. That leads us to the following problem.

Can a customized plant be created using only the generic P&IDs mentioned above? The answer is negative. We need a special type of P&ID - connectors absorbing the plant customization.

The PI&D connectors are Source #1 of errors in megaprojects where limitations on the equipment size force designers to use identical systems connected in parallel. The reason is the triviality of the P&ID connectors - they mostly include only piping. So they are always at the bottom of the priority list.

The second reason is that the P&ID connectors are just the interpretations of the equipment layout, which is normally available long after the P&IDs are ready.

Below are the P&ID connector and corresponding layout excerpt from the SAMPLE project that comes with PlantDesigner software. As seen, P&ID accurately replicates the layout.

Injection of mechanical design content into P&ID

The upper image contains an unusual P&ID symbol - a miniature cross-section of a cartridge filter. Injection of such oversized non-standard symbols with multiple tie-in points into P&ID is a usual practice of setting requirements for spacious configuration and orientation of bulky equipment. In other words, P&ID images are not limited to the process, piping or instrumentation domains; they may contain generic mechanical designs as well. Below is a pictorial explanation of how Crenger.com bridges both worlds.