The case for engineering calculation management

It is not particularly difficult or expensive to begin managing calculations as a business process, and engineering companies can gain significant ROI, as this article explains.

It is not particularly difficult or expensive to begin managing calculations as a business process, and engineering companies can gain significant ROI, as this article explains.

Applied maths calculations are the backbone of engineering-computing critical product parameters, analysing test data, and predicting product performance.

When Ben Rich of the legendary Lockheed Skunk Works recalled his early days there as a thermodynamicist, he said: "We were the analytical experts, the elite of the plant, who decreed sizes and shapes and told the draftsmen what to draw".

Virtually every design decision is indeed informed by numerous calculations during each step of product development.

Yet these calculations are not managed, per se.

Product data management (PDM) and product life-cycle management (PLM) systems keep track of the results of calculations, but not of the full picture, which includes underlying assumptions.

Calculations in the engineering-based enterprise today are performed by hand, on calculators and spreadsheets, by writing programs, or using mathematical software; and they are scattered on desks, personal hard drives, document management systems, and file cabinets.

In failing to manage calculations (through technology standardisation, centralised storage, and global availability, for starters), enterprises risk needless redesign and disastrous errors, meanwhile squandering intellectual capital.

Calculation management is itself a mission critical business process, complementary to PDM and knowledge management.

Happily, it is neither particularly difficult nor expensive to begin managing calculations as a business process, so engineering-based enterprises have an opportunity to gain a significant return on their investment.

The collapse of the Tay Bridge in a howling gale on 28th December 1879, less than two years after it was completed, is one of the most famous bridge collapses.

Its designer, Thomas Bouch, was knighted when it was completed, but the glory was fleeting.

A court of inquiry held him responsible for the collapse by using an inadequate wind loading assumption of 10lb/ft2.

Enterprises support engineering calculations today on the level of task automation.

Individual engineers typically choose their own tools ranging from spiral notebooks and calculators to maths software based on personal preference, well under the radar of their IT organisations.

This is certainly a path of least resistance: engineers keep their artistic freedom, and IT doesn't have to worry about them.

So why change? Because calculation is a critical engineering process, and once companies begin to treat it that way, they stand to gain substantially greater benefits from their investments in calculation tools.

The ABET defines design engineering as "the process of devising a system, component, or process to meet desired needs.

It is a decision-making process (often iterative), in which the basic sciences, mathematics, and engineering sciences are applied to convert resources optimally to meet a stated objective".

Mathematics and analysis follow an iterative decision-making process.

Individual calculations are concoctions of raw data, practical know-how, mathematical laws, and scientific principles that answer the mundane but vital questions, such as "How thick will this wall need to be to handle the required load?".

The calculation process-advancing over the course of a project-is fundamentally what elevates engineering from trial-and-error and guesswork.

Imagine a flowchart of the engineering process for a large project.

It would be enormous, with loops and branches among subprojects.

The decision stream would have numerous tributaries, including the flow of calculations.

Indeed, it is safe to say that the history of the calculation process-inputs, assumptions, methods and results-is among the most important records of an engineering project.

Any unusual successes or failures are likely to be reflected conclusively within it.

It is common sense and well known in the business world that automating a business process offers a much higher return on the technology investment than automating discrete tasks.

In the late-1980s, the consulting firm Nolan, Norton and Co put some numbers on the phenomenon, citing modest (10 to 20%) ROI for task automation and impressive returns of up to 300% for business process automation.

Today's leading expert on IT investment as it relates to worker productivity is Erik Brynjolfsson, a professor at the Centre for Business at the Massachusetts Institute of Technology's Sloan School of Management.

He has spent years studying the business process changes, small and large, that are precipitated by new technologies, and which he believes account for the really impressive productivity gains.

Brynjolfsson's study of 1167 large companies in 41 industries concluded that productivity growth arises from the combination of new technologies and new business processes.

Addressing an e-business conference in Cambridge, Massachusetts in 2002, Brynjolfsson said companies' productivity and market value were not as closely linked with IT spending as with "how they used technology.

It was what they were doing with it.

And it was their corporate culture, their attitudes towards a whole set of information-related decisions".

Calculation management is not rocket science.

It is largely a matter of applying sound principles and technologies from the realms of information management and old-fashioned good housekeeping to the current calculation clutter.

Consider a group of engineers brainstorming the redesign of an axle in a conference room.

After the meeting breaks up, Fred and Mary work up calculations on two key components.

Mary creates a spreadsheet to forecast the type and number of bearings to be used, while Fred writes a program to calculate stresses on a modified U-joint for off-road use.

In the scenario above, we see a formula from Mary's spreadsheet and some of Fred's code.

These are essentially the only enterprise records of that brainstorming session.

The whiteboard was erased.

Fred's notes were on the legal pad he took to the meeting, and Mary's notes are in a spiral notebook on her desk.

As the axle design proceeds, the answers produced by the spreadsheet and the program form the basis for ensuing calculations.

Mary's bearings spreadsheet is eventually adapted for calculations of other assemblies.

A random audit later uncovers errors in 5% of the cells.

The company has been overspending on bearings as a result, but at least the over design poses no danger to customers.

Meanwhile, the new U-joint has failed repeatedly when driven over logs.

A recall may be necessary, and a government investigator wants to know what assumptions were used for flexing of the axle on uneven terrain.

Fred's program is still in the project document management system, but Fred himself has trouble reverse-engineering his code.

Scenarios like this are unnervingly common in engineering-based organisations, with varying degrees of impact.

Unmanaged calculations contribute to the following enterprise-level problems.

Computer programs are difficult to reverse-engineer, so the probability that they will be verified formally or informally declines almost to zero; that means errors are likely to show up downstream, or worst case, in the final product.

Poorly documented calculations are dangerous to reuse; this means the same problems get solved over and over again, and schedules slip.

If the lessons Fred and Mary learned were exposed in their work, they could be refined and used for the benefit of many projects.

In an emergency, an engineering team that cannot retrace its steps quickly risks adding public relations insult to existing injury.

The following are the basic principles of calculation management.

First, managing calculations includes managing underlying assumptions.

While many tools, such as spreadsheets and programs, can successfully automate calculations, the assumptions and data underlying those calculations are effectively obscured by most of them.

And without the information behind the answers in some readable and manageable form, how much is a company willing to risk reusing those answers in the future? Assumptions don't record themselves, but companies can provide a calculating environment (through technology and training) that is conducive to thorough documentation.

Secondly, liberate calculations from particular media and applications.

Maths is a particularly challenging data type.

Often when equations are represented most accessibly (as in printed materials and handwritten notes), they don't work; they're just pictures.

Meanwhile, working formulas in most applications are forbidding to the eye.

We all know the frustration of entering the same information into an array of databases or other sprawling applications that are intended to save labour.

Your calculation management tools should offer a user-friendly display, produce hard copy that is easy to work with, and interface neatly with relevant software.

Thirdly, establish rules and procedures for calculation management.

Successful calculation management strategies are a logical and natural outgrowth of current practices.

These systems should improve the enterprise's control over valuable intellectual capital and make it easier for engineers to do their jobs by: promoting reuse and creating libraries of in-house standards and methods; centralising key parameters; and making calculations and standards available online.

The Internet is the eye-level shelf for the information consumer, and remote access is essential for many engineering firms whose employees are frequently out in the field.

What are the benefits of managing the calculation process with Mathcad? Several Mathsoft customers have already profited from building processes around calculation.

One inherent benefit of Mathcad is that engineers like using it, whereas they tend to view most "knowledge management" technologies as overhead.

The following examples show how several leading engineering-based companies succeeded with independent calculation management efforts using Mathcad.

An aerospace contractor identified over 2000 critical design calculations needed to produce large aircraft fuel tanks.

These calculations are now housed on a central server, and used routinely.

The requirement for engineers to read, understand and customise these applications, and then disseminate complete analyses, made Mathcad the only choice for this application.

A major multinational engineering firm wanted to standardise their methods for calculating parameters of bolted girder joints and other structural elements for large building projects.

Because these calculations are performed daily by hundreds of entry-level engineers, requiring downstream validation and review, this firm chose to create a central repository in the form of Mathcad worksheets.

The firm then built a web service with custom Visual Basic forms that also allows engineers to use the worksheets remotely.

Each time a worksheet is used, the calculated results and validity checks on the parameters are fed back to the engineer's browser, and the worksheet itself remains as the "audit trail" for the calculation in the usual very readable Mathcad format.

A major OEM of auto parts has developed Mathcad-based templates containing large numbers of functions and data sets useful in their design engineering activities.

These templates ensure that all the engineering staff use the same functions and data sets, and the correct information in doing their calculations.

At this site, parts of Mathcad worksheets are often found in PowerPoint presentations, memos, emails, and other forms of communication.

A major aerospace company used to deliver engineering analyses to their US Government clients in the form of Excel spreadsheets.

Government engineers spent significant time checking each one, and invariably found at least one error in each.

This group has transitioned to using Mathcad exclusively for these deliverables, sharply reducing validation time by the Government and reducing the error rate almost to zero.

For the vast majority of calculations engineers perform, Mathsoft's flagship product Mathcad is the right application to use, both for personal productivity and enterprise-level management of a valuable asset.

Millions of engineers already use Mathcad in more than 90% of the Fortune 1000.

Calculation management with Mathsoft begins with standardising on Mathcad, with its undisputed strengths in calculation and user-friendliness at the individual level.

From there, Mathcad Enterprise provides central calculation storage and management, scalability, interfaces to systems such as Microsoft Office and SharePoint, and training in best practices for deploying Calculation Management at the group, departmental, and enterprise levels.

The following features can be based on SharePoint repositories supported directly by Mathcad Enterprise, or integrated with other document management, PDM, or PLM systems: libraries of reusable methods; centralised access to key parameters; web deployment of designs and applications; and electronic handbooks.

In conclusion, Mathsoft takes the broadest view in supporting engineers, beginning with calculation tasks and continuing through calculation management at the group, department and enterprise levels.

Calculation management is open territory for business process automation in engineering.

Mathsoft has created a medium for expressing, performing, and managing engineering calculations that opens up a range of high-value applications to an organisation.

Back to top