by DSI Staff | Published 03/22/2007 | Systems Engineering
Although eXpress analyses can be used to assess the diagnostic capability of a fixed system design, eXpress really comes into its own when the results are used to improve a design’s diagnostic capability. Unlike spreadsheet-based approaches, eXpress has been specifically designed to facilitate both early and iterative analysis; in fact, it can be effectively employed throughout the product development cycle, beginning in the earliest phases of system design. The value of this early design feedback cannot be over-estimated—if diagnostic analysis is postponed until later phases of the design process, it becomes much less cost-effective to implement the recommended changes. Additional costs can result not only from duplicated effort as the design is re-engineered, but also from situations where existing funds have already been allocated and therefore “new” funds must be found for the design changes. Unfortunately, when postponed analyses provide unattractive assessments of a system’s diagnostic capability, there is a temptation either to sweep the results under the metaphorical rug or to somehow develop positive “spin” on the “negative” numbers.
When diagnostic analysis is performed using eXpress, the design can be optimized early on, when modifications still fit within the expected cost profile for design development. eXpress offers the ability to not only perform, but also easily share assessments of a system’s projected diagnostic behavior. In addition to text reports, eXpress data can be exported using other formats (such as spreadsheets or XML files) that can be easily read by other processes. Because the eXpress design can be set up to resemble engineering schematics or management diagrams, the graphical representation of the design becomes an important aide when sharing the results of diagnostic analysis. Furthermore, within a study, the Diagnostic Flow Diagram can be examined interactively to demonstrate the reasoning performed by the diagnostics.
When the design changes (as is inevitable when diagnostic engineering is employed concurrent to the design process), the eXpress model can be updated and analyses re-generated with a minimum of effort. This is all very simple for eXpress, since the model itself both contains and maintains definitions of the interrelationships between the various components that comprise the system. This is a major advantage over the use of spreadsheets as the sole method of performing a diagnostic assessment, since the spreadsheets are typically not linked to a model and are therefore cumbersome to update as a design matures. Spreadsheets also lack a standard form—the equations themselves often needing to be updated each time that the data fields are updated—so there are additional costs associated with ensuring that the updated spreadsheet is still valid.
Because of all this, spreadsheet-based approaches to diagnostic assessment are typically postponed until the design is relatively mature (to cut down on the time spent re-populating and re-validating the spreadsheet each time the design changes). This is a situation, however, where money not spent is not money saved. By eliminating iterative analysis, the “wait-until-the-design-matures” approach typically results in additional costs. This can happen, when design modifications are introduced after the window of “cost-effectiveness” has already passed or, worse yet, when costs arise from not updating the design at all. Furthermore, because the spreadsheets are not integrated with graphic representations of either the design or the diagnostics, attempts to use spreadsheets as the foundation for communication between different specialists may result in less-than-optimal communication (which, of course, also increases upgrade costs).
Because the system’s complex interrelationships have been fully incorporated into its set of linked models, eXpress can reveal when an analyst’s projection of the overall diagnostic capability of a system may have been overly optimistic. This is because eXpress is able to easily determine the impact of each test across multiple levels of design hierarchy—something that is difficult for even the most seasoned Subject Matter Expert (SME).
For example, eXpress calculations accurately take into consideration situations in which one or more tests have overlapping coverage at lower-levels of the design hierarchy. Because the definition of each test’s coverage is part of the eXpress model, eXpress is able to accurately calculate BIT coverage or detection coverage, regardless of the complexity of the design. This differs greatly from spreadsheet-based calculations, where it is difficult to account for overlapping coverage in equations or macros. Moreover, for non-trivial designs, the SME who must populate the spreadsheet may even have difficulty determining the precise diagnostic usefulness of certain tests.
Another major drawback of spreadsheet-based approaches is that it is difficult to translate analysis results into recommendations. The spreadsheet equations may indicate that the fault isolation levels do not meet system requirements, but they do not point out why. In eXpress, all calculations are backed up with reports that can easily be used to determine the problem areas of the design.
There are thus many drawbacks to using spreadsheets as the primary method of diagnostic assessment. It is difficult to re-populate the spreadsheets as the design changes, to ensure the validity of equations as the spreadsheet is modified, to determine the best actions to remedy diagnostic inadequacies and to communicate the results with other SMEs. It comes as no surprise, then, that analysts who use spreadsheet-based approaches are tempted to resort to creative means of getting the appropriately “attractive” numbers.