Original Date: 11/01/2004
Revision Date: 01/18/2007

Electronic Part Obsolescence Forecasting plays a crucial role in managing system obsolescence and life-cycle costs. Current tools are prescriptive and have significant limits when it comes to predicting future part obsolescence. The Center for Advanced Life-Cycle Engineering has developed a methodology that is more accurate at predicting part obsolescence. This is a huge step toward enabling proactive and strategic life-cycle planning.

The advent of acquisition reform has caused a shift from traditional Mil-Spec parts to commercial-off-the-shelf (COTS) parts. The government, with longer life-cycle systems, is about 1% of COTS micro-electronic business. However, the majority of COTS micro-electronic business is driven by industry, which has shorter life-cycle systems. The shift has raised the need for solutions that help mitigate the inherent risk of sustaining and supporting COTS in legacy systems. The Center for Advanced Life-Cycle Engineering (CALCE) has developed algorithms to forecast electronic part obsolescence for use in proactive obsolescence management.

The electronics industry has a continuous pattern of growth and change. The tremendous growth of the industry causes electronic parts to have shorter life-cycles than the assemblies they are used within – one reason why a longer life-cycle system becomes obsolete. Part obsolescence occurs when there is a slow demand for the part or a shortage of materials to make the part. There are significant cost impacts associated with electronic part obsolescence in a system’s life-cycle, including procurement of new parts, storing of parts, upgrading the system, and mitigating risks in the system.

The objective of Part Obsolescence Forecasting is to track and archive obsolete parts and predict when existing parts will become obsolete. Current forecasting tools excel at articulating the current state of a part’s availability and identifying alternative options. However, limitations exist in the capability to forecast future obsolescence dates and provide quantitative confidence limits when predicting future obsolescence. Most forecasting is based on the development of models for the part’s life cycle. These models compute risk of obsolescence using methods based on ordinal scales where risk is a probability and the scales are not based on probability data.

In the basic CALCE method, sales data for an electronic part is curve fit. The attributes of the curve fits are plotted, and trend equations are created that can be used for predicting the life-cycle curve of future versions of the part type. In order to determine “windows of obsolescence” in the forecasted part life-cycle curves, CALCE has developed a methodology based on data-mining historical last-order dates for selected part families. Using the data supplied by part database suppliers, CALCE has demonstrated the development of data-mining-based algorithms for electronic part vendor-specific windows of obsolescence that can be used in conjunction with the life-cycle curve forecasting approach. Together they substantially increase the predictive capabilities of obsolescence forecasting approaches. This methodology enables more accurate obsolescence forecasts and can be generated for user-specified confidence levels.

For more information see the Point of Contact for this survey.