Design limit tests are those tests performed to ensure that a weapon system will provide adequate performance characteristics when exposed to environmental conditions expected at the extremes of the operating envelope. Many past development programs have attempted to satisfy these requirements by promoting MIL-STD or MIL-SPEC tests which did not duplicate actual operational environments, resulting in poor performance and excessive failures during field operations.
Mission profile data is essential to sound design limit testing of systems and subsystems. Early in the project, studies to determine the "real" mission profile and subsequent "worst case" environments must be performed. Care must be taken to allocate the overall worst case system mission profile environments to all subsystems. Some subsystems will see either increased stress due to amplification factors and cooling problems or decreased stress due to shock and vibration isolation or efficient cooling systems. Subsystem test environments must be adjusted accordingly to ensure that proper test levels are used. The contractor should be required to develop test environments using measured environmental data, where possible. Contractual arrangements should be made for the modification of test environments as additional data becomes available.
The overall test plan should integrate design limit with other planned testing. Design limit testing should be planned at multiple levels from parts through critical assemblies to subsystems to the system level. This will give a complete view of potential problems and allow data comparison from several sources. This planning will also increase test efficiency. Testing can be minimized using this approach, and a design proven at many levels will result. Integration of test results may often show problems which otherwise would be obscured by lack of data until field deployment. The field environment is a complex combination of environments. The best design limit test will result if these combined environments are simulated.
Many of these design limit tests are viewed as a hurdle which must be "passed" and as such, the test program becomes a success-oriented exercise. This in turn leads to unrealistic test schedules with little attention paid to the engineering information provided by the test results. Design engineers should be provided feedback concerning the results of their design efforts so that future designs may benefit from the knowledge gained or current designs improved. The objective is to establish knowledge of the design, not simply to pass the test.
All failure incidents are important and should be analyzed in detail, regardless of when they occurred. They may be symptoms of a future serious deficiency. Early analysis should prove otherwise or result in changes to prevent possible recurrence. High confidence, not test completion, is the goal of successful testing and the program should be structured to create this atmosphere. Verified effective corrective action is an intrinsic part of the qualification process. Redesign for all detected problems, and trend analysis to ensure effectiveness in later contractor and field testing, is necessary.
For subsystems which are being designed for multiple-use applications, tests must be designed to simulate worst case conditions. If a subsystem has been previously qualified, the environments previously used must be compared with those expected in the new application. Any case result in a more severe environment should require additional (delta) qualification testing to the more severe environment.