This paper will discuss the problem of improving reliability prediction. It will consider the analytical advantages and disadvantages of the current situation in both the theoretical and practical aspects of reliability prediction for industry. This problem relates to many different areas of engineering.
The author will detail how much of the current predicting is inaccurate when it is based on information obtained by using traditional approaches of accelerated life testing (ALT) where the data related to the degradation (failure) process differs substantially from the product’s degradation process during its service life under real world conditions.
Through analysis of publications coupled with this author’s thinking the paper will consider the accuracy of currently used reliability prediction methods. It will analyze why currently considered reliability prediction approaches are not successful in industry and leads to many recalls, decreased reliability, durability, maintainability, and profit, and higher life cycle cost of the product.
The author will introduce a new approach to reliability prediction, which eliminates many of the aspects considered to be negative in current engineering situations.