Study On Acceptance Sampling Plan Based On Different Quality Characteristics Index

The good quality of the product brings a good reputation to an industry in the global market competition. In order to ensure the quality of the products, companies need sampling inspection for the raw materials, components and products. But in the actual manufacturing process, the different raw materials, parts or products often have different quality characteristics. How to establish sampling plans according to the different quality characteristics is a very practical problem and urgently needed to solve. Recently, Sherman originally proposed the repetitive group sampling (RGS) plan. The RGS plans need less sample size and save inspection time, so many scholars domestic and abroad are attracted.In this thesis, using the lifetime performance index, the coefficient of variation, and the reliable life as quality parameters, we respectively propose acceptance sampling plans combining with the repetitive group sampling for different life models.The main works are listed as follows:1. For larger-the-better products, the lifetime performance index can be used to assess the lifetime of the products. In this paper, we first utilize the lifetime performance index as a quality parameter to the sampling inspection of the product quality. Since there is a close relationship between the lifetime performance index and the non-conforming rate (or conforming rate), we can control the non-conforming rate of the products by means of controlling the lifetime performance index. Firstly, we have derived the estimator of the lifetime performance index based on the censored samples for exponential products, and then a single sampling plan based on the lifetime performance index is developed. Moreover, the sample size and the acceptance constant of the plan are obtained. Secondly, combining the idea of the repetitive group sampling, we have proposed repetitive group sampling plans based on the lifetime performance index for Rayleigh products and Weibull products. And the parameters of the plans are given under different quality levels and risk levels. So the consumers can make a decision of accepting or rejecting a lot according to the proposed sampling plans, while the producers can use these plans to control the nonconforming rate of products so as to improve the product quality.2. The coefficient of variation (CV) is an important index in the statistical quality management, which can be used to check the stability of products quality. When the mean and variance of the normal distribution are both unknown, the estimator of the CV follows the non-central t distribution. In this paper, we will use the estimator of CV as the test statistic and propose four acceptance sampling plans for normal products based on the CV, such as variable two-stage sampling plan, repetitive group sampling plan, multiple dependent state sampling plan and multiple state repetitive group sampling plan. In order to reduce the cost of inspection, the required sample size is better to be small. So we can choose the minimum of the average sample size as the objective function to solve the plan parameters,which can make the proposed plan to be economical and effective. Finally, two criteria named OC curves and sample size are used to compare the proposed plans with other sampling plans.The results show that the poposed four plans are all better than the single sampling plan proposed by Tong Yi. Moreover, the multiple state repetitive group sampling plan requires the smallest sample size. And the multiple dependent state sampling plan and the multiple state repetitive group sampling plan not only consider the sample of current lot but also consider the sample results of the preceeding lots, so they are very suitable to be used to inspect the continous lots.3. Normally, for a symmetric distribution, the average life is preferable to use as a quality parameter. But the majority of life distribution is not symmetric, so it is not adequate to describe the central of the distribution using the average life. In this paper, we use the reliable life of the products as the quality index to propose a sampling plan for Lomax products under the truncated life test. We firstly assume that the shape parameter is known and the lot is large enough, then the required minimum sample size is determined when the consumer’s risk is specified. Secondly, we have discussed the OC value of the proposed plan and the producer’s risk, and found that there is a relationship between the producer’s risk and the ratio of the true value and expected value of the reliable life. Finally, an example is given to illustrate that the proposed sampling plan is simple and easy to operate.4. When the in-control process mean fluctuates time to time, the process standard deviation is a linear function of the process mean,that is, the coefficient of variation is a constant. In this case, EWMA-CV control chart can be used to monitor the process variability. We all know that repetitive sampling can reduce sample size and save the cost and time of inspection. So we will combine the repetitive sampling with the EWMA-CV control chart. Then a new control chart with double control limit is presented to monitor the process coefficient of variation using an EWMA statisitc based on the estimate of the coefficient of variation. When there is a small shift, the average run length of the proposed control chart decreases. Moreover, the larger the shift the shorter the average run length. This demonstrates the proposed control chart performs well for small shifts as well as large shifts.