Unique optimal continuous sampling inspection scheme for risk control improvement

Abstract

The invention discloses a unique optimal continuous sampling inspection scheme for risk control improvement. The method comprises the following steps: giving an average detection quality limit and second-class risks; collecting data; carrying out normality test on the latest data, if the fitting degree exceeds a threshold value, calculating the qualified rate and the lower confidence limit of the process, otherwise, stopping the machine for adjustment; comparing the qualified rate estimated value with the effective interval, comparing the average detection quality limit with the process capability critical value, and making a decision; if the number of the continuous qualified products is in the effective working interval, calculating the number ioj of the continuous qualified products and the test score foj, and determining the sample size n and the process yield index key value s0; and executing the optimal continuous sampling inspection scheme. Newest data is kept and recorded, a process capability index and a qualified rate estimator are calculated, and after comparison, a decision for executing an integrated control scheme or shutdown adjustment is made. The method is suitable for a scene without inspection capability limitation under large-scale continuous production, and can meet quality requirements at the lowest cost, eliminate risk control difference and control two types of risks at a set level.

Description

Technical Field

[0001] The present invention relates to the technical field of quality management and statistical process control, and in particular to a unique optimal continuous sampling inspection scheme for risk control improvement, which uses OCSP to accurately control process quality and Monitor process capability at all times. The solution determines the only optimal OCSP solution based on the current quality requirements and process capability, optimizes the frequency and scope of sampling inspection, and optimizes the production process risk control based on the process capability index. The present invention is suitable for large-scale continuous production environments. It optimizes the inspection process through risk control, which is convenient for practitioners to make flexible choices and ensures the stability and consistency of product quality. The present invention is suitable for scenarios without inspection capability restrictions, and can meet quality requirements at the lowest cost, while controlling the two types of risks at a given level. This technology is widely used in fields with extremely high quality requirements, such as aviation, aerospace, automotive, electronics, and pharmaceutical industries. Background Art

[0002] In the production process of manufacturing enterprises, the stability and consistency of product quality are key factors to ensure production efficiency and market competitiveness. In order to monitor and control process quality, Dodge first proposed a process quality control scheme of continuous sampling inspection (CSP), expounded the working principle of CSP, and used the mathematical method of series and sequence to give the calculation formulas of average outgoing quality (AOQ) and long-term average inspection (AFI). The CSP method is a process quality control scheme that integrates monitoring and improvement. It was quickly recognized and promoted. Li Chunzhi et al. proposed to determine the scheme parameters of CSP based on the process qualified product rate estimate of the controlled process. CSP-1, CSP-2, CSP-V, CSP-T and CSP-3 were improved in turn and named OCSP-1, OCSP-2, OCSP-V, OCSP-T and OCSP-3. These OCSPs have the following advantages: For a specified control process, there is only one optimal solution in each OCSP. Selecting the most appropriate scheme parameters from countless feasible CSP schemes solves the confusion of practitioners. When process capability fluctuates in the direction of improvement or deterioration, OCSP can adaptively strengthen quality control to keep the output quality of multiple concurrent processes controllable and consistent. The OCSP process introduces a lower confidence limit for the process yield estimate to limit process changes within a predetermined range. Subsequently, OCSP was reformed for the second time. In response to limited inspection capabilities, OCSP was combined with the process yield indicator S pkThe risk control schemes of the estimated quantities are combined to form an integrated control scheme that can simultaneously meet four types of constraints, namely quality, cost, first-class and second-class risks. Based on the above research, the present invention uses three types of performance indicators (average outgoing quality AOQ, long-term average inspection number AFI, acceptance probability L(p)) to compare the control effects of the OCSP series of schemes, and proposes an integrated control scheme of OCSP and process yield index without inspection capability restrictions. The integrated control strategy pointed out by the scheme is: select OCSP-3 for quality control under high process capability, select OCSP-1 under medium and low process capabilities, and in all cases, risk control is performed using a risk control scheme based on the process yield index. Summary of the invention

[0003] To achieve the above object, the present invention provides the following technical solutions:

[0004] A unique optimal continuous sampling inspection scheme for risk control improvement includes the following steps:

[0005] (1) Given the average outgoing quality limit AOQL, the first-category risk α, and the second-category risk β.

[0006] (2) Collect n (initial value n0=100) test data.

[0007] (3) Use the latest n production data to perform a normality test. If the normal fitting P-value is ≥ 0.05, calculate the qualified product rate estimate and the lower confidence limit q * , otherwise stop and adjust.

[0008] Calculate the process yield estimate using n data:

[0009]

[0010] in: is the maximum likelihood estimate of the process yield, q * is the lower confidence limit, USL is the upper tolerance limit, LSL is the lower tolerance limit, Using the sample standard deviation s and sample mean Estimate the process standard deviation σ and process mean μ respectively, where n is the number of samples, x i =(i,...,n) is the sample observation value.

[0011] Compute a lower confidence limit for the process yield estimate:

[0012]

[0013] in: is the value of the alpha quantile of the chi-square distribution with n-1 degrees of freedom.

[0014] (4) Compare with the effective working range (1-10AOQL, 1-AOQL] and make a judgment: Exemption from inspection; Stop production for repairs; If AOQL≤0.001, OCSP-3 is implemented; If AOQL>0.001, OCSP-1 is implemented.

[0015] (5) Calculate the number of consecutive qualified products i in the optimal continuous sampling inspection plan oj (j=1 or 3), the test score f in the score test phase oj (j=1 or 3), determine the sample size n and the key value s0 of the process yield index.

[0016] OCSP-3 scheme quality control parameters i o3 and f o3 The calculation formula is:

[0017]

[0018] OCSP-1 scheme quality control parameters i o1 and f o1 The calculation formula is:

[0019]

[0020] The risk control parameters n and s0 of the risk control scheme are the solutions of the following system of equations:

[0021]

[0022] Among them, S AOQL and S 10AOQL When p = AOQL and p = 10AOQL respectively, S pk Value Φ -1 (·) is the inverse function of the normal cumulative distribution function. The calculation formulas for a and b in formula (8) are: b AOQL =0, the calculation formulas for a and b in formula (9) are: b 10AOQL =0.

[0023] (6) Execute the integrated control scheme (n, s0, i oj ,f oj )(j=1 or 3). At the same time, keep the latest n inspection data and use n data to calculate the process capability index estimate and process yield estimator and will With s0, With q * For comparison, if And S pk ≥s0, then continue to step (3); if Then stop and adjust; if S pk ≥s0 and Then return to step (3).

[0024] The calculation formula is:

[0025]

[0026] Compared with the prior art, the present invention has the following advantages:

[0027] 1. Based on The risk control scheme of the estimated quantity can be equivalently integrated into each OCSP, so that the risk control effect of each OCSP scheme is the same, eliminating the risk control differences between OCSPs.

[0028] 2. Real-time monitoring of process status effectively prevents cost increases caused by process deterioration. Establish a process capability fluctuation control monitor so that the plan can not only reflect the capability level of the production process in real time, but also ensure timely stop-loss when the process capability deteriorates and causes cost increases, thereby improving the accuracy of process control.

[0029] 3. Achieve quality control with minimum inspection cost. By optimizing the design of sampling inspection, the integrated solution can reduce unnecessary inspection resource investment and ensure the minimum inspection cost while meeting quality constraints.

[0030] 4. Quantitatively control the first and second risks. This solution can simultaneously control two types of risks (accepting unqualified products and rejecting qualified products), ensure the reliability of product quality, and reduce costs and losses caused by quality problems.

[0031] 5. Improve inspection efficiency. This solution can significantly reduce unnecessary comprehensive inspections, optimize sampling strategies, reduce inspection costs and improve overall production efficiency.

[0032] 6. Dynamically adjust the sampling plan. Improve the accuracy of quality management. By introducing relevant indicators, the plan can adjust the inspection plan in real time according to the capacity level of the production process, thereby ensuring that the sampling inspection matches the actual production situation and further improving the accuracy and effectiveness of quality management.

[0033] 7. Applicable to complex production environments. This solution has the unique advantage of adapting to quality control under high variation and complex production conditions. It is especially suitable for production lines that need to frequently adjust inspection strategies to ensure that high quality standards can be maintained in a changing environment.

[0034] 8. Make full use of the counting and metering characteristics of the verification data. OCSP operates through the counting characteristics of the verification data. The risk control scheme is implemented through data measurement features. The operation of the integrated control scheme does not increase the inspection workload. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the appended Figure 1 , combined with examples, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of invention protection.

[0037] See also Figure 1 As shown, a technical solution provided by the present invention is:

[0038] A unique optimal continuous sampling inspection scheme for risk control improvement. When using it, follow the steps below.

[0039] (1) Given the current average detection quality upper limit AOQL = 0.00143, the first type of risk α = 0.01 and the second type of risk β = 0.01.

[0040] (2) Collect the inner diameter finishing data of 100 cylinder heads on the grinder. The specified inner diameter is

[0041] Data of fine grinding of 100 cylinder heads on the grinding machine Unit: mm

[0042]

[0043] (3) Perform a normality test on the data. The normal fitting P-value is 0.371 ≥ 0.05. The process has a high normal fitting degree, indicating that the process approximately obeys the normal distribution. Calculate the estimator of the qualified product rate and the lower confidence limit q * ,get:

[0044]

[0045] q* =0.995224

[0046] (4) Compared with the effective working range (1-10AOQL, 1-AOQL], according to Make a judgment: And AOQL=0.00143>0.001, so OCSP-1 is implemented. Calculate the scheme parameters of OCSP-1 and get:

[0047] i o1 =2527

[0048] f o1 =0.002725

[0049] n = 158

[0050] s0=0.9236,

[0051] (5) Execute the integrated control scheme (n, s0, i o1 ,f o1 ).

[0052] (6) Keep records of the latest n inspection data and use n data to calculate the process capability index estimate and process yield estimator and will With s0, With q * For comparison, if And S pk ≥s0, then continue to step (3); if Then stop and adjust; if S pk ≥s0 and Then return to the normality test step.

[0053] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A unique optimal continuous sampling inspection scheme for risk control improvement, comprising the following steps: (1) Given the average outgoing quality limit AOQL, the first-category risk α, and the second-category risk β. (2) Collect n (initial value n0=100) test data. (3) Use the latest n production data to perform a normality test. If the normal fitting P-value is ≥ 0.05, calculate the qualified product rate estimate and the lower confidence limit q * , otherwise stop and adjust. Calculate the process yield estimate using n data: in: is the maximum likelihood estimate of the process yield, q * is the lower confidence limit, USL is the upper tolerance limit, LSL is the lower tolerance limit, Using the sample standard deviation s and sample mean Estimate the process standard deviation σ and process mean μ respectively, where n is the number of samples, x i =(i,...,n) is the sample observation value. Compute a lower confidence limit for the process yield estimate: in: is the value of the alpha quantile of the chi-square distribution with n-1 degrees of freedom. (4) Compare with the effective working range (1-10AOQL, 1-AOQL] and make a judgment: Exemption from inspection; Stop production for repairs; If AOQL≤0.001, OCSP-3 is implemented; If AOQL>0.001, OCSP-1 is implemented. (5) Calculate the number of consecutive qualified products i in the optimal continuous sampling inspection plan oj (j=1 or 3), the test score f in the score test phase oj (j=1 or 3), determine the sample size n and the key value s0 of the process yield index. OCSP-3 scheme quality control parameters i o3 and f o3 The calculation formula is: OCSP-1 scheme quality control parameters i o1 and f o1 The calculation formula is: The risk control parameters n and s0 of the risk control scheme are the solutions of the following system of equations: Among them, S AOQL and S 10AOQL When p = AOQL and p = 10AOQL respectively, S pk Value Φ -1 (·) is the inverse function of the normal cumulative distribution function. The calculation formulas for a and b in formula (8) are: b AOQL =0, the calculation formulas for a and b in formula (9) are: b 10AOQL =0. (6) Execute the integrated control scheme (n, s0, i oj ,f oj ). At the same time, keep a record of the latest n inspection data and use the n data to calculate the process capability index estimate and process yield estimator and will With s0, With q * For comparison, if And S pk ≥s0, then continue to step (3); if Then stop and adjust; if S pk ≥s0 and Then return to step (3). The calculation formula is:

2. The unique optimal continuous sampling inspection scheme for risk control improvement according to claim 1 is suitable for quality control of production processes with upper and lower tolerance requirements and no inspection capacity restrictions, and realizes simultaneous control of quality, cost and two types of risks.

Images