Quality control, also known as ensuring the effectiveness of results, or quality assurance. Quality control is divided into process quality control and result quality control; Results quality control was divided into internal quality control and external quality control. This paper only discusses internal quality control. The laboratory shall formulate and implement to monitor the effectiveness of verification, calibration and test results. The data obtained shall be recorded in such a way as to facilitate the discovery of its development trend. Such monitoring shall be planned and reviewed.
Internal quality control can include (but not limited to) the following means:
a) Use of reference materials or quality control substances;
b) Use other calibrated instruments that can provide traceable results;
c) Functional verification of measuring and testing equipment;
d) When applicable, use verification or work standards and prepare control charts;
e) Intermediate verification of measuring equipment;
f) Repeated verification, calibration or detection using different methods;
g) Repeat verification, calibration or testing of stored samples;
h) The correlation between the results of different characteristics of articles;
i) The results of the review report;
j) Intra laboratory comparison;
k) Blind sample test.
We divide these methods into three categories: method confirmation; Intermediate verification; Review class.
Method validation class
We all know that non-standard methods or self-defined methods need to be confirmed, that is, to confirm whether the formulated calibration and detection methods are correct by means of tests. If some means can confirm that the calibration and detection methods are correct, this means can be used as the most effective means of result quality control.
1) Use of reference materials or quality control substances
Reference materials and quality control materials belong to the measurement standard. If the difference between the detection value obtained by the detection equipment after detecting the reference materials and the nominal value of the reference materials is not greater than the maximum allowable error specified in the detection standard, it can be proved that this detection method is feasible, and here it can be proved that the internal quality control is qualified.
Attachment: the calibrated weights and gauge blocks are all quality control substances.
2) Use different methods for repeated verification, calibration or detection (hereinafter referred to as method comparison)
Whether the newly developed method in the laboratory is feasible or not, the test value Y1 obtained by this method can be compared with the test value Y2 obtained by the traditional method. If the uncertainty of the test results of the two methods is assumed to be U1 and U2 respectively, there is no correlation between U1 and U2. If | y1-y2 | ≤ sqrt (U1 ^ 2 + U2 ^ 2), the new method is proved to be feasible. It can also be used as an effective internal quality control means.
3) Correlation between the results of different characteristics of articles
There is a positive correlation between article characteristic 1 and characteristic 2, that is, when characteristic 1 increases, characteristic 2 should also increase. However, if the positive correlation is not met after detection, there is a problem with the method. This method can be used as an auxiliary internal quality control method without strict verification.
4) Blind sample test
The blind sample can be the tested equipment, but the stability must be guaranteed. Assign a value to the blind sample, which is equivalent to the reference material. The determination method is shown in 1).
In addition to the correct method, the accuracy of testing equipment is equally important. Instrument traceability is the most effective verification means to ensure the accuracy of testing equipment and the credibility of other characteristics. However, the effectiveness of calibration cannot be guaranteed to be effective in the whole calibration cycle. Therefore, the interim verification of testing equipment should be added in the laboratory, which is an economic and applicable guarantee scheme. The following means belong to the broad term of interim verification.
1) Interim verification of measuring equipment
It mainly includes high-level measurement standard calibration method and assignment comparison method. For interim verification, please refer to my previous article “measurement series 15: interim verification”.
The main verification is indication error, sometimes stability, and sometimes repeatability and other characteristics.
The judgment standard is the maximum allowable deviation specified in the test standard. If the test standard is not specified, the maximum allowable deviation is deduced according to the target uncertainty, or the customer’s requirements are taken as the maximum allowable deviation.
2) Use of reference materials or quality control substances
This method is equivalent to the high-level measurement standard calibration method.
3) Use other calibrated instruments that can provide traceable results (equipment comparison)
Under the same measurement method, two different devices are used to detect the same tested samples respectively, and Y1 and Y2 are obtained. The uncertainty of the two values is basically the same, which is set as U. however, almost all of the methods have systematic errors that cannot be reflected in the mathematical model (know but don’t know where they exist). Therefore, it is necessary to eliminate the systematic error component introduced by the method in U, because | y1-y2 | offsets the systematic error.
The judgment standard becomes: | y1-y2| ≤ sqrt (2) × U’（U'<U)
4) Functional verification of measuring and testing equipment
Before the equipment is put into use or after repair, the equipment shall be inspected. The verification here refers to Measurement Confirmation (tracing the source first, and then determining whether it meets the use technical indicators); At the same time, other functions of non metering performance characteristics shall be checked, such as screen display.
5) Where applicable, use verification or work standards and prepare control charts
The traditional interim verification only checks the changes in one calibration cycle rather than cross cycle assessment. However, the best scheme is to control the whole life cycle of the equipment, that is, use the detection equipment to regularly measure a stable verification standard, draw the measured values on the quality control chart, and analyze its change law. The standard deviation mean or variance mean can be used.
Recheck means cannot confirm the correctness of the method, nor can they identify whether there is a problem with the testing equipment. They can only check whether a single test is correct.
1) Intra laboratory comparison (personnel comparison)
Under the same measurement method, using the same equipment and two different personnel to test the same tested sample respectively, the uncertainties of Y1 and Y2 are basically the same, and set as U. in addition to the systematic error component introduced by the method, it is also necessary to eliminate the uncertainty component introduced by the bias of the testing equipment, because | y1-y2 | offsets these two systematic errors.
The judgment standard becomes: | y1-y2| ≤ sqrt (2) × U”（U”<U'<U)。
If more than one person participates in the personnel comparison, the arithmetic mean value can be used as the reference value and the normalized deviation en can be used for judgment according to the technical specification of JJF 1117 measurement comparison, but u “can also be used for calculation.
2) Repeat verification, calibration or testing of stored samples
The same sample can be tested, and the judgment standard is the same as 1) personnel comparison, but the uncertainty here is calculated by repeatability, while the personnel comparison is calculated by reproducibility. Therefore, u “‘< U” determined here is usually less than reproducibility.If sampling is involved, ensure the uniformity of the sample. The uncertainty u at this time shall include the uncertainty component introduced by sampling.
3) Results of the review report
At present, all test reports need to be reviewed (verified), so this means is used in every report. It is a kind of review, and the detection process cannot be reproduced. In fact, the correlation between the results of different characteristics can be used to review the report.
4) Correlation between the results of different characteristics of articles
It can be used as a logical judgment method for single detection. If processed by computer, it will be more effective. Recheck means mainly check whether there are errors in the detection, that is, abnormal values. Different quality control methods shall be selected according to the needs and conditions for different detection parameters. Considering the economy, for the same parameter, the method comparison can be used this year and the personnel comparison can be used next year to achieve the quality control effect.