Method for detecting quality stability of rolled cigarette
Technical field
The present invention relates to a kind of method for detecting quality stability of rolled cigarette.
Background technology
In recent years, cigarette industry robotization, informationization, intelligent level are more and more higher, and the movable qualitative data that produces such as process control, quality inspection is more and more abundanter, for the Stability Control of process, quality index provides a large amount of basic data.But owing to the frequency of data acquisition, the difference of method, can cause the statistic of dissimilar qualitative datas and degree of freedom that very big difference is arranged, increase difficulty for the application of statistical technique.Therefore, how to utilize the qualitative data of these " magnanimity " come objective, accurately reflect the stability of process, quality index, and between different qualitative datas, make up a statistical study platform with comparability, with the reflection product batch in, batch between steadiness, being used for the mass discrepancy of comparison identical product at different enterprises, different workshop, different platform, is the basic content of cigarette quality stability study.
The unsettled generation reason of cigarette quality, mainly from: the one, in producing batch in quality fluctuation, the 2nd, the quality fluctuation between different production batch; The 3rd, different mass discrepancies of producing between board.In indexs such as current main employing franchise or undulating quantity are weighed batch, batch between quality, with the degree of stability of reflection cigarette quality, its core quantizating index is an extreme difference.But extreme difference is represented the degree of scatter of sample, and data user rate is lower, is generally used for the little situation of sample number.Therefore, quantize the fluctuation of quality index with extreme difference, the stability of reflection product quality has significant limitation, is necessary to explore new statistical technique, reflects the degree of stability of product quality.
Summary of the invention
Purpose of the present invention is exactly big in order to solve present range method detection cigarette quality stability limitation, the problem that can not correctly reflect cigarette quality stability, provide a kind of and have simple and easy to doly, can reflect the method for detecting quality stability of rolled cigarette of cigarette quality stability accurately and effectively.
For achieving the above object, the present invention has adopted following technical scheme:
A kind of method for detecting quality stability of rolled cigarette, its method be,
(1) according to Pneumatic silk feeding distance by as far as near order, selected successively some cigarette machines are the test board, and to its numbering;
(2) selected some batches is that test is criticized with trade mark pipe tobacco, and numbers respectively by the date of manufacture order, and a batch interval is not less than 7 days; Every batch of test every the some hrs sampling once, is called the test group of this batch; Take a sample some groups for every batch, by the sampling serial number;
(3) starting material and the cigarette physical index that each test group is used taken a sample, detected;
(4) use method of analysis of variance, the starting material of analysis of control board and cigarette prop up each index batch in, batch between cigarette stability;
(5) with the test group be unit, calculate the average of test result, adopt t to check paired analytic approach then, the difference value between the analysis different platform on same index consistency and the acceptable level;
(6) be vertical hurdle project with the index in the test, with the average of contrast board, standard deviation, batch in P value, batch between paired analysis between P value and board average, suppose that poor, one-sided P value is the horizontal bar project, make a quality stability observation table;
(7) table according to the observation, between judging batch, batch in and each the index stability between board, and determine to influence the principal element of rolled cigarette index.
In the described step (1), selected cigarette machine is no less than 3, and wherein a cigarette machine is the contrast board, and other cigarette machines are the contrast board; All boards must be from same tobacco rag feeding estrade for silk.
In the described step (2), test batch or group number are no less than 3 batches or group, with 5 batches or group for good.
In the described step (3), the sampling of the different tests group of same index, the necessary unanimity of detection rule, and duplicate detection is no less than 2 times.
In described step (4), (5), the also available degree of fluctuation of the degree of stability of index is represented, the corresponding relation of they and P value scope must clearly be stipulated, generally speaking, the cut off value of P value scope gets 0.01,0.05,0.20, and testing crew also can be according to producing the actual suitable cut off value of determining.
In the described step (4), adopting the true probability of abandoning in the method for analysis of variance is P value method, the reflection quality of stability, it with the test that contrasts board criticize, test group is district's group factor, do the dual factors non-interaction action and can repeat variance analysis, in determining batch, batch between degree of stability.
In the described step (5), the average of same index of test board, contrast board must be corresponding one by one by test batch, group; Carry out t check and analyze in pairs, suppose that difference at first is made as zero, whether observation bilateral P value is greater than 0.05 or setting value; Bilateral P value greater than zero index, is needed to adjust the hypothesis extent, one-sided P value is equaled or near 0.05 or setting value.
The present invention uses variance analysis, assumed statistical inspection etc. to be means, to abandon true probability is criterion, cigarette quality index, procedure parameter are carried out significance of difference analysis, quality fluctuation effectively in the inspection lot, batch, search the reason that produces quality fluctuation, again effectively between proof machine interstation, teams and groups and even the consistance of the same cigarette trade mark quality of different productions.
The invention has the beneficial effects as follows: method is easy, can effectively find the performance of quality fluctuation, and can intuitively find the reason of quality fluctuation rapidly, and the controlling level of improving the quality of products reduces defect ware and occurs.
Description of drawings
Fig. 1 is for generally speaking, P value and quality stability relation table;
Fig. 2 does not have alternately for dual factors can repeat analysis of variance table;
Fig. 3 is each batch test group index average corresponding tables one by one;
Fig. 4 is the t-check: in pairs two sample average analytical tables;
Fig. 5 is one-sided test methods analyst table in the t-check;
Fig. 6 observes table for the quality stability test figure.
Embodiment
The invention will be further described below in conjunction with accompanying drawing and embodiment.
The present invention is with certain trade mark pipe tobacco, obbbin, and filter tip, many cigarette machines, QTM cigarette combined test stand, 0.001g precision electronic balance, QTM nozzle rod sucting resistance test board, CJ-301 pipe tobacco vibrations separating sieve, the cigarette length tester is an instrument, its method is:
(1) according to Pneumatic silk feeding distance by as far as nearly order, selected successively 5,8, No. 10 cigarette machines are the test board; Selected 5 batches of J trade mark pipe tobaccos are numbered respectively by the date of manufacture order, and a batch interval is not less than 7 days; In the process of the test, serve as the contrast board with No. 10 machines, No. 8 board participates in the 1st, 2 batch of contrast test, and No. 5 board participates in the 3rd, 4,5 batch of test.No. 10 machine is fed silk with another test board from same tobacco rag feeding estrade.
(2) every batch of test every sampling in 1 hour once, is called the test group of this batch.Every batch totally 5 groups, by the sampling serial number.The sampling and testing rule: 1. each gets pipe tobacco 2 times continuously from every board Pneumatic silk feeding blanking port, and about at every turn 1kg detects its tobacco structure; 2. simultaneously, get 30 of mouth rods at random, detect its individual cigarette weight and resistance to suction from cigarette machine mouth bar bucket in the pipe tobacco sampling; 3. after the sampling of mouth rod, get cigarette at random from the cigarette machine and prop up 30 in sample.Wherein 10 are used for the cigarette length test, and all the other 20 are used for individual cigarette weight, resistance to suction, circumference, hardness and the test of pipe tobacco section ventilation rate; 4. in the every batch of pilot time slot, get 3 dish obbbins at random from the cigarette board, every dish is got 1 of the obbbin bar of 1m.Evenly get 10 points for every, do the air permeability test, write down every some detected value.
(3) processing method of data
(a) criticize the internal stability analytical approach: adopt the true probabilistic method of abandoning in the method for analysis of variance.Abandon the P value in the true probabilistic method, represent the probability of making mistakes when there were significant differences (being undulatory property) between each level of factor.Therefore the size of P value, reality reflected each level (as batch between, batch interior etc.) the size of stability.Therefore, adopt and to abandon true probabilistic method and reflect that quality of stability is very directly perceived, convenient and practical.The boundary of generally getting the P value is 0.20,0.05 and 0.01 three class, sees Fig. 1.According to the principle of variance analysis, with No. 10 board length detection data instances: with the board is unit, is district's group factor with test batch, test group, does dual factors and does not have alternately and can repeat variance analysis, the results are shown in Figure 2.As can be seen, it is significantly weak that the cigarette length index is criticized interior difference, better stable, and stable extreme difference between criticizing.
(b) analytical approach of mass discrepancy between board: adopt " t-check: in pairs two sample average analyses ".The same with variance analysis, the t check analysis is also adopted and is abandoned true probabilistic method, and it is not influenced by sample size and degree of freedom, and good comparability is convenient directly perceived.And t check not only can contrast the difference degree of two levels, and finds the difference on acceptable level.With No. 10 and No. 5 cigarette machine interstation cigarette length indexs is example, with the test result of each board, is that its mean value (see figure 3) is got by unit with the test group, adopts the paired analysis (supposing that difference is 0) of t check then, and the result as shown in Figure 4.Fig. 4 shows, between two boards cigarette length average hypothesis difference be 0 (indifference) deny true probability P (two tail) much smaller than 0.01, prove that the board differences is extremely significant.Further the value of supposing difference is adjusted, allowed P (single tail) near 0.05 (see figure 5).Prove that No. 5, No. 10 board cigarette length differences are that 0.103mm is reliable.And the equal value difference of No. 5, No. 10 board cigarette lengths is 0.152mm.As seen, with the difference of the mean value quantizating index as level difference, reliability is relatively poor.
(4) utilization method of analysis of variance and the t method of inspection carried out comprehensive analysis to all detection indexs: 1. with 5 batches of No. 10 boards, 4600 of totally 25 test group measurement data are sample, to No. 10 boards batch between the fluctuation, batch in the fluctuation situation analyze; 2. the mean value to every index of 15 test group of 25 test group of No. 10 boards and 10 test group of No. 8 boards, No. 5 boards carries out " t check: paired two sample average analyses ".Wherein, in batch, batch between quality fluctuation and board comparing result adopt without exception and abandon true probability, analysis result as shown in Figure 6, wherein tiltedly black font be fluctuation ratio significantly and extremely significant content.Hypothesis difference wherein is that No. 10 boards of expression of negative are lower than other boards; For positive number, represent No. 10 the board height.As can be seen from Table 6, the starting material of this trade mark cigarette and cigarette quality index batch in, batch between steadiness:
(a) cigarette machine supplied materials
Fluctuation in batch: it is not remarkable to fluctuate in tobacco structure is criticized, good stability; It is remarkable that the resistance to suction of mouth rod, weight indicator and obbbin air permeability are criticized the interpolation heteropole, stable extreme difference.
Fluctuation between batch: it is extremely remarkable to fluctuate between tobacco structure, the resistance to suction of mouth rod, weight indicator are criticized, stable extreme difference; Though obbbin air permeability P value less than 0.05, very near 0.05, can think that stability better.
The board contrast: supplied materials nozzle rod sucting resistance, weight and obbbin air permeability index are not remarkable at the board differences; Compare with No. 10 boards, a little less than No. 5 board tobacco structure significances of difference, No. 8 board tobacco structure long filament rates, whole cut rate index are extremely remarkable, and filament broken rate is not remarkable.
(b) cigarette physical quality
Fluctuation in batch: significantly, stability better a little less than length, weight, the resistance to suction difference; Circumferential differences is extremely remarkable, poor stability; It is good that hardness, ventilation rate are criticized internal stability.
Fluctuation between batch: the differences between batches of cigarette quality index are all extremely remarkable.
Board contrast: with No. 8 boards utmost point significant difference is arranged on circumference, hardness No. 10, length index difference is not remarkable, and other index differences are significantly weak; With No. 5 boards utmost point significant difference is arranged on length, weight, resistance to suction and ventilation rate No. 10, circumference index difference is weak significantly, and hardness number difference is not remarkable.
Utilize method of the present invention as can be seen,
(1) integral body is seen, it is better that the cigarette physical index is criticized internal stability, poor stability between batch.Supplied materials situation and cigarette quality control ability difference between No. 10 and No. 8 boards are little, and the supplied materials situation of No. 5 boards and cigarette quality control ability with No. 10 with to compare difference No. 8 bigger.
(2) from supplied materials, nozzle rod sucting resistance, weight and obbbin air permeability batch in, batch between fluctuation extremely significantly or significantly, in tobacco structure is criticized fluctuation significantly and batch between fluctuation extremely remarkable, in batch inner wrap strip physical quality stability batch better overall and batch between quality fluctuation very big.Show that the fluctuation of cigarette physical quality is because the fluctuation of tobacco structure causes between batch, supplied materials tobacco structure stable be batch in, batch between the prerequisite of cigarette physical quality good stability.
(3) from the contrast between No. 8 and No. 10 boards, both are only extremely remarkable at circumference and difference in hardness, and other difference are not remarkable; And No. 5 all not obvious with the circumference and the difference in hardness of No. 10 boards.Show that the correlativity that cigarette props up hardness and cigarette Zhi Yuanzhou is stronger.
(4) from the contrast between No. 5 and No. 10 boards, under the situation of two boards " supplied materials tobacco structure difference is extremely remarkable; obbbin, the basic indifference of mouth rod ", a plurality of physical quality index significant differences of cigarette, thus proved that once more tobacco structure is the key factor that influences the cigarette physical index really.
(5) from test findings, the emphasis of control cigarette quality stability work should be the mass discrepancy between quality fluctuation and board between control batch.The fluctuation root of cigarette quality mainly comes from the difference of supplied materials tobacco structure, and concerns not fairly obvious with obbbin, mouth rod.The difference of tobacco structure, mainly be batch between fluctuation and board differences.Therefore, reduce supplied materials tobacco structure between board difference and batch between the fluctuation of tobacco structure, be to guarantee cigarette physical quality stability important prerequisite.