Online identification method for cigarette weight stability

Abstract

The present invention discloses an online identification method for cigarette weight stability. The main design concept of the present invention is to set a sliding window for the cigarette weights collected by the online automatic sampling and detection link, determine the actual normal distribution fitting curve of the cigarette weight data obtained within the sliding window, and obtain its simulated normal distribution curve. Then, a correlation analysis is performed on the actual normal distribution fitting curve and the simulated normal distribution curve. The control stability of the cigarette weight index is determined based on the correlation between the two. This method can reflect the stability of cigarette weight during the production process, facilitate the timely detection of anomalies, and alert operators to follow up and prevent unqualified cigarettes from entering the next process.

Description

Technical Field

[0001] The present invention relates to the field of cigarette manufacturing, and in particular to an online identification method for cigarette weight stability. Background Art

[0002] Cigarette-making equipment experiences a certain degree of vibration during the production process, and the fabric belts also experience a certain degree of wear and stretching, which can cause fluctuations in cigarette weight. Under normal circumstances, the weight distribution of cigarettes follows a normal distribution. Cigarette weight also has a certain correlation with draw resistance. If the weight is too heavy, the draw resistance will increase, which will greatly affect the consumer experience. When the equipment is unstable, the weight of cigarettes fluctuates greatly. If the operator fails to detect this in time, cigarettes that do not meet the weight standard will flow into the next process, posing a significant quality risk. Therefore, it is necessary to pay close attention to the distribution of cigarette weight and determine the stability of cigarette weight during the production process. Therefore, it is necessary to promptly determine the stability of cigarette weight during the production process to prevent unqualified cigarettes from flowing into the next process, while also reducing the consumption of raw materials and lowering production costs.

[0003] In this regard, analysis has determined that the weight of cigarettes is closely related to the stem content and moisture content of tobacco. Therefore, some current processing ideas are to establish a neural network model, collect the process parameters of different batches of tobacco, and import them into the neural network model training to determine the parameters in the production process of cigarette equipment to achieve the purpose of controlling the weight of cigarettes; on the other hand, or use the cigarette equipment itself with a cigarette weight control system to detect the weight of cigarettes, and adjust the position of the leveling plate or suction belt according to the test results to make the weight of cigarettes meet the process standards, and at the same time remove cigarettes that are too light or too heavy to ensure that the weight of cigarettes meets the requirements.

[0004] However, after analyzing the prior art, the present inventors believe that the determination of cigarette weight stability still has the following defects:

[0005] (1) Tobacco production is a batch production, and the weight of tobacco required to produce a cigarette is quite different. Therefore, the process parameters of tobacco cannot represent the parameters of tobacco for a cigarette, and cannot reflect the stability of the equipment in controlling the weight of cigarettes.

[0006] (2) The weight control system of the cigarette making equipment itself has a certain lag from detection to control action to the qualified weight of the cigarette, which cannot reflect the weight of the cigarette in real time and cannot intuitively reflect the stability of the cigarette weight.

[0007] (3) The weight control system of the cigarette making equipment itself may also suffer from misdetection, misrejection, or even damage.

[0008] Based on this, the present invention believes that the existing method for determining the weight stability of cigarettes needs to be adaptively optimized and improved. Summary of the Invention

[0009] In view of the above, the present invention aims to provide an online identification method for cigarette weight stability to solve the above-mentioned technical problems.

[0010] The technical solution adopted in the present invention is as follows:

[0011] The present invention provides an online identification method for cigarette weight stability, which includes:

[0012] The preset sliding window is used to sample the weight data of cigarettes detected online;

[0013] Obtaining a normal distribution fitting curve of actual cigarette weight based on cigarette weight data within the sliding window;

[0014] According to the cigarette weight data in the sliding window, a simulated normal distribution curve of cigarette weight is obtained;

[0015] Performing correlation analysis on the actual normal distribution fitting curve and the simulated normal distribution curve;

[0016] Based on the results of the correlation analysis, the cigarette weight stability is determined.

[0017] In at least one possible implementation, obtaining a simulated normal distribution curve of cigarette weight includes:

[0018] Obtaining a weight mean and a standard deviation based on the cigarette weight data within the sliding window;

[0019] The simulated normal distribution curve is obtained by using the weight mean and the standard deviation in combination with a normal distribution algorithm.

[0020] In at least one possible implementation, the utilizing the weight mean and the standard deviation in combination with a normal distribution algorithm includes:

[0021] Based on the weight mean and the standard deviation, the preset cigarette weight process standard range is divided into several equal parts to obtain corresponding several simulation value arrays.

[0022] In at least one possible implementation, the correlation analysis includes:

[0023] Normalizing the actual normal distribution fitting curve and the simulated normal distribution curve;

[0024] Perform correlation analysis on the normalized actual normal distribution fitting curve array and the simulated normal distribution curve array to obtain the correlation coefficient.

[0025] In at least one possible implementation, determining cigarette weight stability based on the result of the correlation analysis includes:

[0026] By comparing the numerical correlation analysis results with different preset thresholds, it is determined whether the weight of the cigarettes is stable or whether preset warning levels of different levels are triggered.

[0027] In at least one possible implementation, comparing the numerical correlation analysis results with different preset thresholds to determine whether the cigarette weight is stable or triggering preset warning levels includes:

[0028] If the result of the correlation analysis is numerically greater than or equal to the first threshold, it is determined to be a stable state;

[0029] If the result of the correlation analysis is numerically less than the first threshold and greater than or equal to the second threshold, a lower level warning is triggered;

[0030] If the result of the correlation analysis is numerically smaller than a second threshold, a higher level warning is triggered, wherein the first threshold is greater than the second threshold.

[0031] In at least one possible implementation, the length of the sliding window is constant, and the sliding window moves at a predetermined step length to obtain new cigarette weight data.

[0032] The key design concept of the present invention lies in setting a sliding window for the cigarette weights collected during online automatic sampling and testing. The system then determines the actual normal distribution fitting curve for the cigarette weight data captured within the sliding window and obtains its simulated normal distribution curve. Correlation analysis is then performed between the actual normal distribution fitting curve and the simulated normal distribution curve. The stability of cigarette weight control is determined based on the correlation between the two, and the degree of correlation. This system can reflect the stability of cigarette weights during the production process, facilitating timely detection of anomalies and alerting operators to follow up, preventing unqualified cigarettes from entering the next process. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described below with reference to the accompanying drawings, in which:

[0034] Figure 1 A flowchart of the method for online identification of cigarette weight stability provided by an embodiment of the present invention;

[0035] Figure 2 Examples of actual normal distribution fitting curves and simulated normal distribution curves provided by embodiments of the present invention Figure 1 ;

[0036] Figure 3Examples of actual normal distribution fitting curves and simulated normal distribution curves provided by embodiments of the present invention Figure 2 . DETAILED DESCRIPTION

[0037] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0038] The present invention proposes an embodiment of a method for online identification of cigarette weight stability. Specifically, Figure 1 shown, including:

[0039] Step S1: Sampling the weight data of cigarettes detected online using a preset sliding window;

[0040] During implementation, the existing online automatic sampling and detection device can be used to automatically sample cigarettes at preset time intervals and perform physical index detection, including the weight of the cigarettes. The sliding window is set to N. Every time a new value is detected, the sliding window moves one step, and N remains unchanged.

[0041] Step S2: Obtaining a normal distribution fitting curve of the actual cigarette weight based on the cigarette weight data in the sliding window;

[0042] Specifically, the normal distribution of cigarette weights is calculated within the sliding window to obtain an array Y of actual normal distribution fitting curves of the cigarettes.

[0043] Step S3: deriving a simulated normal distribution curve of cigarette weight based on the cigarette weight data within the sliding window;

[0044] In actual operation, the weight mean and standard deviation can be obtained based on the cigarette weight data in the sliding window, and the simulated normal distribution curve can be obtained using the weight mean and standard deviation. The implementation details of this concept are as follows:

[0045] (1) Calculate the average value μ of the cigarette weight within the sliding window.

[0046]

[0047] In formula (1), x(n) is the sample value of cigarette weight in the sliding window, n=1,...,N, and n is the sampling sequence number.

[0048] (2) Calculate the standard deviation σ of the cigarette weight within the sliding window.

[0049]

[0050] In formula (2), x(n) is the sampled value of cigarette weight in the sliding window, and μ is the average value of the sampled value of cigarette weight in the sliding window.

[0051] (3) Using the mean value and standard deviation obtained by equations (1) and (2), the cigarette weight process standard range is divided into K equal parts, and the simulated value array S is obtained using the known normal distribution curve formula to obtain the simulated normal distribution curve array F.

[0052]

[0053] In formula (3), s(k) is the simulated value after the cigarette weight is divided into K equal parts within the process standard range, k = 1, ..., K, k is the sequence number of the simulated value; μ is the average value of the cigarette weight in the sliding window, and σ is the standard deviation of the sampled values ​​of the cigarette weight in the sliding window.

[0054] Step S4, performing correlation analysis on the actual normal distribution fitting curve and the simulated normal distribution curve;

[0055] Here are the following preferred solutions:

[0056] (1) Normalize the actual normal distribution fitting curve array Y of cigarettes and the simulated normal distribution curve array F.

[0057]

[0058]

[0059] In formula (4), y(k) is the value in the actual normal distribution fitting curve array Y, y min is the minimum value of the actual normal distribution fitting curve array, y max is the maximum value of the actual normal distribution fitting curve array; in formula (5), f(k) is the value in the simulated normal distribution fitting curve array F, f min is the minimum value of the actual normal distribution fitting curve array, f max The maximum value of the array of actual normal distribution fitting curves.

[0060] (2) Perform correlation analysis on the normalized actual normal distribution fitting curve array YN and the simulated normal distribution curve array FN to obtain the correlation coefficient r.

[0061]

[0062] In formula (6), yn(k) is the value in the normalized actual normal distribution fitting curve array YN, and fn(k) is the value in the normalized simulated normal distribution fitting curve array FN.

[0063] Step S5: Determine the cigarette weight stability based on the results of the correlation analysis.

[0064] Different thresholds can be set, and the correlation analysis results can be compared with the different thresholds to determine whether the cigarette weight is stable or to trigger different levels of preset warnings. Specifically, if the correlation analysis result is numerically greater than or equal to a first threshold, a stable state is determined; if the correlation analysis result is numerically less than the first threshold and greater than or equal to a second threshold, a lower-level warning is triggered; if the correlation analysis result is numerically less than the second threshold, a higher-level warning is triggered, where the first threshold is greater than the second threshold.

[0065] Based on the previous example, a correlation coefficient between 0.8 and 1.0 is considered extremely strong, and between 0.6 and 0.8 is considered strong. Therefore, 0.8 and 0.6 are used as the cutoff points for the alarm signal, respectively. When r ≥ 0.8, the cigarette weight is considered stable. When 0.8 > r ≥ 0.6, the cigarette weight is reported as an orange warning signal. When r < 0.6, the cigarette weight is reported as a red warning signal.

[0066] In order to facilitate understanding of the aforementioned embodiments, specific examples are provided below for illustration.

[0067] Example 1: First, set the sliding window N to 50 and calculate the normal distribution fitting curve array Y of the cigarette weight within the sliding window, such as Figure 2 The solid line shows this. Next, the mean and standard deviation of the cigarette weight within the sliding window are calculated, which are 0.656 and 0.014, respectively. The cigarette weight process standard (0.61-0.71) is divided into 100 equal parts to obtain the simulated value, i.e., K = 100. The simulated normal distribution curve array F is calculated according to the normal distribution formula, as shown in the following example: Figure 2 As shown by the dotted line, the normal distribution fitting curve array Y and the simulated normal distribution curve array F of the cigarette weight in the sliding window were normalized and correlation analysis was performed, and the correlation coefficient was 0.88. Finally, it was determined to be stable according to the stability judgment standard.

[0068] Example 2: First, set the sliding window N to 50 and calculate the normal distribution fitting curve array Y of the cigarette weight within the sliding window, such as Figure 3 The solid line shows this. Next, the mean and standard deviation of the cigarette weight within the sliding window are calculated, which are 0.652 and 0.0186, respectively. The cigarette weight process standard (0.61-0.71) is divided into 100 equal parts to obtain the simulated value, i.e., K = 100. The simulated normal distribution curve array F is calculated according to the normal distribution formula, as shown in the following example: Figure 3As shown by the dotted line; the normal distribution fitting curve array Y and the simulated normal distribution curve array F of the cigarette weight in the sliding window are normalized respectively, and a correlation analysis is performed, and the correlation coefficient is 0.70. Finally, it is determined as an orange warning based on the stability judgment standard.

[0069] In summary, the main design concept of the present invention lies in setting a sliding window for the cigarette weight data collected during online automatic sampling and testing, determining the actual normal distribution fitting curve of the cigarette weight data obtained within the sliding window, and obtaining a simulated normal distribution curve for the actual normal distribution fitting curve and the simulated normal distribution curve. Correlation analysis is then performed between the actual normal distribution fitting curve and the simulated normal distribution curve, and the control stability of the cigarette weight indicator is determined based on the correlation between the two. This invention can reflect the stability of cigarette weight during the production process, facilitate timely detection of anomalies, alert operators to follow up, and prevent unqualified cigarettes from entering the next process.

[0070] In the embodiment of the present invention, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can be represented by: a, b, c, a and b, a and c, b and c, or a, b and c, where a, b, c can be single or multiple.

[0071] The above describes in detail the structure, features and effects of the present invention based on the embodiments shown in the drawings, but the above is only a preferred embodiment of the present invention. It should be noted that the technical features involved in the above embodiments and their preferred modes can be reasonably combined and matched into a variety of equivalent schemes by those skilled in the art without departing from or changing the design ideas and technical effects of the present invention; therefore, the scope of implementation of the present invention is not limited to what is shown in the drawings. Any changes made in accordance with the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which still do not exceed the spirit covered by the description and drawings, should be within the scope of protection of the present invention.

Claims

1. A method for online identification of cigarette weight stability, characterized in that: include: Sampling the weight data of the cigarettes detected online using a preset sliding window includes: using a sliding window of constant length to sample the cigarettes online at preset intervals to obtain the cigarette weight data, and moving the sliding window according to a predetermined step length to obtain new cigarette weight data; Obtaining a normal distribution fitting curve of actual cigarette weight based on cigarette weight data within the sliding window; Obtaining a simulated normal distribution curve of cigarette weights based on cigarette weight data within a sliding window, including: obtaining a weight mean and a standard deviation based on the cigarette weight data within the sliding window; utilizing the weight mean and the standard deviation in combination with a normal distribution algorithm to obtain the simulated normal distribution curve, specifically including: dividing a preset cigarette weight process standard range into a plurality of equal portions based on the weight mean and the standard deviation, and obtaining a corresponding plurality of simulated value arrays; performing a correlation analysis on the actual normal distribution fitting curve and the simulated normal distribution curve, including: normalizing the actual normal distribution fitting curve and the simulated normal distribution curve; performing a correlation analysis on the normalized array of the actual normal distribution fitting curve and the array of the simulated normal distribution curve to obtain a correlation coefficient; Based on the results of the correlation analysis, the stability of the cigarette weight is determined, including: using the comparison relationship between the numerical correlation analysis results and different preset thresholds to determine whether the cigarette weight is stable or trigger preset warnings of different levels.

2. The method for online identification of cigarette weight stability according to claim 1, characterized in that: The method of comparing the numerical correlation analysis results with different preset thresholds to determine whether the cigarette weight is stable or triggering preset warning levels includes: If the result of the correlation analysis is numerically greater than or equal to the first threshold, it is determined to be a stable state; If the result of the correlation analysis is numerically less than the first threshold and greater than or equal to the second threshold, a lower level warning is triggered; If the result of the correlation analysis is numerically smaller than a second threshold, a higher level warning is triggered, wherein the first threshold is greater than the second threshold.

Images