阅读说明：本技术 基于加工前后同步近红外分析的烟草热加工强度及波动性在线监测方法 (Tobacco hot processing strength and volatility online monitoring method based on synchronous near-infrared analysis before and after processing ) 是由 朱文魁 范子彦 陈良元 王兵 李斌 刘朝贤 李军 于 2020-08-26 设计创作，主要内容包括：一种基于加工前后同步近红外分析的烟草热加工强度及波动性在线监测方法,特征是,提出了一种反映物料热加工过程质量稳定性的监控指标,采用该指标随批次加工时间变化而生成的时间序列图,在线实时监控烟草加工质量稳定性的变化。其优点是：1、对烟草物料经历热加工前后的近红外光谱数据进行时间同步化处理,可尽量消除加工过程配方烟草物料质量波动对热加工强度实时评价的影响。2、通过提取出烟草红外光谱中对热加工最为敏感的波段数据,可有效提升后续计算得到的类间距离对热加工条件变化的灵敏度。能够区分出不同热加工条件下类间距离的变化；实现烟草热加工强度的在线实时监测,同时可实现对各批次烟草物料热加工强度及其波动性的定量评价。(An on-line monitoring method for the hot processing strength and volatility of tobacco based on the synchronous near infrared analysis before and after processing is characterized by providing a monitoring index for reflecting the quality stability of the hot processing process of materials, and adopting a time sequence chart generated by the index along with the change of batch processing time to monitor the change of the quality stability of the tobacco processing on line in real time. The advantages are that: 1. the time synchronization processing is carried out on the near infrared spectrum data before and after the tobacco material is subjected to the thermal processing, so that the influence of the quality fluctuation of the tobacco material in the formula in the processing process on the real-time evaluation of the thermal processing strength can be eliminated as much as possible. 2. By extracting the band data which is most sensitive to thermal processing in the tobacco infrared spectrum, the sensitivity of the inter-class distance obtained by subsequent calculation to the change of thermal processing conditions can be effectively improved. The change of the distance between the classes under different hot processing conditions can be distinguished; the on-line real-time monitoring of the hot processing strength of the tobacco is realized, and meanwhile, the quantitative evaluation of the hot processing strength and the volatility of each batch of tobacco materials can be realized.)

1. A tobacco hot working intensity and volatility on-line monitoring method based on synchronous near infrared analysis before and after processing is characterized in that: respectively installing online near-infrared detection devices before and after the thermal processing production process, and detecting the near-infrared spectrum of the tobacco material before thermal processing and the near-infrared spectrum of the tobacco material after thermal processing on line in real time; according to the transmission time of the tobacco material between the two detection points, the computer carries out time synchronization processing on near infrared spectrum data before and after the tobacco material is subjected to thermal processing; based on the infrared spectrum of the two synchronous detection points, respectively extracting characteristic waveband data, and calculating and determining the inter-class distance of the infrared spectrum data of the waveband before and after processing by adopting a linear discriminant analysis method; the distance between classes is used as a parameter for reflecting the hot processing intensity of the tobacco, the computer calculates the distance between classes in different processing time intervals in a rolling mode in real time, and the stability of the hot processing quality of the tobacco is monitored in real time by adopting a trend graph of the index changing along with the processing time.

2. The on-line monitoring method according to claim 1, wherein: the time synchronization treatment specifically refers to the time t required by the transmission of the tobacco material between two infrared spectrum detection points before and after the thermal processing₀In time series, the infrared spectrum data of the material before thermal processing at the time t is compared with (t + t)₀) And aligning the infrared spectrum data of the materials subjected to thermal processing one by one.

3. The on-line monitoring method according to claim 1, wherein: the characteristic wave band data are extracted, and specifically, the infrared spectrum data in the wavelength ranges of 1360-1480nm and 1100-1200nm are extracted according to the difference of the influence degree of the thermal processing process on each wave band in the full wavelength range of the infrared spectrum of the tobacco.

4. The on-line monitoring method according to claim 1, wherein: the method comprises the steps of adopting a linear discriminant analysis method to calculate and determine the inter-class distance of infrared spectrum data of the wave band before and after processing, and specifically projecting the extracted characteristic wave band data of the infrared spectrum of the two types before and after processing on a one-dimensional straight line, so that the projection point of each type of the spectrum data is as close as possible, and the distance between the centers of the two types of the spectrum data meets the conditions that the intra-class variance is as minimum as possible and the inter-class variance is maximum after projection, and the inter-class distance is the inter-class variance of the two types of the spectrum data after projection.

5. The on-line monitoring method according to claim 1, wherein: the different processing time intervals can be specifically set within the range of 10s-120s according to the production monitoring frequency requirement, and the primary inter-class distance is output at each time interval delta t to reflect the tobacco hot processing strength D in the processing time period.

6. The on-line monitoring method according to claim 1, wherein: the hot processing procedure comprises tobacco shred roller drying, tobacco shred airflow drying, stem fluidized bed drying, tobacco flake roller moisture regaining and tobacco flake vacuum moisture regaining.

Technical Field

The invention relates to an online monitoring method for tobacco hot processing strength volatility based on synchronous near-infrared analysis before and after processing, belonging to the technical field of tobacco processing.

Background

The heat processing process of the tobacco raw materials runs through the whole tobacco processing technology from primary baking, threshing and redrying after tobacco leaf picking to tobacco shred manufacturing, and is also the key point of continuous attention in the field of tobacco technology research. The technological function of the tobacco heat processing process is that on one hand, the temperature and humidity state of a product is adjusted through a heat treatment environment of drying (primary baking, secondary baking, cut tobacco drying and the like) or warming and humidifying (leaf moistening, loosening and moisture regaining and the like) so as to meet the subsequent processing requirements; on the other hand, because the tobacco as a heat-sensitive material is easy to generate a series of thermophysical and chemical change processes under the heat treatment environment, the thermal processing procedure has a remarkable effect on the physical and chemical properties of the treated tobacco. In this respect, a great deal of research carried out in the tobacco processing field based on the process evaluation method in the past also shows that the hot processing process in the tobacco shred making process is a key processing link influencing the physical and chemical properties of tobacco, and the change degree of the physical and chemical properties is closely related to the process processing strength.

However, an effective quantitative characterization method for the hot working strength of the hot working procedure in the tobacco production process of the cigarette enterprise is still lacked at present, and the strength of the hot working strength is mostly reflected by the setting level of each processing parameter in the existing procedure evaluation method, so that the hot working strength corresponding to different process parameters is lacked in comparability, and the on-line monitoring and evaluation of the hot working strength of the tobacco in the production process of the cigarette enterprise are also difficult to realize. Based on this, the invention patent CN106690391A proposes a method for regulating consistency of multipoint processing strength of a product dried by a tobacco roller, in which a tobacco material roller heat and moisture treatment online monitoring system is used to detect information of tobacco shred temperature change in the roller drying process, and then the heat processing strength is expressed in the form of integral of material temperature change. The detection method and the device adopted by the invention need to be installed and disassembled in the roller equipment, are inconvenient to operate and use, have influence on normal production, and are only suitable for the tobacco roller drying process. Similarly, the invention patent CN110286197A proposes a method for representing consistency of tobacco processing strength in a drum drying process, in which a moving temperature measuring device is used to detect information about temperature change of tobacco in the drum drying process, and then a function of the inlet and outlet temperature and water content of tobacco is used to express its hot processing strength. The motion detection device adopted by the invention has difference with the motion behavior of the tobacco shreds in the roller, the measured temperature is difficult to truly reflect the temperature change of the solid-phase tobacco shreds, and meanwhile, the method is only suitable for the drying process of the tobacco roller.

On the other hand, tobacco, which is a typical heat-sensitive material, inevitably undergoes a change in chemical properties during heat processing. Therefore, the change of the spectral characteristics of the tobacco before and after thermal processing can be detected and analyzed by adopting a near infrared spectrum detection means so as to reflect the change degree of the chemical characteristics of the tobacco after thermal processing and further reflect the thermal processing strength of the tobacco. In this respect, the invention patent CN201310131867.6 and the invention patent cn200910059487.x respectively provide a method for accurately characterizing the material processing strength in a tobacco shred drying process and a near-infrared characterization method for quality change in a tobacco shred manufacturing process. The invention patent CN201310131867.6 collects the cut tobacco before and after drying respectively, carries out spectrum scanning after balancing, and obtains the inter-class average value of the mahalanobis distance from each drying processing gradient cut tobacco to the non-drying cut tobacco class model through principal component analysis and calculation so as to reflect the processing intensity. The invention patent CN200910059487.X is to prepare a blank sample and a comparison sample respectively before and after adjusting the technological parameters in the tobacco shred manufacturing process; scanning the ground blank sample and the comparison sample to obtain respective near-infrared spectrograms; and subtracting the near infrared spectrogram of the blank sample from the near infrared spectrogram of the comparison sample to obtain a near infrared difference spectrogram, and reflecting the internal quality change of the tobacco in the thermal processing process by the change of the near infrared difference spectrogram. The two invention patents have the common problem that the practical application is difficult to meet, firstly, the tobacco material is agricultural products, and the tobacco processing objects are mixed formulas formed by tobacco leaves of different grades, so that the quality of the tobacco products at each processing point can fluctuate along with time sequence due to factors such as formula proportion fluctuation at different moments in practical processing, and the fluctuation can cover the influence of the thermal processing process on the tobacco quality. Secondly, in the infrared spectrum of the tobacco material, not all wave bands are sensitive to the thermal processing procedure, and actually only individual wave bands can be obviously changed along with the adjustment of the thermal processing conditions. Finally, the two inventive methods are off-line detection methods, and cannot be monitored on line.

Disclosure of Invention

The invention aims to solve the problems of the method for detecting and analyzing the processing quality stability of the tobacco in the thermal processing process, and provides an online monitoring method for the volatility of the thermal processing strength of the tobacco based on synchronous near-infrared analysis before and after processing. The invention provides a monitoring index for reflecting the quality stability of a material thermal processing process, which is defined as the distance obtained by linear discriminant analysis of synchronously analyzed characteristic waveband data.

The purpose of the invention is realized by the following technical scheme:

an online monitoring method for fluctuation of tobacco hot processing strength based on synchronous near-infrared analysis before and after processing is characterized in that online near-infrared detection devices are respectively installed before and after a hot processing production procedure, and a near-infrared spectrum of a tobacco material before being subjected to hot processing and a near-infrared spectrum of the tobacco material after being subjected to hot processing are detected in real time on line; according to the transmission time of the tobacco material between the two detection points, the computer carries out time synchronization processing on near infrared spectrum data before and after the tobacco material is subjected to thermal processing; based on the infrared spectrum of the two synchronous detection points, respectively extracting characteristic waveband data, and calculating and determining the inter-class distance of the infrared spectrum data of the waveband before and after processing by adopting a linear discriminant analysis method; the distance between classes is used as a parameter for reflecting the hot processing intensity of the tobacco, the computer calculates the distance between classes in different processing time intervals in a rolling mode in real time, and the stability of the hot processing quality of the tobacco is monitored in real time by adopting a trend graph of the index changing along with the processing time.

The time synchronization treatment specifically refers to aligning infrared spectrum data of the material before thermal processing at the time t and infrared spectrum data of the material after thermal processing at the time (t + t 0) one by one on a time sequence according to the time t0 required by transmission of the tobacco material between two infrared spectrum detection points before and after thermal processing. After the processing method, the infrared spectrum data before and after processing are basically from the material at the same processing point when the inter-class distance is calculated, so that the influence of the quality fluctuation of the tobacco material in the processing process on the real-time evaluation of the hot processing strength can be eliminated as much as possible.

The characteristic wave band data are extracted, and specifically, the infrared spectrum data in the wavelength ranges of 1360-1480nm and 1100-1200nm are extracted according to the difference of the influence degree of the thermal processing process on each wave band in the full wavelength range of the infrared spectrum of the tobacco. By extracting the wave band data which is most sensitive to thermal processing in the tobacco infrared spectrum, the sensitivity of the inter-class distance obtained by subsequent calculation to the change of thermal processing conditions can be effectively improved.

The method comprises the steps of adopting a linear discriminant analysis method to calculate and determine the inter-class distance of infrared spectrum data of the wave band before and after processing, and specifically projecting extracted characteristic wave band data of the infrared spectrum of the two classes before and after processing on a one-dimensional straight line, so that the projection point of each class of spectrum data is as close as possible, and the distance between the centers of the two classes of spectrum data meets the conditions that the intra-class variance is as small as possible and the inter-class variance is as large as possible after projection. The inter-class distance is the inter-class variance of the two types of spectral data after projection. When the linear discriminant analysis method is used for analyzing the characteristic waveband data of the infrared spectrums before and after processing, compared with other analysis methods such as principal component analysis, the method can better distinguish the change of the distance between the infrared spectrums under different thermal processing conditions.

The different processing time intervals can be specifically set within the range of 10s-120s according to the production monitoring frequency requirement, and the primary inter-class distance is output at each time interval delta t to reflect the tobacco hot processing strength D in the processing time period.

The method comprises the following specific steps:

1. respectively arranging online near-infrared detection devices at a material incoming section and a material outgoing section of a specific hot processing procedure;

2. under the normal production condition, detecting and recording the transmission time t0 of the tobacco material between two detection points;

3. the process is normal in material passing production, and a near infrared detection device detects near infrared spectrum data of incoming tobacco and discharged tobacco in the processing process in real time on line;

4. the computer collects the near infrared spectrum data of the tobacco which is fed and discharged in real time according to the time sequence and continuously extracts the characteristic infrared spectrum data in the wavelength ranges of 1360-;

5. aligning the infrared spectrum data of the material before thermal processing at the time t and the infrared spectrum data of the material after thermal processing at the time (t + t 0) one by one on a time sequence;

6. grouping and recording the extracted near infrared spectrum characteristic waveband data of the incoming material tobacco according to a set time interval delta t; meanwhile, the extracted characteristic wave band data of the discharged material near infrared spectrum are processed in the same way;

7. calculating and determining the distance between the two groups of data before and after hot working, which are detected and obtained in the delta t time interval, by adopting a linear discriminant analysis method;

8. in the production process, every time delta t time interval passes, the computer calculates the primary inter-class distance in real time according to the collected feeding and discharging near-infrared detection data, reflects the thermal processing strength D of the tobacco material at the moment, and monitors the thermal processing quality stability of the tobacco in real time by adopting a trend chart of the index changing along with the processing time;

9. when the batch production is finished, the computer statistically analyzes the mean value and the variation coefficient of the distance between the output classes in the batch processing process so as to quantitatively reflect the hot processing intensity and the fluctuation of the batch materials.

The invention has the following advantages:

1. the time synchronization processing is carried out on the near infrared spectrum data before and after the tobacco material is subjected to the thermal processing, so that the influence of the quality fluctuation of the tobacco material in the formula in the processing process on the real-time evaluation of the thermal processing strength can be eliminated as much as possible.

2. In the real-time analysis of the infrared spectrum of the tobacco, by extracting the band data which is most sensitive to the thermal processing in the infrared spectrum of the tobacco (as shown in fig. 1, the difference of the band data of 1360 + 1480nm and 1100 + 1200nm in the infrared spectrum first-order wave spectrum after the thermal processing of the tobacco is the largest), the sensitivity of the inter-class distance obtained by the subsequent calculation to the change of the thermal processing condition can be effectively improved.

3. When the linear discriminant analysis method is used for analyzing the infrared spectrum characteristic waveband data before and after processing, the change of the distance between the classes under different thermal processing conditions can be better distinguished.

4. The on-line real-time monitoring of the hot processing strength of the tobacco in the hot processing process can be realized, and the quantitative evaluation of the hot processing strength and the volatility of each batch of tobacco materials can be realized.

Drawings

FIG. 1 shows the extraction of characteristic bands of infrared spectra.

FIG. 2 is a schematic diagram of the linear discriminant analysis method for determining the distance between classes before and after drying.

FIG. 3 shows the monitoring and analysis results of the D value of the hot working strength of the tobacco shred roller drying process.

Detailed Description

The invention is further described by the following specific examples and by the accompanying drawings, but the invention content is not limited thereto.

Aiming at the tobacco shred roller drying process in the tobacco shred production line of a cigarette factory, an on-line monitoring method for the hot processing strength and the volatility of tobacco is established, and the main process steps are as follows:

1. aiming at a first-zone production line for making cut tobacco at a certain cigarette processing point, an online near-infrared detection device is respectively arranged in front of an HT (high-temperature) device at a feeding section and behind an on-site winnowing device at a discharging section of a cut tobacco roller drying process;

2. detecting the transmission time of the tobacco material between two detection points under normal production conditions, and recording as 425 s;

3. the process is normal in material passing production, and a near infrared detection device detects near infrared spectrum data of incoming cut tobacco and discharged cut tobacco in the processing process in real time on line;

4. the computer collects the near infrared spectrum data of the tobacco which is fed and discharged in real time according to the time sequence and continuously extracts the characteristic infrared spectrum data in the wavelength ranges of 1360-;

5. aligning the infrared spectrum data of the material before thermal processing at the time t and the infrared spectrum data of the cut tobacco after cut tobacco drying at the time (t + 425 s) one by one on a time sequence;

6. grouping and recording the extracted near infrared spectrum characteristic wave band data of the incoming cut tobacco according to a set time interval of 30 s; meanwhile, the extracted characteristic wave band data of the discharged material near infrared spectrum are processed in the same way;

7. calculating and determining the distance between two groups of data before and after the cut tobacco drying, which are detected and obtained within 30s time interval, by adopting a linear discriminant analysis method;

8. in the production process, every 30s time interval, the computer calculates the primary inter-class distance in real time according to the collected feeding and discharging near-infrared detection data, reflects the thermal processing strength D of the tobacco material at the moment, and monitors the drying thermal processing quality stability of the tobacco roller in real time by adopting a trend chart of the index changing along with the processing time;

9. when the batch production is finished, the computer statistically analyzes the mean value of the distance between the output classes in the batch production process

And a coefficient of variation CV value for quantitatively reflecting the hot working strength and the volatility of the batch material. The results are shown in FIG. 3.