WO2020098261A1 - 一种控制烘丝入口含水率的方法和系统 - Google Patents
一种控制烘丝入口含水率的方法和系统 Download PDFInfo
- Publication number
- WO2020098261A1 WO2020098261A1 PCT/CN2019/089598 CN2019089598W WO2020098261A1 WO 2020098261 A1 WO2020098261 A1 WO 2020098261A1 CN 2019089598 W CN2019089598 W CN 2019089598W WO 2020098261 A1 WO2020098261 A1 WO 2020098261A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- moisture
- moisture content
- stage
- loose
- variable
- Prior art date
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A24—TOBACCO; CIGARS; CIGARETTES; SIMULATED SMOKING DEVICES; SMOKERS' REQUISITES
- A24B—MANUFACTURE OR PREPARATION OF TOBACCO FOR SMOKING OR CHEWING; TOBACCO; SNUFF
- A24B3/00—Preparing tobacco in the factory
- A24B3/04—Humidifying or drying tobacco bunches or cut tobacco
Definitions
- the invention relates to the technical field of wire baking, in particular to a method and system for controlling the moisture content of the wire inlet.
- the control of tobacco moisture content is an important issue of concern to the tobacco industry.
- the main difficulty is that it is difficult to control the moisture content of each process section before drying the silk. Ratio, steam ratio, etc.
- the pre-production work orders of tobacco yarns often only issue a water addition ratio, and the water addition ratio is often set by experienced employees. Different employees may have different settings.
- the water addition ratio is often set by experienced employees. Different employees may have different settings.
- Due to many factors that affect the moisture content Leading to unstable tobacco quality. Therefore, it is necessary to propose a low-cost, safe, and convenient intelligent assistant decision-making scheme to help the tobacco silk workshop to provide a more accurate and reliable water addition ratio, so as to achieve the control of the moisture content of the baked silk inlet.
- the present invention provides a method and system for controlling the moisture content of the drying wire inlet. It does not change any structure and principle of the silk production line, does not add additional measuring points, and does not affect normal production. Learning methods, providing safe, convenient, and reasonable water addition ratio suggestions, and predicting the changes in water content at various stages, as well as the water content prediction at the four positions of the intermediate loose outlet, the moistening leaf inlet, the moistening leaf outlet, and the drying wire inlet In order to achieve the control of the moisture content of the entrance of the baking wire.
- a method for controlling the moisture content of the wire inlet collecting the basic operating conditions of historical operations, and establishing different working models for different basic operating conditions.
- the basic working conditions include the corresponding parameters of the main factors that affect the moisture content of the wire inlet at the processing stage before the wire drying inlet.
- the processing stage includes: a loosening stage, a premix cabinet stage, a moistening leaf feeding stage, and a leaf storage stage, where
- the main factors of the loose phase include: water addition ratio, steam ratio, loose return air temperature, loose material flow rate, air pressure to loose feed ratio, loose ambient temperature and humidity, and loose moisture change;
- the premix cabinet stage The main factors include: the environmental temperature and humidity of the premixing cabinet, the length of the premixing cabinet, and the amount of moisture change in the premixing cabinet;
- the main factors in the leaf-wetting and feeding stage include: the return air temperature of the feed, the flow rate of the feed material, the actual proportion of the feed steam compensation, The proportion of air pressure to feed, the environmental temperature and humidity of the feed, and the moisture change of the leaf feed;
- the main factors of the leaf storage stage include: the temperature and humidity of the leaf storage cabinet, the temperature and humidity of the leaf storage, the length of the leaf storage cabinet, the storage Changes in leaf moisture.
- the machine learning model includes a coding unit, an optimization target, and variable coding; wherein, the coding unit is the main information unit of the model and a knowledge point of machine learning, including basic working condition information and each group of basic The zero value of the loose outlet moisture meter corresponding to the working condition information, the zero value of the feed inlet moisture meter, the zero value of the feed outlet moisture meter, and the zero value of the dry wire inlet moisture meter; controlling the moisture content of the dry wire inlet is the optimization goal; the variable coding It is a coding of model positioning, which is calculated by each variable in each stage to realize the mapping of each variable to the model, that is, the corresponding model can be quickly found according to the variable, and the variable is one or more of the basic working condition information
- the working model corresponds to the coding unit through variable coding, and the corresponding relationship instructs the machine learning task to which coding unit the working model is assigned, or from which coding unit the current working model obtains historical operation information.
- the beginning stage is often a gradual process, which is the dynamic data of the "U", which cannot stabilize the relationship between the actual variables and related factors.
- the previous N Minute data is not processed for variable coding, that is, the unstable data in the production line is eliminated during the learning process of the machine learning model, and is not processed for variable coding.
- the unstable data includes data at the beginning of the production line work 1. The data within the production interruption period and the abnormal data far beyond the normal range set, the data screening greatly improves the accuracy of the final optimization plan.
- the machine learning model establishes different models according to different processing stages to form a multi-stage tandem model, and conducts Pearson correlation analysis for each processing stage to screen out variables with strong correlations.
- the selected variables are subjected to variable coding. That is, in each processing stage, the variables with less correlation are not considered in the model.
- the model only considers how to optimize the variables with greater correlation to achieve the control of the moisture content of the baking inlet, which greatly improves the prediction efficiency and prediction accuracy.
- the variables that can be obtained before the production line work are used to predict the moisture change of each processing stage, and then through the moisture change of each stage, the moisturizing leaf feed material is reversed from the set value of the moisture content of the baked silk inlet.
- the water content at the outlet, the water content at the inlet of the moistening leaf, the water content at the loose outlet, and the amount of water change during the loose phase are used to predict the moisture change of each processing stage, and then through the moisture change of each stage, the moisturizing leaf feed material is reversed from the set value of the moisture content of the baked silk inlet.
- variable code is stored as an integer data type, which represents the dependent variables of each stage such as the proportion of water added, the proportion of steam, the return air temperature, the material flow rate, the proportion of compressed air to loose feed, the temperature and humidity of the loose environment, etc.
- variable coding sample corresponding to the current variable, it is used directly. If it does not exist, it is calculated, predicted, and generated by the method of least squares, linear regression, and support vector machine. When the same situation is encountered again It can reduce the training process and directly match the historical value as the prediction of the water addition ratio, which greatly improves the efficiency and improves the accuracy.
- the calculation formula for collecting the moisture change in each processing stage is:
- the loose water change amount (loose outlet moisture content-loose outlet moisture meter zero value)-default moisture content of loose inlet;
- the amount of moisture change in the premixing cabinet (moisture content at the inlet of the leaf-zero value of the moisture meter at the leaf inlet)-(moisture content at the loose outlet-zero value of the moisture meter at the loose outlet);
- the amount of change in the moisture content of the moisturizing leaf feed (moisture content at the moistening leaf outlet-zero point of the moisturizing leaf outlet moisture meter)-(moisture content at the moistening leaf inlet-zero point value of the moisturizing leaf inlet moisture meter);
- the leaf storage moisture change amount (moisture content at the drying wire inlet-zero value of the moisture meter at the drying wire inlet)-(moisture content at the outlet of the moistening leaf-zero value of the moisture meter at the moistening leaf outlet);
- the water content is measured by a moisture meter.
- Moisture content at the outlet of the moist leaf (moisture content at the entrance of the dried wire-zero value of the moisture meter at the entrance of the dried wire)-change in moisture content of the stored leaf + zero value of the moisture meter at the outlet of the moistened leaf;
- Moisture content at the inlet of the moist leaf (moisture content at the outlet of the moist leaf-zero value of the moisture meter at the moist leaf outlet)-amount of moisture change at the moist leaf + zero value of the moisture meter at the moist leaf inlet;
- Loose outlet moisture content (moisture inlet moisture content-zero moisture value of the moisturizer inlet)-moisture change of the premixing cabinet + zero value of the loose moisture meter;
- Loose moisture change (loose outlet moisture content-loose outlet moisture meter zero value)-default moisture content of loose inlet.
- the machine learning model establishes a variable coding traceability mechanism.
- Each optimization scheme can be traced to the source of the variable coding.
- the basis for users to query the optimization scheme is such as brand, batch, time, production line, basic working conditions, The operation status and other information make the optimization plan more reasonable and safe and credible.
- a system for controlling the moisture content of the drying wire inlet includes a basic working condition information acquisition module and a machine learning model.
- the basic working condition information collection module includes a detection device installed on the silk production line.
- the basic working condition information includes the corresponding parameters of the main factors that affect the moisture content of the wire inlet at the processing stage before the wire drying inlet.
- the processing stage includes: a loosening stage, a pre-mixing cabinet stage, a moistening leaf feeding stage, and a leaf storage stage;
- the main factors of the loose phase include: water addition ratio, steam ratio, loose return air temperature, loose material flow rate, air pressure to loose feed ratio, loose ambient temperature and humidity, loose moisture change;
- the premix cabinet stage The main factors include: the ambient temperature and humidity of the premixing cabinet, the length of the premixing cabinet, and the amount of water change in the premixing cabinet;
- the main factors of the leaf replenishing stage include: the return air temperature of the feed, the flow rate of the feed material, and the actual proportion of the feed steam compensation 1.
- the proportion of air pressure to feed, the environmental temperature and humidity of the feed, and the moisture change of the leaf feed; the main factors of the leaf storage stage include: the temperature and humidity of the leaf storage cabinet, the temperature and humidity of the leaf storage cabinet, the length of the leaf storage cabinet, Leaf storage moisture change amount; one basic working condition information corresponds to one working model;
- the machine learning model includes variable coding, coding unit and optimization goal.
- the coding unit includes basic working condition information and a loose outlet moisture meter zero value corresponding to each group of basic working condition information, a feed inlet moisture meter zero value, a feed outlet moisture meter zero value, and a bake wire inlet moisture meter zero value.
- the moisture content of the inlet of the baked wire is controlled to the optimization goal.
- the working model has a one-to-one correspondence with the coding unit through variable coding, and the variable coding is calculated by one variable or a combination of multiple variables in the basic working condition information to realize the mapping of the basic working condition information to the coding unit and indicating
- the variable coding is obtained, the coding unit is obtained, the water change of the three stages of the premixing cabinet stage, the wet leaf feeding stage and the leaf storage stage is calculated, and finally the water addition ratio in the loose stage is obtained To generate an optimization plan.
- the machine learning model establishes different models according to different processing stages to form a multi-stage tandem model, and conducts Pearson correlation analysis for each processing stage to screen out variables with strong correlations.
- the selected variables are subjected to variable coding.
- the present invention has the following advantages:
- variable coding a large number of historical data samples are compiled into a non-repetitive, high-value coding library, which greatly reduces the memory space and improves the efficiency of model training.
- FIG. 1 is a schematic diagram of an implementation of a method for controlling the moisture content of a baked wire inlet of the present invention
- FIG. 2 is a schematic diagram of the composition of the basic working condition information of the present invention.
- Embodiment 1 a method for controlling the moisture content of the inlet of a bake wire, please refer to FIG. 1 to collect the basic working condition information of historical operation, and different working models are established for different basic working conditions;
- the machine learning model includes a coding unit, an optimization target, and variable coding; control the moisture content of the drying wire inlet to the optimization target, and the working model corresponds to the coding unit through variable coding;
- the following table is a test and inspection effect table.
- the basic working conditions include the corresponding parameters of the main factors that affect the moisture content of the wire inlet at the processing stage before the wire drying inlet.
- the process stages include: the loosening stage, the pre-mixing cabinet stage, the wet leaf feeding stage and The leaf storage stage, wherein the main factors of the loose phase include: water addition ratio, steam ratio, loose return air temperature, loose material flow rate, air pressure to loose feed ratio, loose environment temperature and humidity, loose moisture change amount;
- the main factors in the pre-mixing cabinet stage include: the ambient temperature and humidity of the pre-mixing cabinet, the length of the pre-mixing cabinet, and the amount of moisture change in the pre-mixing cabinet;
- the main factors in the leaf-wetting and feeding stage include: the return air temperature of the feed, the flow rate of the feed material, and the feed
- the main factors of the leaf storage stage include: the temperature and humidity of the
- the coding unit is the main information unit of the model and is the knowledge point of machine learning, including the basic working condition information and the zero value of the loose outlet moisture meter corresponding to each group of basic working condition information, the zero value of the feed inlet moisture meter, and the feed outlet moisture
- the zero value of the instrument and the zero value of the moisture meter of the wire inlet; controlling the moisture content of the inlet of the wire drying is the optimization goal;
- the variable coding is a coding of the model positioning, which is calculated by each variable at each stage to realize the variable to the model Mapping, that is, you can quickly find the corresponding model according to the variable, which is one or more combinations of the basic working condition information;
- the working model corresponds to the coding unit through variable coding, and indicates the machine through the corresponding relationship To which coding unit the working model is assigned by the learning task, or from which coding unit the current working model obtains historical operation information.
- the calculation formula for collecting the moisture change in each processing stage is:
- the loose water change amount (loose outlet moisture content-loose outlet moisture meter zero value)-default moisture content of loose inlet;
- the amount of moisture change in the premixing cabinet (moisture content at the inlet of the leaf-zero value of the moisture meter at the leaf inlet)-(moisture content at the loose outlet-zero value of the moisture meter at the loose outlet);
- the amount of change in the moisture content of the moisturizing leaf feed (moisture content at the moistening leaf outlet-zero point of the moisturizing leaf outlet moisture meter)-(moisture content at the moistening leaf inlet-zero point value of the moisturizing leaf inlet moisture meter);
- the leaf storage moisture change amount (moisture content at the drying wire inlet-zero value of the moisture meter at the drying wire inlet)-(moisture content at the outlet of the moistening leaf-zero value of the moisture meter at the moistening leaf outlet);
- the water content is measured by a moisture meter.
- Moisture content at the outlet of the moist leaf (moisture content at the entrance of the dried wire-zero value of the moisture meter at the entrance of the dried wire)-change in moisture content of the stored leaf + zero value of the moisture meter at the outlet of the moistened leaf;
- Moisture content at the inlet of the moist leaf (moisture content at the outlet of the moist leaf-zero value of the moisture meter at the moist leaf outlet)-amount of moisture change at the moist leaf + zero value of the moisture meter at the moist leaf inlet;
- Loose outlet moisture content (moisture inlet moisture content-zero moisture value of the moisturizer inlet)-moisture change of the premixing cabinet + zero value of the loose moisture meter;
- Loose moisture change (loose outlet moisture content-loose outlet moisture meter zero value)-default moisture content of loose inlet.
- Embodiment 2 In the actual production of silk, the starting stage is often a gradual process, which is the dynamic data of the "U", which cannot stabilize the relationship between the actual variables and related factors.
- the previous N Minute data is not processed for variable coding, that is, the unstable data in the production line is eliminated during the learning process of the machine learning model, and is not processed for variable coding.
- the unstable data includes data at the beginning of the production line work 1. The data within the production interruption period and the abnormal data far beyond the normal range set, the data screening greatly improves the accuracy of the final optimization plan.
- Embodiment 3 The machine learning model establishes different models according to different processing stages to form a multi-stage tandem model, and conducts Pearson correlation analysis for each processing stage to screen out variables with strong correlations.
- the selected variables are subjected to variable coding. That is, in each processing stage, the variables with less correlation are not considered in the model.
- the model only considers how to optimize the variables with greater correlation to achieve the control of the moisture content of the baking inlet, which greatly improves the prediction efficiency and prediction accuracy.
- the variables that can be obtained before the production line work are used to predict the moisture change of each processing stage, and then through the moisture change of each stage, the moisturizing leaf feed material is reversed from the set value of the moisture content of the baked silk inlet.
- the water content at the outlet, the water content at the inlet of the moistening leaf, the water content at the loose outlet, and the amount of water change during the loose phase are used to predict the moisture change of each processing stage, and then through the moisture change of each stage, the moisturizing leaf feed material is reversed from the set value of the moisture content of the baked silk inlet.
- variable code is stored as an integer data type, which represents various stages such as water addition ratio, steam ratio, return air temperature, material flow rate, air pressure to loose feed ratio, loose environment temperature and humidity, etc.
- integer data type represents various stages such as water addition ratio, steam ratio, return air temperature, material flow rate, air pressure to loose feed ratio, loose environment temperature and humidity, etc.
- variable coding rounding function ((variable-minimum value of variable) / variable step).
- variable coding sample corresponding to the current variable, it is used directly. If it does not exist, it is calculated, predicted, and generated by the method of least squares, linear regression, and support vector machine. When the same situation is encountered again It can reduce the training process and directly match the historical value as the prediction of the water addition ratio, which greatly improves the efficiency and improves the accuracy.
- Embodiment 4 correspondingly, a system for controlling the moisture content of the drying wire inlet, with the goal of controlling the moisture content of the drying wire inlet, including a basic working condition information acquisition module and a machine learning model, different working models are established for different basic working conditions ,
- the machine learning model includes variable coding, coding unit and optimization goal.
- the working model corresponds to the coding unit through variable coding
- the variable coding is calculated by one variable or a combination of multiple variables in the basic working condition information, Realize the mapping of the basic working condition information to the coding unit, instruct the machine learning task to which coding unit the working model is assigned, and from which coding unit the current basic working condition information obtains the historical operation information.
- the basic working condition information collection module includes a detection device originally installed on the silk production line.
- the basic working condition information includes the corresponding parameters of the main factors that affect the moisture content of the wire inlet at the processing stage before the wire drying inlet.
- the processing stage includes: a loosening stage, a pre-mixing cabinet stage, a moistening leaf feeding stage, and a leaf storage stage;
- the main factors of the loose phase include: water addition ratio, steam ratio, loose return air temperature, loose material flow rate, air pressure to loose feed ratio, loose ambient temperature and humidity, loose moisture change;
- the premix cabinet stage The main factors include: the ambient temperature and humidity of the premixing cabinet, the length of the premixing cabinet, and the amount of water change in the premixing cabinet;
- the main factors of the leaf replenishing stage include: the return air temperature of the feed, the flow rate of the feed material, and the actual proportion of the feed steam compensation 1.
- the proportion of air pressure to feed, the environmental temperature and humidity of the feed, and the moisture change of the leaf feed; the main factors of the leaf storage stage include: the temperature and humidity of the leaf storage cabinet, the temperature and humidity of the leaf storage cabinet, the length of the leaf storage cabinet, The amount of stored leaf moisture changes.
- the coding unit includes basic working condition information and a loose outlet moisture meter zero value corresponding to each group of basic working condition information, a feed inlet moisture meter zero value, a feed outlet moisture meter zero value, and a bake wire inlet moisture meter zero value.
- the variable coding is obtained, the coding unit is obtained, the water change of the three stages of the premixing cabinet stage, the wet leaf feeding stage and the leaf storage stage is calculated, and finally the water addition ratio in the loose stage is obtained To generate an optimization plan.
Landscapes
- Manufacture Of Tobacco Products (AREA)
- Drying Of Solid Materials (AREA)
Abstract
一种控制烘丝入口含水率的方法和系统，在不改变制丝生产线任何结构和原理、不增加额外测点、不影响正常生产的前提下，通过采集历史操作的基础工况信息和机器学习的方法，提供安全、便捷、合理的加水比例建议，并预测出各阶段含水量的变化量，以及中间松散出口、润叶入口、润叶出口、烘丝入口这四个位置的含水率预测，从而达到烘丝入口含水率的控制。
Description
本发明涉及烘丝技术领域，尤其涉及一种控制烘丝入口含水率的方法和系统。
烟草含水率控制是烟草行业关注的重要课题，其主要难点在于烘丝之前各工艺段含水率控制难度较大，主要是影响烟草含水率的因素较多，如环境温湿度，留柜时长，加水比例、蒸汽比例等。
目前烟草制丝的生产前的工单往往只下达了一个加水比例，而该加水比例往往由有经验的员工设定，不同的员工，设定值可能不同，同时，由于影响含水率的因素多，导致烟草质量不稳定。因此有必要提出一种低成本、安全、便捷的智能化的辅助决策方案，帮助烟草制丝车间提供一个较准确可靠的加水比例，从而达到烘丝入口含水率的控制。
发明内容
本发明为解决上述技术问题，提供了一种控制烘丝入口含水率的方法和系统，在不改变制丝生产线任何结构和原理、不增加额外测点、不影响正常生产的前提下，通过机器学习的方法，提供安全、便捷、合理的加水比例建议，并预测出各阶段含水量的变化量，以及中间松散出口、润叶入口、润叶出口、烘丝入口这四个位置的含水率预测，从而达到烘丝入口含水率的控制。
一种控制烘丝入口含水率的方法，采集历史操作的基础工况信息，不同的基础工况建立不同的工作模型。所述基础工况包括烘丝入口前的加工阶段影响烘丝入口含水率的主要因素的相应参数，所述加工阶段包括：松散阶段、预混柜阶段、润叶加料阶段和贮叶阶段，其中，所述松散阶段的主要因素包括：加水比例、蒸汽比例、松散回风温度、松散物料流量、压空到松散的加料比例、松散环境温湿度、松散水分变化量；所述预混柜阶段的主要因素包括：预混柜的环境温湿度、预混柜时长、预混柜水分变化量；所述润叶加料阶段的主要因素包括：加料回风温度、加料物料流量、加料蒸汽补偿实际比例、压空到加料的加料比例、加料的环境温湿度、润叶加料水分变化量；所述贮叶阶段的主要因素包括：贮叶柜环境温湿度、贮叶环境温湿度、贮叶柜时长、贮叶水分变化量。
建立机器学习模型，所述机器学习模型包括编码单元、优化目标和变量编码；其中，所述编码单元是模型的主要信息单元，是机器学习的知识点，包括基础工况信息及与每组基础工况信息对应的松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值和 烘丝入口水分仪零点值；控制烘丝入口含水率为所述优化目标；所述变量编码是模型定位的一种编码，由各阶段各变量计算得出，实现各变量对模型的映射，即根据变量可以快速找到相应的模型，所述变量为基础工况信息中的一种或多种的组合；所述工作模型通过变量编码与编码单元一一对应，通过对应关系指示机器学习任务将工作模型派给哪个编码单元，或者当前工作模型从哪个编码单元获取历史操作信息。
给出下单时产线上各个测点获得的当前的加工阶段的基础工况信息、松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值、烘丝入口水分仪零点值和烘丝入口含水率设定值，得到变量编码，匹配到编码单元，计算得到预混柜阶段、润叶加料阶段和贮叶阶段三个阶段的阶段水分变化量，最后得到松散阶段的加水比例，生成优化方案。
可选的，在制丝实际生产中，开始阶段往往是个循序渐进的过程，是倒“U”的动态数据，无法稳定反应实际变量与相关因素的关系，为进一步提升优化方案的可靠度，前面N分钟数据是不加以处理为变量编码的，即所述机器学习模型的学习过程中剔除产线中非稳定数据，不加以处理为变量编码，所述非稳定数据包括产线工作起始阶段的数据、生产中断料时间段内的数据、远超设定的正常范围内的异常数据，数据的筛选大大提高了最后优化方案的精准性。
可选的，所述机器学习模型根据不同的加工阶段建立不同的模型，形成多阶段串联模型，并针对各个加工阶段分别进行Pearson相关性分析，筛选出相关性显著性强的变量，再对所筛选出的变量进行变量编码。即在各个加工阶段中，相关性较小的变量不在模型的考虑范围，模型只考虑如何优化相关性较大的变量来达到控制烘丝入口含水率，大大提高了预测效率和预测准确度。
在获得优化方案时，先通过产线工作前就能获得的变量去预测各个加工阶段的水分变化量，再通过各阶段水分变化量，从烘丝入口含水率设定值依次反推出润叶加料出口含水率、润叶加料入口含水率、松散出口含水率以及松散阶段水分变化量。
可选的，所述变量编码是整型的数据类型存储的，它代表了加水比例、蒸汽比例、回风温度、物料流量、压空到松散的加料比例、松散环境温湿度等各阶段因变量，具体计算公式为：变量编码＝取整函数((变量–变量最低值)/变量步长)。
生成优化方案时，如果存在当前变量对应的变量编码样本则直接使用，如果不存在则通过最小二乘法、线性回归、支持向量机方法去计算，预测，生成优化方案，当再次遇到相同情况时，可以减少训练的过程，直接匹配到历史值作为加水比例的预测，大大提高了效率，且精确度也有所提高。
可选的，采集各个加工阶段水分变化量的计算公式为：
所述松散水分变化量＝(松散出口含水率-松散出口水分仪零点值)-松散入口默认含水率；
所述预混柜水分变化量(润叶入口含水率-润叶入口水分仪零点值)-(松散出口含水率-松散出口水分仪零点值)；
所述润叶加料水分变化量＝(润叶出口含水率-润叶出口水分仪零点)-(润叶入口含水率-润叶入口水分仪零点值)；
所述贮叶水分变化量＝(烘丝入口含水率-烘丝入口水分仪零点值)-(润叶出口含水率-润叶出口水分仪零点值)；
其中，含水率通过水分仪测得。
在通过烘丝入口含水率逆推计算各个出入口含水率时，计算公式为：
润叶出口含水率＝(烘丝入口含水率-烘丝入口水分仪零点值)-贮叶水分变化量+润叶出口水分仪零点值；
润叶入口含水率＝(润叶出口含水率-润叶出口水分仪零点值)-润叶水分变化量+润叶入口水分仪零点值；
松散出口含水率＝(润叶入口含水率-润叶入口水分仪零点值)-预混柜水分变化量+松散出口水分仪零点值；
松散水分变化量＝(松散出口含水率-松散出口水分仪零点值)-松散入口默认含水率。
可选的，所述机器学习模型，建立变量编码溯源机制，每条优化方案都可追溯到变量编码的源头，用户查询该优化方案的依据如牌号、批次、时间、生产线、基础工况、操作状态等信息，使优化方案更具有合理性和安全可信性。
对应的，一种控制烘丝入口含水率的系统，以控制烘丝入口含水率为目标，包括基础工况信息采集模块和机器学习模型。
所述基础工况信息采集模块包括安装在制丝生产线上的检测装置。
所述基础工况信息包括烘丝入口前的加工阶段影响烘丝入口含水率的主要因素的相应参数，所述加工阶段包括：松散阶段、预混柜阶段、润叶加料阶段和贮叶阶段；其中，所述松散阶段的主要因素包括：加水比例、蒸汽比例、松散回风温度、松散物料流量、压空到松散的加料比例、松散环境温湿度、松散水分变化量；所述预混柜阶段的主要因素包括：预混柜的环境温湿度、预混柜时长、预混柜水分变化量；所述润叶加料阶段的主要因素包括：加料回风温度、加料物料流量、加料蒸汽补偿实际比例、压空到加料的加料比例、加料的环境温湿度、润叶加料水分变化量；所述贮叶阶段的主要因素包括：贮叶柜环境温湿度、贮叶环境 温湿度、贮叶柜时长、贮叶水分变化量；一种基础工况信息对应一种工作模型；
所述机器学习模型包括变量编码、编码单元和优化目标。
所述编码单元包括包括基础工况信息及与每组基础工况信息对应的松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值和烘丝入口水分仪零点值。
控制烘丝入口含水率为所述优化目标。
所述工作模型通过变量编码与编码单元一一对应，所述变量编码由基础工况信息中的一种变量或多种变量的组合计算得出，实现基础工况信息对编码单元的映射，指示机器学习任务将工作模型派给哪个编码单元以及当前基础工况信息从哪个编码单元获取历史操作信息。
给出下单时产线上各个测点获得的当前的加工阶段的基础工况信息、松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值、烘丝入口水分仪零点值和烘丝入口含水率设定值，得到变量编码，得到编码单元，计算得到预混柜阶段、润叶加料阶段和贮叶阶段三个阶段的阶段水分变化量，最后得到松散阶段的加水比例，生成优化方案。
可选的，所述机器学习模型根据不同的加工阶段建立不同的模型，形成多阶段串联模型，并针对各个加工阶段分别进行Pearson相关性分析，筛选出相关性显著性强的变量，再对所筛选出的变量进行变量编码。
可选的，所述变量编码是整型的数据类型存储的，具体计算公式如为：变量编码＝取整函数((变量–变量最低值)/变量步长)；，生成优化方案时，如果存在当前变量对应的变量编码样本则直接使用，如果不存在则通过最小二乘法、线性回归、支持向量机方法去计算，预测，生成优化方案。
由上述对本发明的描述可知，和现有技术相比，本发明具有如下优点：
1、通过历史操作，获取相应基础工况信息的加水比例，达到对烘丝入口水分的精确控制的目的；
2、建立多阶段串联模型，只考虑相关性强的因素，大大提高预测的效率；
3、记录历史操作，使得每个优化方案都可以追溯到源头，使优化方案更具有合理性和安全可信性。
4、通过变量编码，将大量的历史数据样本精编为一个不重复、高价值的编码库，大大降低了内存空间，且提高了模型训练效率。
此处所说明的附图用来提供对本发明的进一步理解，构成本发明的一部分，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。
其中：
图1是本发明一种控制烘丝入口含水率的方法的实现示意图；
图2是本发明基础工况信息的组成示意图。
为了使本发明所要解决的技术问题、技术方案及有益效果更加清楚、明白，以下结合附图和实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本发明，并不用于限定本发明。
实施例一：一种控制烘丝入口含水率的方法，请参阅图1，采集历史操作的基础工况信息，不同的基础工况建立不同的工作模型；
建立机器学习模型，所述机器学习模型包括编码单元、优化目标和变量编码；控制烘丝入口含水率为所述优化目标，所述工作模型通过变量编码与编码单元一一对应；
给出下单时产线上各个测点获得的当前的加工阶段的基础工况信息、松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值、烘丝入口水分仪零点值和烘丝入口含水率设定值，得到变量编码，匹配到编码单元，计算得到预混柜阶段、润叶加料阶段和贮叶阶段三个阶段的阶段水分变化量，最后得到松散阶段的加水比例，生成优化方案。
如下表为测试检验效果表，通过给定的烘丝入口含水率设定值，即可到推出润叶出口含水率、润叶入口含水率、松散出口含水率和加水比例，用所述方法预测出的水分变化量推出来的各出入口含水率与实测值误差极小，基本控制在0.5允差范围内，甚至大部分达到了0.2的优秀允差范围内，且加水比例也是与实测值偏差极小。
表1.测试检验效果
请参阅图2，所述基础工况包括烘丝入口前的加工阶段影响烘丝入口含水率的主要因素的相应参数，所述加工阶段包括：松散阶段、预混柜阶段、润叶加料阶段和贮叶阶段，其中，所述松散阶段的主要因素包括：加水比例、蒸汽比例、松散回风温度、松散物料流量、压空到松散的加料比例、松散环境温湿度、松散水分变化量；所述预混柜阶段的主要因素包括： 预混柜的环境温湿度、预混柜时长、预混柜水分变化量；所述润叶加料阶段的主要因素包括：加料回风温度、加料物料流量、加料蒸汽补偿实际比例、压空到加料的加料比例、加料的环境温湿度、润叶加料水分变化量；所述贮叶阶段的主要因素包括：贮叶柜环境温湿度、贮叶环境温湿度、贮叶柜时长、贮叶水分变化量。
所述编码单元是模型的主要信息单元，是机器学习的知识点，包括基础工况信息及与每组基础工况信息对应的松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值和烘丝入口水分仪零点值；控制烘丝入口含水率为所述优化目标；所述变量编码是模型定位的一种编码，由各阶段各变量计算得出，实现各变量对模型的映射，即根据变量可以快速找到相应的模型，所述变量为基础工况信息中的一种或多种的组合；所述工作模型通过变量编码与编码单元一一对应，通过对应关系指示机器学习任务将工作模型派给哪个编码单元，或者当前工作模型从哪个编码单元获取历史操作信息。
采集各个加工阶段水分变化量的计算公式为：
所述松散水分变化量＝(松散出口含水率-松散出口水分仪零点值)-松散入口默认含水率；
所述预混柜水分变化量(润叶入口含水率-润叶入口水分仪零点值)-(松散出口含水率-松散出口水分仪零点值)；
所述润叶加料水分变化量＝(润叶出口含水率-润叶出口水分仪零点)-(润叶入口含水率-润叶入口水分仪零点值)；
所述贮叶水分变化量＝(烘丝入口含水率-烘丝入口水分仪零点值)-(润叶出口含水率-润叶出口水分仪零点值)；
其中，含水率通过水分仪测得。
在通过烘丝入口含水率逆推计算各个出入口含水率时，计算公式为：
润叶出口含水率＝(烘丝入口含水率-烘丝入口水分仪零点值)-贮叶水分变化量+润叶出口水分仪零点值；
润叶入口含水率＝(润叶出口含水率-润叶出口水分仪零点值)-润叶水分变化量+润叶入口水分仪零点值；
松散出口含水率＝(润叶入口含水率-润叶入口水分仪零点值)-预混柜水分变化量+松散出口水分仪零点值；
松散水分变化量＝(松散出口含水率-松散出口水分仪零点值)-松散入口默认含水率。
实施例二，在制丝实际生产中，开始阶段往往是个循序渐进的过程，是倒“U”的动态数 据，无法稳定反应实际变量与相关因素的关系，为进一步提升优化方案的可靠度，前面N分钟数据是不加以处理为变量编码的，即所述机器学习模型的学习过程中剔除产线中非稳定数据，不加以处理为变量编码，所述非稳定数据包括产线工作起始阶段的数据、生产中断料时间段内的数据、远超设定的正常范围内的异常数据，数据的筛选大大提高了最后优化方案的精准性。
实施例三，所述机器学习模型根据不同的加工阶段建立不同的模型，形成多阶段串联模型，并针对各个加工阶段分别进行Pearson相关性分析，筛选出相关性显著性强的变量，再对所筛选出的变量进行变量编码。即在各个加工阶段中，相关性较小的变量不在模型的考虑范围，模型只考虑如何优化相关性较大的变量来达到控制烘丝入口含水率，大大提高了预测效率和预测准确度。
在获得优化方案时，先通过产线工作前就能获得的变量去预测各个加工阶段的水分变化量，再通过各阶段水分变化量，从烘丝入口含水率设定值依次反推出润叶加料出口含水率、润叶加料入口含水率、松散出口含水率以及松散阶段水分变化量。
本实施例中，所述变量编码是整型的数据类型存储的，它代表了加水比例、蒸汽比例、回风温度、物料流量、压空到松散的加料比例、松散环境温湿度等各阶段因变量，具体计算公式为：变量编码＝取整函数((变量–变量最低值)/变量步长)。
生成优化方案时，如果存在当前变量对应的变量编码样本则直接使用，如果不存在则通过最小二乘法、线性回归、支持向量机方法去计算，预测，生成优化方案，当再次遇到相同情况时，可以减少训练的过程，直接匹配到历史值作为加水比例的预测，大大提高了效率，且精确度也有所提高。
实施例四，对应的，一种控制烘丝入口含水率的系统，以控制烘丝入口含水率为目标，包括基础工况信息采集模块和机器学习模型，不同的基础工况建立不同的工作模型，所述机器学习模型包括变量编码、编码单元和优化目标。
控制烘丝入口含水率为所述优化目标；所述工作模型通过变量编码与编码单元一一对应，所述变量编码由基础工况信息中的一种变量或多种变量的组合计算得出，实现基础工况信息对编码单元的映射，指示机器学习任务将工作模型派给哪个编码单元以及当前基础工况信息从哪个编码单元获取历史操作信息。
本实施例中，所述基础工况信息采集模块包括原本就安装在制丝生产线上的检测装置。所述基础工况信息包括烘丝入口前的加工阶段影响烘丝入口含水率的主要因素的相应参数，所述加工阶段包括：松散阶段、预混柜阶段、润叶加料阶段和贮叶阶段；其中，所述松散阶段的主要因素包括：加水比例、蒸汽比例、松散回风温度、松散物料流量、压空到松散的加 料比例、松散环境温湿度、松散水分变化量；所述预混柜阶段的主要因素包括：预混柜的环境温湿度、预混柜时长、预混柜水分变化量；所述润叶加料阶段的主要因素包括：加料回风温度、加料物料流量、加料蒸汽补偿实际比例、压空到加料的加料比例、加料的环境温湿度、润叶加料水分变化量；所述贮叶阶段的主要因素包括：贮叶柜环境温湿度、贮叶环境温湿度、贮叶柜时长、贮叶水分变化量。
所述编码单元包括包括基础工况信息及与每组基础工况信息对应的松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值和烘丝入口水分仪零点值。
给出下单时产线上各个测点获得的当前的加工阶段的基础工况信息、松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值、烘丝入口水分仪零点值和烘丝入口含水率设定值，得到变量编码，得到编码单元，计算得到预混柜阶段、润叶加料阶段和贮叶阶段三个阶段的阶段水分变化量，最后得到松散阶段的加水比例，生成优化方案。
综上所述，和现有技术相比，本申请提出的一种控制烘丝入口含水率的方法和系统，通过历史操作，获取相应基础工况信息的加水比例，达到对烘丝入口水分的精确控制的目的，且由于全部优化方案均来自历史操作，使得每个优化方案都可以追溯到源头，使优化方案更具有合理性和安全可信性。本申请通过建立多阶段串联模型，只考虑相关性强的因素，大大提高预测的效率，同时，还通过变量编码，将大量的历史数据样本精编为一个不重复、高价值的编码库，大大降低了内存空间，且提高了模型训练效率
上面结合附图对本发明进行了示例性描述，显然本发明具体实现并不受上述方式的限制，只要采用了本发明的方法构思和技术方案进行的各种非实质性的改进，或未经改进将本发明的构思和技术方案直接应用于其它场合的，均在本发明的保护范围之内。
Claims (10)
- 一种控制烘丝入口含水率的方法，其特征在于，采集历史操作的基础工况信息，不同的基础工况建立不同的工作模型；所述基础工况包括烘丝入口前的加工阶段中影响烘丝入口含水率的主要因素的相应参数，所述加工阶段包括：松散阶段、预混柜阶段、润叶加料阶段和贮叶阶段；其中，所述松散阶段的主要因素包括：加水比例、蒸汽比例、松散回风温度、松散物料流量、压空到松散的加料比例、松散环境温湿度、松散水分变化量；所述预混柜阶段的主要因素包括：预混柜的环境温湿度、预混柜时长、预混柜水分变化量；所述润叶加料阶段的主要因素包括：加料回风温度、加料物料流量、加料蒸汽补偿实际比例、压空到加料的加料比例、加料的环境温湿度、润叶加料水分变化量；所述贮叶阶段的主要因素包括：贮叶柜环境温湿度、贮叶环境温湿度、贮叶柜时长、贮叶水分变化量；建立机器学习模型，所述机器学习模型包括编码单元、优化目标和变量编码；其中，所述编码单元是模型的主要信息单元，是机器学习的知识点，包括基础工况信息及与每组基础工况信息对应的松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值和烘丝入口水分仪零点值；控制烘丝入口含水率为所述优化目标；所述变量编码是模型定位的一种编码，由各阶段各变量计算得出，实现各变量对模型的映射，即根据变量可以快速找到相应的模型，所述变量为基础工况信息中的一种或多种的组合；所述工作模型通过变量编码与编码单元一一对应，通过对应关系指示机器学习任务将工作模型派给哪个编码单元，或者当前工作模型从哪个编码单元获取历史操作信息；给出下单时产线上各个测点获得的当前的加工阶段的基础工况信息、松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值、烘丝入口水分仪零点值和烘丝入口含水率设定值，得到变量编码，匹配到编码单元，计算得到预混柜阶段、润叶加料阶段和贮叶阶段三个阶段的阶段水分变化量，最后得到松散阶段的加水比例，生成优化方案。
- 根据权利要求1所述的一种控制烘丝入口含水率的方法，其特征在于，所述机器学习模型的学习过程中剔除产线中非稳定数据，不加以处理为变量编码，所述非稳定数据包括产线工作起始阶段的数据、生产中断料时间段内的数据、远超设定的正常范围内的异常数据。
- 根据权利要求1所述的一种控制烘丝入口含水率的方法，其特征在于，所述机器学习模型根据不同的加工阶段建立不同的模型，形成多阶段串联模型，并针对各个加工阶段分别进行Pearson相关性分析，筛选出相关性显著性强的变量，再对所筛选出的变量进行变量编码。
- 根据权利要求1所述的一种控制烘丝入口含水率的方法，其特征在于，所述变量编码是整 型的数据类型存储的，具体计算公式为：变量编码＝取整函数((变量–变量最低值)/变量步长)。
- 根据权利要求4所述的一种控制烘丝入口含水率的方法，其特征在于，生成优化方案时，如果存在当前变量对应的变量编码样本则直接使用，如果不存在则通过最小二乘法、线性回归、支持向量机方法去计算、预测，生成优化方案。
- 根据权利要求1所述的一种控制烘丝入口含水率的方法，其特征在于，所述松散水分变化量＝(松散出口含水率-松散出口水分仪零点值)-松散入口默认含水率；所述预混柜水分变化量(润叶入口含水率-润叶入口水分仪零点值)-(松散出口含水率-松散出口水分仪零点值)；所述润叶加料水分变化量＝(润叶出口含水率-润叶出口水分仪零点)-(润叶入口含水率-润叶入口水分仪零点值)；所述贮叶水分变化量＝(烘丝入口含水率-烘丝入口水分仪零点值)-(润叶出口含水率-润叶出口水分仪零点值)。
- 根据权利要求1所述的一种控制烘丝入口含水率的方法，其特征在于，所述机器学习模型建立变量编码溯源机制，每条优化方案都可追溯到变量编码的源头。
- 一种控制烘丝入口含水率的系统，以控制烘丝入口含水率为目标，其特征在于，包括基础工况信息采集模块和机器学习模型；所述基础工况信息采集模块包括安装在制丝生产线上的检测装置；所述基础工况信息包括烘丝入口前的加工阶段影响烘丝入口含水率的主要因素的相应参数，所述加工阶段包括：松散阶段、预混柜阶段、润叶加料阶段和贮叶阶段；其中，所述松散阶段的主要因素包括：加水比例、蒸汽比例、松散回风温度、松散物料流量、压空到松散的加料比例、松散环境温湿度、松散水分变化量；所述预混柜阶段的主要因素包括：预混柜的环境温湿度、预混柜时长、预混柜水分变化量；所述润叶加料阶段的主要因素包括：加料回风温度、加料物料流量、加料蒸汽补偿实际比例、压空到加料的加料比例、加料的环境温湿度、润叶加料水分变化量；所述贮叶阶段的主要因素包括：贮叶柜环境温湿度、贮叶环境温湿度、贮叶柜时长、贮叶水分变化量；一种基础工况信息对应一种工作模型；所述机器学习模型包括变量编码、编码单元和优化目标；所述编码单元包括包括基础工况信息及与每组基础工况信息对应的松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值和烘丝入口水分仪零点值；控制烘丝入口含水率为所述优化目标；所述工作模型通过变量编码与编码单元一一对应，所述变量编码由基础工况信息中的一种变量或多种变量的组合计算得出，实现基础工况信息对编码单元的映射，指示机器学习任务将工作模型派给哪个编码单元以及当前基础工况信息从哪个编码单元获取历史操作信息； 给出下单时产线上各个测点获得的当前的加工阶段的基础工况信息、松散出口水分仪零点值、加料入口水分仪零点值、加料出口水分仪零点值、烘丝入口水分仪零点值和烘丝入口含水率设定值，得到变量编码，得到编码单元，计算得到预混柜阶段、润叶加料阶段和贮叶阶段三个阶段的阶段水分变化量，最后得到松散阶段的加水比例，生成优化方案。
- 根据权利要求8所述的一种控制烘丝入口含水率的系统，其特征在于，所述机器学习模型根据不同的加工阶段建立不同的模型，形成多阶段串联模型，并针对各个加工阶段分别进行Pearson相关性分析，筛选出相关性显著性强的变量，再对所筛选出的变量进行变量编码。
- 根据权利要求8所述的一种控制烘丝入口含水率的系统，其特征在于，所述变量编码是整型的数据类型存储的，具体计算公式为：变量编码＝取整函数((变量–变量最低值)/变量步长)；生成优化方案时，如果存在当前变量对应的变量编码样本则直接使用，如果不存在则通过最小二乘法、线性回归、支持向量机方法去计算、预测，生成优化方案。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811353612.3 | 2018-11-14 | ||
| CN201811353612.3A CN111184246B (zh) | 2018-11-14 | 2018-11-14 | 一种控制烘丝入口含水率的方法和系统 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2020098261A1 true WO2020098261A1 (zh) | 2020-05-22 |
Family
ID=70684195
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2019/089598 WO2020098261A1 (zh) | 2018-11-14 | 2019-05-31 | 一种控制烘丝入口含水率的方法和系统 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111184246B (zh) |
| WO (1) | WO2020098261A1 (zh) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112800671A (zh) * | 2021-01-26 | 2021-05-14 | 联想(北京)有限公司 | 一种数据处理方法、装置及电子设备 |
| CN115336780A (zh) * | 2021-08-26 | 2022-11-15 | 张家口卷烟厂有限责任公司 | 基于神经网络模型和双重参数修正的松散回潮加水控制系统 |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112021626B (zh) * | 2020-07-10 | 2021-08-17 | 张家口卷烟厂有限责任公司 | 烟用制丝环节智能化控制系统及方法 |
| CN112132316A (zh) * | 2020-08-19 | 2020-12-25 | 张家口卷烟厂有限责任公司 | 制丝环节在线设备异常监控系统及方法 |
| CN111938187A (zh) * | 2020-08-25 | 2020-11-17 | 红塔烟草(集团)有限责任公司 | 一种润叶加料机新风温度智能调节装置及调节方法 |
| CN112273695A (zh) * | 2020-10-30 | 2021-01-29 | 红云红河烟草(集团)有限责任公司 | 松散回潮出口含水率预测方法、装置以及设备 |
| CN112273696B (zh) * | 2020-10-30 | 2022-04-01 | 红云红河烟草(集团)有限责任公司 | 切丝后水分控制方法、装置以及设备 |
| CN113515045B (zh) * | 2021-07-12 | 2024-02-20 | 浙江中烟工业有限责任公司 | 一种基于历史数据预判的松散润叶出口水分控制方法 |
| CN115707397A (zh) * | 2021-08-19 | 2023-02-21 | 秦皇岛烟草机械有限责任公司 | 一种烘丝前水分控制方法 |
| CN115251445B (zh) * | 2022-08-15 | 2023-05-23 | 北京航天拓扑高科技有限责任公司 | 一种松散回潮机出口烟叶含水率的控制方法 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2264354A (en) * | 1992-02-18 | 1993-08-25 | Koerber Ag | Method of and apparatus for operating a dryer for fibrous material |
| US6227205B1 (en) * | 1997-12-17 | 2001-05-08 | Brown & Williamson Tobacco Corporation | Method for treatment of tobacco fine cut |
| CN102090708A (zh) * | 2010-09-10 | 2011-06-15 | 龙岩烟草工业有限责任公司 | 一种提高烟丝含水率稳定性的控制方法 |
| CN103750528A (zh) * | 2011-12-31 | 2014-04-30 | 贵州中烟工业有限责任公司 | 一种设定烘丝入口水分值的方法 |
| CN105341985A (zh) * | 2015-12-10 | 2016-02-24 | 龙岩烟草工业有限责任公司 | 烘丝机入口叶丝含水率控制方法和系统 |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101488024A (zh) * | 2009-01-23 | 2009-07-22 | 秦皇岛烟草机械有限责任公司 | 烟草加工过程参数的在线质量评估与实时智能控制方法 |
| CN102708296A (zh) * | 2012-05-15 | 2012-10-03 | 红云红河烟草(集团)有限责任公司 | 一种基于灰色多因素预测模型的能源供需预测方法 |
| EP2821865A1 (en) * | 2013-07-01 | 2015-01-07 | International Tobacco Machinery Poland Sp. z o.o. | Method and system for establishing a data set for identification of parts, modules and/or activities for conversion of flexible production system |
| CN105243435B (zh) * | 2015-09-15 | 2018-10-26 | 中国科学院南京土壤研究所 | 一种基于深度学习元胞自动机模型的土壤含水量预测方法 |
| CN108652066B (zh) * | 2018-05-31 | 2021-07-13 | 福建中烟工业有限责任公司 | 预测松散回潮工序加水量的方法及装置 |
| CN108760668B (zh) * | 2018-06-01 | 2019-10-29 | 南京林业大学 | 基于加权自动编码器的马尾松苗木根部水分快速测量方法 |
-
2018
- 2018-11-14 CN CN201811353612.3A patent/CN111184246B/zh active Active
-
2019
- 2019-05-31 WO PCT/CN2019/089598 patent/WO2020098261A1/zh active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2264354A (en) * | 1992-02-18 | 1993-08-25 | Koerber Ag | Method of and apparatus for operating a dryer for fibrous material |
| US6227205B1 (en) * | 1997-12-17 | 2001-05-08 | Brown & Williamson Tobacco Corporation | Method for treatment of tobacco fine cut |
| CN102090708A (zh) * | 2010-09-10 | 2011-06-15 | 龙岩烟草工业有限责任公司 | 一种提高烟丝含水率稳定性的控制方法 |
| CN103750528A (zh) * | 2011-12-31 | 2014-04-30 | 贵州中烟工业有限责任公司 | 一种设定烘丝入口水分值的方法 |
| CN105341985A (zh) * | 2015-12-10 | 2016-02-24 | 龙岩烟草工业有限责任公司 | 烘丝机入口叶丝含水率控制方法和系统 |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112800671A (zh) * | 2021-01-26 | 2021-05-14 | 联想(北京)有限公司 | 一种数据处理方法、装置及电子设备 |
| CN112800671B (zh) * | 2021-01-26 | 2024-05-31 | 联想(北京)有限公司 | 一种数据处理方法、装置及电子设备 |
| CN115336780A (zh) * | 2021-08-26 | 2022-11-15 | 张家口卷烟厂有限责任公司 | 基于神经网络模型和双重参数修正的松散回潮加水控制系统 |
| CN115336780B (zh) * | 2021-08-26 | 2023-09-26 | 张家口卷烟厂有限责任公司 | 基于神经网络模型和双重参数修正的松散回潮加水控制系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111184246A (zh) | 2020-05-22 |
| CN111184246B (zh) | 2021-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2020098261A1 (zh) | 一种控制烘丝入口含水率的方法和系统 | |
| CN102393882B (zh) | 在线监控和诊断室内空气质量传感器的方法 | |
| CN108652066A (zh) | 松散回潮工序的加水方法及预测该工序加水量的装置 | |
| CN107677709A (zh) | 具有SnO2气敏材料制备步骤的可控湿度的半导体气敏元件的测试方法 | |
| CN111734670B (zh) | 一种磁悬浮鼓风机的测试系统及测试方法 | |
| CN103557991B (zh) | 一种用于压力测量的真空计校准方法及其装置 | |
| CN113128764A (zh) | 烘丝机出口含水率预测模型的生成及基于其的调控方法 | |
| CN104886751A (zh) | 一种基于切叶丝含水率的烘丝机筒壁温度预测模型 | |
| CN102095831A (zh) | 水分仪管理方法和系统 | |
| CN105795497A (zh) | 一种提高烘丝机故障分析的数据采集方法 | |
| CN105183027A (zh) | 一种用于云室的气体湿度自动控制系统及方法 | |
| JP2009020824A (ja) | 空調機の省エネルギー性分析装置、空調システム、空調機の省エネルギー性分析方法および分析プログラム | |
| CN102419593B (zh) | 基于数据挖掘的传感器故障诊断方法 | |
| Zhang et al. | Parameters online detection and model predictive control during the grain drying process | |
| CN112683564B (zh) | 一种不同风速下计算空调器额定制冷量偏差的方法 | |
| CN102128860A (zh) | 湿度传感器动态参数测量系统 | |
| CN110431496A (zh) | 评估装置、评估系统和评估方法 | |
| CN105136339A (zh) | 一种提高入风口温度检测精度的方法 | |
| CN107270955B (zh) | 使用燃烧率信号的燃料分表计量 | |
| CN105136979A (zh) | 湿度传感器时间常数测试装置 | |
| Iskandarovich et al. | Metrological support of moisture measurements of agricultural materials | |
| US11934157B2 (en) | Method for analyzing energy used for producing a unit of mass or volume of compressed gas (specific energy consumption) | |
| CN112283891B (zh) | 用于空调器的调试方法 | |
| CN102393469B (zh) | 一种全自动煤、焦炭对二氧化碳化学反应的测定方法 | |
| CN202362309U (zh) | 一种全自动煤、焦炭对二氧化碳化学反应性测定仪 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19884696 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 19884696 Country of ref document: EP Kind code of ref document: A1 |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 29/11/2021) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 19884696 Country of ref document: EP Kind code of ref document: A1 |