技术领域Technical Field
本发明涉及煮浆控制技术领域，尤其是涉及一种基于数字孪生的全自动煮浆系统。The present invention relates to the technical field of pulp cooking control, and in particular to a fully automatic pulp cooking system based on digital twins.
背景技术Background Art
我国豆制品有2000多年的历史，豆制品营养丰富，蛋白质不仅含量高，还含有丰富的维生素、矿物质、必需脂肪酸、膳食纤维等。在制作豆浆时，首先要对其进行清洗，然后再进行磨浆，将磨好的生浆再放到煮浆罐里进行煮浆。常见的煮浆方式有四种，分别是慢速煮浆保温、慢速传导煮浆、快速煮浆保温和慢速蒸汽煮浆，其中慢速煮浆保温，用夹层蒸气锅煮浆，煮沸豆糊后，保持10min；慢速传导煮浆：采用夹层蒸汽锅，在20min内煮沸豆糊；快速煮浆保温：采用蒸汽喷射煮浆，40s煮沸豆糊后维持80s；慢速蒸汽煮浆：采用蒸汽喷射煮浆，10至12min煮沸豆糊，经实验表明，相比于直接加热的煮浆方式，间接加热所制得豆浆色泽不佳，出现蒸煮味严重，蛋白质提取率偏低；而相比于没有保温效果的煮浆方式（慢速传导煮浆和慢速蒸汽煮浆），具有保温过程的煮浆手段（慢速煮浆保温和快速煮浆保温）所制得的豆浆胰蛋白酶抑制剂残留率较低，分别为13.89%和6.70%，均低于15%，而采用蒸汽加热保温的方式效果更好，其残留率比直接加热低60.8%，所以快速蒸汽保温的蒸煮效果最好。my country's soy products have a history of more than 2,000 years. Soy products are rich in nutrition. Not only are they high in protein, but they are also rich in vitamins, minerals, essential fatty acids, dietary fiber, etc. When making soy milk, it must first be cleaned and then ground. The ground raw pulp is then placed in a cooking tank for cooking. There are four common ways of cooking, namely slow cooking and insulation, slow conduction cooking, fast cooking and insulation, and slow steam cooking. Among them, slow cooking and insulation uses a sandwich steam pot to cook the pulp. After boiling the bean paste, it is kept for 10 minutes; slow conduction cooking: using a sandwich steam pot to boil the bean paste within 20 minutes; fast cooking and insulation: using steam jet cooking, boiling the bean paste for 40 seconds and then maintaining it for 80 seconds; slow steam cooking: using steam jet cooking, boiling the bean paste for 10 to 12 minutes. Experiments have shown that compared with direct heating cooking methods, The soy milk produced by indirect heating has poor color, a strong cooking smell, and a low protein extraction rate. Compared with the cooking methods without heat preservation effect (slow conduction cooking and slow steam cooking), the cooking methods with heat preservation process (slow cooking and heat preservation and fast cooking and heat preservation) have lower trypsin inhibitor residual rates in soy milk, which are 13.89% and 6.70% respectively, both lower than 15%. The steam heating and heat preservation method has better effect, and its residual rate is 60.8% lower than that of direct heating, so the fast steam heat preservation cooking effect is the best.
在传统生产过程中，从制浆机下来后液体在缓冲槽中液位高低均由人来控制，前后管理好多设备，人很容易疲劳，出错。在煮浆过程中人员必须严格遵照工艺和设备操作的要求，确保品质，对人员的培训非常重要，不同的人员操作容易造成产品质量不稳定，容易出现操作失误造成产品质量不合格。在煮浆时温度不能实时检测，靠人目测和经验来控制蒸气阀门，出浆时会手动操作阀门，并且时时观察下游储浆容器是否会满或溢出，很容易造成溢出现象和烫伤，污染食品工作环境，造成不必要的损失。In the traditional production process, the liquid level in the buffer tank after the pulping machine is controlled by humans. There are many devices to manage before and after, and people are easily tired and make mistakes. During the pulping process, personnel must strictly follow the requirements of process and equipment operation to ensure quality. Personnel training is very important. Different personnel operations can easily lead to unstable product quality and unqualified product quality due to operational errors. The temperature cannot be detected in real time during pulping. People rely on visual inspection and experience to control the steam valve. The valve is manually operated when the pulp is discharged, and the downstream pulp storage container is always observed to see if it is full or overflowing. It is easy to cause overflow and burns, pollute the food working environment, and cause unnecessary losses.
数字孪生技术是一个实体产品的数字化表达，以便于我们能够在数字化产品上看到实体产品可能发生的情况。数字化产品在数字化设计与数字化生产的过程中，仿真分析模型的参数，可以传递到数字化产品定义的全三维几何模型，再传递到实体中加工成实际产品，实际产品又通过在线的数字化检测、测量系统反映到产品定义数字化模型中，进而又反馈到仿真分析模型中。Digital twin technology is a digital expression of a physical product, so that we can see what may happen to the physical product on the digital product. In the process of digital design and digital production of digital products, the parameters of the simulation analysis model can be transferred to the full three-dimensional geometric model defined by the digital product, and then transferred to the entity to be processed into the actual product. The actual product is reflected in the digital model of the product definition through the online digital detection and measurement system, and then fed back to the simulation analysis model.
传统煮浆设备在自动化程度、精准控制和智能化方面存在一定局限性，难以满足日益提高的生产需求和质量要求。将数字孪生技术与食品加工设备相结合的经验相对来说不是特别丰富，仍处于逐渐发展和积累阶段，借鉴较优的蒸煮方式实验研究成果，保证豆浆的营养成分和口感，并基于数字孪生技术，减少人的不可控因素，设计一款操作简单，智能化的自动煮浆系统十分必要。Traditional pulping equipment has certain limitations in terms of automation, precision control and intelligence, and it is difficult to meet the increasing production needs and quality requirements. The experience of combining digital twin technology with food processing equipment is relatively not rich, and is still in the stage of gradual development and accumulation. It is necessary to design a simple-to-operate, intelligent automatic pulping system based on digital twin technology to ensure the nutritional content and taste of soy milk and reduce human uncontrollable factors by drawing on the experimental research results of better steaming methods.
发明内容Summary of the invention
本发明提供一种基于数字孪生的全自动煮浆系统，鉴于数字化技术的迅猛进步以及智能制造理念的逐步兴起，数字孪生作为把物理世界与数字世界加以融合的新兴技术，和食品生产机械技术相结合，达成对食品加工流程的智能把控，用以解决上述背景技术里所提及的在传统煮浆过程中，因为煮浆设备对工艺要求的严格性，设备的复杂性以及操作人员技术水平的局限，存有生产效率不高、产品质量难以获得保障之类的问题。The present invention provides a fully automatic pulp cooking system based on digital twins. In view of the rapid progress of digital technology and the gradual rise of the concept of intelligent manufacturing, digital twins, as an emerging technology that integrates the physical world and the digital world, are combined with food production machinery technology to achieve intelligent control of the food processing flow, so as to solve the problems mentioned in the above background technology that in the traditional pulp cooking process, due to the strict process requirements of the pulp cooking equipment, the complexity of the equipment and the limitations of the technical level of the operators, there are problems such as low production efficiency and difficulty in ensuring product quality.
第一方面，本发明提供的一种基于数字孪生的全自动煮浆系统，采用如下的技术方案：包括生豆处理模块、煮浆设备模块、控制模块、传感器模块、数字孪生模块和用户模块，所述数字孪生模块包括数字孪生生成器和数字孪生引擎，所述数字孪生生成器生成数字孪生模型，所述数字孪生引擎搭载深度学习算法和数字孪生模型对系统运行数据进行处理并生成运算结果，所述控制模块根据运算结果生成控制指令，并输出至生豆处理模块和煮浆设备模块。In the first aspect, the present invention provides a fully automatic pulp cooking system based on digital twins, which adopts the following technical scheme: it includes a raw bean processing module, a pulp cooking equipment module, a control module, a sensor module, a digital twin module and a user module. The digital twin module includes a digital twin generator and a digital twin engine. The digital twin generator generates a digital twin model. The digital twin engine is equipped with a deep learning algorithm and a digital twin model to process the system operation data and generate calculation results. The control module generates control instructions according to the calculation results and outputs them to the raw bean processing module and the pulp cooking equipment module.
进一步地，所述生豆处理模块包括大豆容器和磨浆机，所述大豆容器和磨浆机内均设有视觉传感器，所述视觉传感器用于检测大豆容器和磨浆机内的物料情况。Furthermore, the green bean processing module includes a soybean container and a pulping machine, and the soybean container and the pulping machine are both provided with visual sensors, and the visual sensors are used to detect the material conditions in the soybean container and the pulping machine.
进一步地，所述煮浆设备模块包括六个连续的煮浆罐、进浆管道单元、出浆管道单元、CIP清洗管道单元和蒸汽管道单元，所述煮浆罐的上部设有进浆口，所述煮浆罐的下部设有出浆口，所述蒸汽管道穿过煮浆罐的上部进入罐体内部，所述蒸汽管道在煮浆罐内部的管道上设有蒸汽孔，所述煮浆罐的上部设有减压阀，所述煮浆罐的顶部设置CIP清洗管道进口，所述煮浆罐的顶部设有搅拌桨。Furthermore, the pulp cooking equipment module includes six continuous pulp cooking tanks, a pulp inlet pipe unit, a pulp outlet pipe unit, a CIP cleaning pipe unit and a steam pipe unit. The upper part of the pulp cooking tank is provided with a pulp inlet, the lower part of the pulp cooking tank is provided with a pulp outlet, the steam pipe passes through the upper part of the pulp cooking tank and enters the interior of the tank body. The steam pipe is provided with steam holes on the pipe inside the pulp cooking tank. The upper part of the pulp cooking tank is provided with a pressure reducing valve, the top of the pulp cooking tank is provided with a CIP cleaning pipe inlet, and the top of the pulp cooking tank is provided with a stirring paddle.
进一步地，所述传感器模块包括若干个传感器，其中，所述煮浆罐的上部设置第一温度传感器，所述煮浆罐的下部设置第二温度传感器，所述煮浆罐的上部设置压力传感器，所述煮浆罐的侧壁设置液位传感器，所述搅拌桨设置转速控制器，所述蒸汽管道单元上设置蒸汽控制阀，所述进浆管道单元上设置进浆控制阀，所述出浆管道单元上设置出浆控制阀，所述CIP清洗管道单元上设置清洗控制阀。Furthermore, the sensor module includes several sensors, wherein a first temperature sensor is arranged on the upper part of the pulp cooking tank, a second temperature sensor is arranged on the lower part of the pulp cooking tank, a pressure sensor is arranged on the upper part of the pulp cooking tank, a liquid level sensor is arranged on the side wall of the pulp cooking tank, a speed controller is arranged on the agitator, a steam control valve is arranged on the steam pipe unit, a pulp inlet control valve is arranged on the pulp inlet pipe unit, a pulp outlet control valve is arranged on the pulp outlet pipe unit, and a cleaning control valve is arranged on the CIP cleaning pipe unit.
进一步地，所述数字孪生生成器包括数据获取单元、数据处理单元、模型构建单元和验证校准单元，所述数据获取单元用于获取系统运行数据，所述数据处理单元用于对设备运行数据进行清洗与整合，所述模型构建单元用于构建三维模型并将处理后的数据与三维模型进行融合，所述验证校准单元用于将三维模型与系统运行数据进行比对，根据比对结果调整三维模型的参数和算法。Furthermore, the digital twin generator includes a data acquisition unit, a data processing unit, a model building unit and a verification and calibration unit. The data acquisition unit is used to acquire system operation data, the data processing unit is used to clean and integrate the equipment operation data, the model building unit is used to build a three-dimensional model and fuse the processed data with the three-dimensional model, and the verification and calibration unit is used to compare the three-dimensional model with the system operation data, and adjust the parameters and algorithms of the three-dimensional model according to the comparison results.
进一步地，所述构建三维模型并将处理后的数据与三维模型进行融合包括以下步骤：Furthermore, the constructing of the three-dimensional model and fusing the processed data with the three-dimensional model comprises the following steps:
利用BIM软件构建三维模型，并对三维模型进行渲染；Use BIM software to build 3D models and render them;
建立运行数据与三维模型属性之间的映射关系；Establishing the mapping relationship between operation data and 3D model attributes;
利用MQTT通信协议对运行数据进行传输；Use MQTT communication protocol to transmit operation data;
根据运行数据变化修改三维模型属性参数并进行可视化动态调整。Modify the 3D model attribute parameters according to the changes in operating data and make visual dynamic adjustments.
进一步地，所述数字孪生引擎包括优化生产工艺单元、优化设备清洗单元和优化设备保养维修单元，所述优化生产工艺单元用于优化学习多种生产配方，所述优化设备清洗单元用于优化清洗设备的时间和周期，所述优化设备保养维修单元用于优化设备保养、维修的时间和频率。Furthermore, the digital twin engine includes a production process optimization unit, an equipment cleaning optimization unit and an equipment maintenance and repair optimization unit. The production process optimization unit is used to optimize and learn multiple production recipes, the equipment cleaning optimization unit is used to optimize the time and cycle of cleaning equipment, and the equipment maintenance and repair optimization unit is used to optimize the time and frequency of equipment maintenance and repair.
进一步地，所述数字孪生引擎搭载深度学习算法和数字孪生模型对系统运行数据进行处理并生成运算结果包括以下步骤：Furthermore, the digital twin engine carries a deep learning algorithm and a digital twin model to process the system operation data and generate a calculation result, including the following steps:
构建深度学习算法模型，通过自动超参数调整和正则化优化模型；Build deep learning algorithm models and optimize them through automatic hyperparameter tuning and regularization;
利用数据集进行模型训练，采用分布式训练和交叉验证评估模型性能；Use the dataset to train the model, and use distributed training and cross-validation to evaluate the model performance;
通过融合深度学习算法模型与数字孪生模型，确定中间数据交互格式；By integrating the deep learning algorithm model with the digital twin model, the intermediate data interaction format is determined;
使用实时流数据处理框架接入和处理实时数据，实现低延迟推理。Use the real-time stream data processing framework to access and process real-time data and achieve low-latency reasoning.
进一步地，所述控制模块包括工艺参数调节单元、设备清洗单元和设备保养维修单元，所述工艺参数调节单元用于生成调整设备的温度、压力、搅拌速度、煮浆时间、进浆和出浆的控制指令，设备清洗单元用于生成开启CIP清洗管道单元上的清洗控制阀的控制指令，设备保养维修单元用于生成提醒设备保养和维修的控制指令。Furthermore, the control module includes a process parameter adjustment unit, an equipment cleaning unit and an equipment maintenance and repair unit. The process parameter adjustment unit is used to generate control instructions for adjusting the temperature, pressure, stirring speed, slurry cooking time, slurry inlet and outlet of the equipment. The equipment cleaning unit is used to generate control instructions for opening the cleaning control valve on the CIP cleaning pipeline unit. The equipment maintenance and repair unit is used to generate control instructions for reminding equipment maintenance and repair.
进一步地，所述用户模块包括终端控制机，所述终端控制机设置显示器，所述终端控制机通过数字孪生模型可视化展示系统运行状态。Furthermore, the user module includes a terminal control machine, the terminal control machine is provided with a display, and the terminal control machine visualizes the system operation status through a digital twin model.
综上所述，本发明具有如下的有益技术效果：In summary, the present invention has the following beneficial technical effects:
1、本发明提出一种基于数字孪生的全自动煮浆系统，依据大量实验室数据，对深度学习算法模型进行训练，采用连续煮浆方式和蒸汽加热保温的方式进行煮浆，充分保证了产品的口感和营养，减少了豆浆胰蛋白酶抑制剂残留率，减少豆腥味，不需投放消泡剂，节约食品添加剂成本，整个煮浆设备模块设计合理，六个不同煮浆罐的煮浆温度不同，本发明实现了工艺要求较高的需求，操作人员通过数字孪生模型可以直观看到设备状态，并进行参数设置，人员操作控制变得简单；1. The present invention proposes a fully automatic pulp cooking system based on digital twins. According to a large amount of laboratory data, a deep learning algorithm model is trained, and continuous pulp cooking and steam heating and heat preservation are used for pulp cooking, which fully guarantees the taste and nutrition of the product, reduces the residual rate of soymilk trypsin inhibitors, reduces the beany smell, does not require the addition of defoaming agents, and saves the cost of food additives. The entire pulp cooking equipment module is reasonably designed, and the pulp cooking temperatures of six different pulp cooking tanks are different. The present invention meets the requirements of higher process requirements. The operator can intuitively see the equipment status and set parameters through the digital twin model, and the personnel operation and control become simple;
2、本发明中的数字孪生模型可以存储多种优化的工艺配方，因为不同品质和种类的豆子以及其之间的配比不同，比如黄豆和黑芝麻配比制作黑芝麻豆浆等，其生产工艺不同，通过深度学习算法模型训练，优化多种产品的生产的工艺，控制不同罐内加热温度、不同罐内压力和不同罐内加热时间等，达到理想和稳定的产品质量，在煮浆还未完全结束时就可以开启下一轮的生豆研磨工序，减少了生产中停机时间，数字孪生模块可给出优化的生产和清洁保养计划；2. The digital twin model in the present invention can store a variety of optimized process formulas, because beans of different qualities and types and their ratios are different, such as the ratio of soybeans and black sesame to make black sesame soy milk, etc., have different production processes. Through deep learning algorithm model training, the production processes of various products are optimized, and the heating temperature, pressure and heating time in different tanks are controlled to achieve ideal and stable product quality. The next round of raw bean grinding process can be started before the pulping is completely completed, reducing downtime in production. The digital twin module can provide optimized production and cleaning and maintenance plans;
3、本发明提出一种基于数字孪生的全自动煮浆系统，通过数字孪生可以检测故障发生，提醒维修人员采取相应的维护措施，降低设备故障和维修成本，提醒维修人员定期保养和更换部件；3. The present invention proposes a fully automatic pulp cooking system based on digital twins, which can detect the occurrence of faults through digital twins, remind maintenance personnel to take corresponding maintenance measures, reduce equipment failures and maintenance costs, and remind maintenance personnel to regularly maintain and replace parts;
4、本发明提出一种基于数字孪生的全自动煮浆系统，数字孪生模块具备自动模型校准和迭代的能力，生产过程中模型不断地学习和经验积累，可以进行工艺参数的优化和控制策略的调整，以实现更高的生产效率和产品质量，同时可以为操作人员提供可视化操作界面，实时查看煮浆系统的状态和性能指标，并生成运行记录、故障报告、清洗记录以及保养信息的报告。4. The present invention proposes a fully automatic pulp cooking system based on digital twins. The digital twin module has the ability of automatic model calibration and iteration. During the production process, the model continuously learns and accumulates experience, and can optimize process parameters and adjust control strategies to achieve higher production efficiency and product quality. At the same time, it can provide operators with a visual operation interface, view the status and performance indicators of the pulp cooking system in real time, and generate operation records, fault reports, cleaning records, and maintenance information reports.
附图说明BRIEF DESCRIPTION OF THE DRAWINGS
图1是本发明实施例1的煮浆流程示意图；FIG1 is a schematic diagram of a pulp cooking process according to Embodiment 1 of the present invention;
图2是本发明实施例1的煮浆罐的结构示意图；FIG2 is a schematic structural diagram of a pulp cooking tank according to Embodiment 1 of the present invention;
图3是本发明实施例1的基于数字孪生的全自动煮浆系统示意图。Figure 3 is a schematic diagram of a fully automatic pulp cooking system based on digital twins in Example 1 of the present invention.
具体实施方式DETAILED DESCRIPTION
以下结合附图对本发明作进一步详细说明。The present invention is further described in detail below in conjunction with the accompanying drawings.
实施例1Example 1
参照图1和图3，本实施例的一种基于数字孪生的全自动煮浆系统，包括生豆处理模块、煮浆设备模块、控制模块、传感器模块、数字孪生模块和用户模块，所述数字孪生模块包括数字孪生生成器和数字孪生引擎，所述数字孪生生成器生成数字孪生模型，所述数字孪生引擎搭载深度学习算法和数字孪生模型对系统运行数据进行处理并生成运算结果，所述控制模块根据运算结果生成控制指令，并输出至生豆处理模块和煮浆设备模块。Referring to Figures 1 and 3, a fully automatic pulping system based on digital twins in this embodiment includes a raw bean processing module, a pulping equipment module, a control module, a sensor module, a digital twin module and a user module. The digital twin module includes a digital twin generator and a digital twin engine. The digital twin generator generates a digital twin model. The digital twin engine is equipped with a deep learning algorithm and a digital twin model to process the system operation data and generate calculation results. The control module generates control instructions according to the calculation results and outputs them to the raw bean processing module and the pulping equipment module.
参照图2，所述生豆处理模块包括大豆容器和磨浆机，所述大豆容器和磨浆机内均设有视觉传感器，所述视觉传感器用于检测大豆容器和磨浆机内的物料情况。2 , the green bean processing module includes a soybean container and a pulping machine. Both the soybean container and the pulping machine are provided with visual sensors for detecting the material conditions in the soybean container and the pulping machine.
参照图2，所述煮浆设备模块包括六个连续的煮浆罐、进浆管道单元、出浆管道单元、CIP清洗管道单元和蒸汽管道单元，所述煮浆罐的上部设有进浆口，所述煮浆罐的下部设有出浆口，所述蒸汽管道穿过煮浆罐的上部进入罐体内部，所述蒸汽管道在煮浆罐内部的管道上设有蒸汽孔，所述煮浆罐的上部设有减压阀，所述煮浆罐的顶部设置CIP清洗管道进口，所述煮浆罐的顶部设有搅拌桨。Referring to Figure 2, the pulp cooking equipment module includes six continuous pulp cooking tanks, a pulp inlet pipe unit, a pulp outlet pipe unit, a CIP cleaning pipe unit and a steam pipe unit. The upper part of the pulp cooking tank is provided with a pulp inlet, the lower part of the pulp cooking tank is provided with a pulp outlet, the steam pipe passes through the upper part of the pulp cooking tank and enters the interior of the tank body. The steam pipe is provided with steam holes on the pipe inside the pulp cooking tank. The upper part of the pulp cooking tank is provided with a pressure reducing valve, the top of the pulp cooking tank is provided with a CIP cleaning pipe inlet, and the top of the pulp cooking tank is provided with a stirring paddle.
所述煮浆罐的上部设置第一温度传感器，所述煮浆罐的下部设置第二温度传感器，所述煮浆罐的上部设置压力传感器，所述煮浆罐的侧壁设置液位传感器，所述搅拌桨设置转速控制器，所述蒸汽管道单元上设置蒸汽控制阀，所述进浆管道单元上设置进浆控制阀，所述出浆管道单元上设置出浆控制阀，所述CIP清洗管道单元上设置清洗控制阀。A first temperature sensor is arranged on the upper part of the pulp cooking tank, a second temperature sensor is arranged on the lower part of the pulp cooking tank, a pressure sensor is arranged on the upper part of the pulp cooking tank, a liquid level sensor is arranged on the side wall of the pulp cooking tank, a speed controller is arranged on the stirring paddle, a steam control valve is arranged on the steam pipe unit, a pulp inlet control valve is arranged on the pulp inlet pipe unit, a pulp outlet control valve is arranged on the pulp outlet pipe unit, and a cleaning control valve is arranged on the CIP cleaning pipe unit.
所述数字孪生生成器包括数据获取单元、数据处理单元、模型构建单元和验证校准单元，所述数据获取单元用于获取系统运行数据，所述数据处理单元用于对设备运行数据进行清洗与整合，所述模型构建单元用于构建三维模型并将处理后的数据与三维模型进行融合，所述验证校准单元用于将三维模型与系统运行数据进行比对，根据比对结果调整三维模型的参数和算法。The digital twin generator includes a data acquisition unit, a data processing unit, a model building unit and a verification and calibration unit. The data acquisition unit is used to acquire system operation data, the data processing unit is used to clean and integrate equipment operation data, the model building unit is used to build a three-dimensional model and fuse the processed data with the three-dimensional model, and the verification and calibration unit is used to compare the three-dimensional model with the system operation data, and adjust the parameters and algorithms of the three-dimensional model according to the comparison results.
所述构建三维模型并将处理后的数据与三维模型进行融合包括以下步骤：The construction of the three-dimensional model and the fusion of the processed data with the three-dimensional model comprises the following steps:
利用BIM软件构建三维模型，并对三维模型进行渲染；Use BIM software to build 3D models and render them;
建立运行数据与三维模型属性之间的映射关系；Establishing the mapping relationship between operation data and 3D model attributes;
利用MQTT通信协议对运行数据进行传输；Use MQTT communication protocol to transmit operation data;
根据运行数据变化修改三维模型属性参数并进行可视化动态调整。Modify the 3D model attribute parameters according to the changes in operating data and make visual dynamic adjustments.
工作时,整个系统运行中，传感器实时采集的设备状态信息、生产过程中的参数，包括煮浆的温度、煮浆罐的压力值、煮浆的时间、进出浆的流量，当数字孪生生成器的数据获取单元获取到这些数据后，会进行整理、筛选和初步分析，以确保数据的准确性和可用性，数字孪生生成器会依据这些数据构建或更新对应的数字孪生模型。数字孪生生成器会利用数据中的设备信息，塑造出与实际物理对象高度相似的数字孪生体，并实时反映煮浆系统的实际状态。数字孪生引擎会利用接收到的数据，并利用深度学习算法进行深入的分析和计算，生成相应的控制指令以及性能评估结果，从而实现对整个煮浆过程的监控、优化和决策支持。When the entire system is running, the sensors collect real-time equipment status information and parameters in the production process, including the temperature of pulp cooking, the pressure value of the pulp cooking tank, the pulp cooking time, and the flow rate of pulp in and out. When the data acquisition unit of the digital twin generator obtains this data, it will sort, screen and conduct preliminary analysis to ensure the accuracy and availability of the data. The digital twin generator will build or update the corresponding digital twin model based on this data. The digital twin generator will use the equipment information in the data to create a digital twin that is highly similar to the actual physical object and reflect the actual status of the pulp cooking system in real time. The digital twin engine will use the received data and use deep learning algorithms for in-depth analysis and calculation to generate corresponding control instructions and performance evaluation results, thereby realizing monitoring, optimization and decision support for the entire pulp cooking process.
工作时，用户模块包括终端控制机，所述终端控制机设置显示器，所述终端控制机通过数字孪生模型可视化展示设备运行状态，终端控制机可以单独操作机器的启停，终端控制机内置控制模块，所述控制模块包括工艺参数调节单元、设备清洗单元和设备保养维修单元，所述工艺参数调节单元用于生成调整设备的温度、压力、搅拌速度、煮浆时间、进浆和出浆的控制指令，设备清洗单元用于生成开启CIP清洗管道单元上的清洗控制阀的控制指令，设备保养维修单元用于生成提醒设备保养和维修的控制指令。When working, the user module includes a terminal control machine, which is provided with a display. The terminal control machine visualizes the operating status of the equipment through a digital twin model. The terminal control machine can independently operate the start and stop of the machine. The terminal control machine has a built-in control module, which includes a process parameter adjustment unit, an equipment cleaning unit and an equipment maintenance and repair unit. The process parameter adjustment unit is used to generate control instructions for adjusting the temperature, pressure, stirring speed, pulp cooking time, pulp inlet and outlet of the equipment. The equipment cleaning unit is used to generate control instructions for opening the cleaning control valve on the CIP cleaning pipeline unit. The equipment maintenance and repair unit is used to generate control instructions for reminding equipment maintenance and repair.
通过控制模块，开启煮浆模式，控制模块根据数字孪生引擎的运算结果，自动控制系统运行，将大豆容器15中的豆类提升倒置在磨浆机16中进行磨浆，六个煮浆罐1温度分别是55℃、65℃、75℃、85℃、95℃、105℃，第一蒸汽阀被打开，到达预设温度后打开第一进浆阀，搅拌桨10旋转，通过管道抽吸至第一煮浆罐1中，蒸煮时间分别为80s、60s、40s、30s、20s、10s，达到预设时间被抽吸至下一个煮浆罐1继续煮浆，对浆液进行搅拌加热并达到预设温度，此时只打开蒸汽管道8上的阀门和排气口的阀门，控制煮浆罐1内气压，罐体上的其他阀门全部关闭，通过蒸汽通孔上排出的高温蒸汽对罐内的生浆进行蒸煮，蒸煮后进行闷浆，关闭所有的阀门，通过罐内的残留的热量继续对豆浆进行闷煮持续60s；此时，如果罐内的压力传感器检测到气压超过正常压力，减压阀9打开释放压力，最后进行放浆，打开出浆控制阀，煮熟的浆液通过出浆管路流出至分浆设备17进行下步工艺，本实施例中的系统和工艺可以不投放消泡剂，实现煮浆过程减少泡沫产生，通过优化煮浆工艺，调整煮浆的温度、时间和搅拌速度参数，使用物理消泡方法，通过机械搅拌来破坏泡沫的稳定性，使其破裂。操作人员通过数字孪生模型可以直观看到设备状态和性能指标，并进行参数设置，人员操作控制变得简单。The pulping mode is turned on through the control module. The control module automatically controls the operation of the system according to the calculation results of the digital twin engine. The beans in the soybean container 15 are lifted and inverted in the pulper 16 for pulping. The temperatures of the six pulping tanks 1 are 55°C, 65°C, 75°C, 85°C, 95°C, and 105°C, respectively. The first steam valve is opened. After reaching the preset temperature, the first pulp inlet valve is opened, and the stirring paddle 10 rotates and is sucked into the first pulping tank 1 through the pipeline. The steaming times are 80s, 60s, 40s, 30s, 20s, and 10s, respectively. When the preset time is reached, it is sucked to the next pulping tank 1 to continue pulping. The slurry is stirred and heated to reach the preset temperature. At this time, only the valve on the steam pipe 8 and the valve at the exhaust port are opened to control the steam pressure in the pulping tank 1. The pressure is set to zero, and all other valves on the tank are closed. The raw pulp in the tank is cooked by the high-temperature steam discharged from the steam through hole. After cooking, the raw pulp is stewed. All valves are closed, and the residual heat in the tank continues to stew the soybean milk for 60 seconds. At this time, if the pressure sensor in the tank detects that the air pressure exceeds the normal pressure, the pressure reducing valve 9 is opened to release the pressure, and finally the pulp is released. The pulp discharge control valve is opened, and the cooked pulp flows out through the pulp discharge pipeline to the pulp separation equipment 17 for the next process. The system and process in this embodiment do not need to add defoaming agents to reduce the generation of foam during the pulping process. By optimizing the pulping process, adjusting the pulping temperature, time and stirring speed parameters, using physical defoaming methods, and destroying the stability of the foam through mechanical stirring to make it break. The operator can intuitively see the equipment status and performance indicators through the digital twin model, and set parameters, which makes personnel operation and control simple.
实施例2Example 2
参照图3，数字孪生引擎还搭载深度学习算法，所述数字孪生引擎包括优化生产工艺单元、优化设备清洗单元和优化设备保养维修单元，所述优化生产工艺单元用于优化学习多种生产配方，所述优化设备清洗单元用于优化清洗设备的时间和周期.Referring to Figure 3, the digital twin engine is also equipped with a deep learning algorithm. The digital twin engine includes a production process optimization unit, an equipment cleaning optimization unit, and an equipment maintenance and repair optimization unit. The production process optimization unit is used to optimize and learn multiple production formulas, and the equipment cleaning optimization unit is used to optimize the time and cycle of cleaning equipment.
所述数字孪生引擎搭载深度学习算法和数字孪生模型对系统运行数据进行处理并生成运算结果包括以下步骤：The digital twin engine is equipped with a deep learning algorithm and a digital twin model to process the system operation data and generate calculation results, including the following steps:
构建深度学习算法模型，通过自动超参数调整和正则化优化模型；Build deep learning algorithm models and optimize them through automatic hyperparameter tuning and regularization;
利用数据集进行模型训练，采用分布式训练和交叉验证评估模型性能；Use the dataset to train the model, and use distributed training and cross-validation to evaluate the model performance;
深度学习模型与数字孪生模型的融合，确定中间数据交互格式；The integration of deep learning models and digital twin models to determine the intermediate data interaction format;
使用实时流数据处理框架接入和处理实时数据，实现低延迟推理。Use the real-time stream data processing framework to access and process real-time data and achieve low-latency reasoning.
工作时,根据系统提示的不同工艺中不同原料的不同配比，提前放置于大豆容器中，磨浆机将黄豆和紫米混合制作出紫米豆浆；将玉米和黄豆搭配，制作玉米豆浆。通过数字孪生引擎，能够优化多种产品的生产工艺，数字孪生引擎可以精确计算控制每个煮浆罐1的温度、时间和搅拌速度，通过控制器控制调节系统运转。During operation, according to the different ratios of different raw materials in different processes prompted by the system, they are placed in the soybean container in advance, and the pulping machine mixes soybeans and purple rice to make purple rice soy milk; corn and soybeans are mixed to make corn soy milk. Through the digital twin engine, the production process of various products can be optimized. The digital twin engine can accurately calculate and control the temperature, time and stirring speed of each pulping tank 1, and control and adjust the operation of the system through the controller.
同时，数字孪生引擎还能够根据生产需求，智能地调整生产计划，优化不同煮浆产品的转换时间，数字孪生模型根据实时数据模拟当前生产状态和进度,利用处理后的运行数据，训练深度学习模型，模型学习不同产品生产之间的转换规律、时间模式以及影响因素,传感器实时监测设备状态、生产进度、原料供应等数据,这些实时数据被传输至数字孪生模型和深度学习算法中,数字孪生模型根据实时数据模拟当前生产状态和进度,深度学习算法基于学习到的转换规律和时间模式，结合数字孪生模型的模拟结果，计算出最佳的产品转换时间,根据计算出的最佳转换时间，智能调整系统中设备的运行先后，当需要从一种豆浆产品转换到另一种时，系统可以自动计算出最佳的转换时间，在煮浆的同时，系统可以自动安排研磨设备进行下一批次的原料研磨，确保生产的连续性，而且，视觉传感器可以检测到不同产品,深度学习算法能够根据不同产品的需求，控制模块智能地调整研磨参数，提高研磨效率和产品质量。At the same time, the digital twin engine can also intelligently adjust the production plan according to production needs and optimize the conversion time of different soy milk products. The digital twin model simulates the current production status and progress based on real-time data, and uses the processed operation data to train the deep learning model. The model learns the conversion rules, time patterns and influencing factors between the production of different products. Sensors monitor equipment status, production progress, raw material supply and other data in real time. These real-time data are transmitted to the digital twin model and deep learning algorithm. The digital twin model simulates the current production status and progress based on real-time data. The deep learning algorithm calculates the optimal product conversion time based on the learned conversion rules and time patterns combined with the simulation results of the digital twin model. According to the calculated optimal conversion time, the operation order of the equipment in the system is intelligently adjusted. When it is necessary to switch from one soy milk product to another, the system can automatically calculate the optimal conversion time. While boiling the soy milk, the system can automatically arrange the grinding equipment to grind the next batch of raw materials to ensure the continuity of production. Moreover, the visual sensor can detect different products. The deep learning algorithm can intelligently adjust the grinding parameters of the control module according to the needs of different products to improve grinding efficiency and product quality.
通过数字孪生引擎搭载深度学习算法，清洗设备时间也得到了优化。数字孪生引擎可以根据生产计划和设备使用情况，合理安排清洗时间，当一款产品煮浆结束前，系统提前进入下一款产品的磨浆工序，同时煮浆设备开启清洗模式，在研磨结束前完成清洗，确保设备始终保持清洁和卫生，同时最大限度地减少清洗对生产的影响,缩短加工到清洗到转换产品等工序的时长。Through the digital twin engine equipped with deep learning algorithms, the time for cleaning equipment has also been optimized. The digital twin engine can reasonably arrange the cleaning time according to the production plan and equipment usage. Before the pulping of a product is completed, the system will enter the pulping process of the next product in advance. At the same time, the pulping equipment will start the cleaning mode and complete the cleaning before the grinding is completed, ensuring that the equipment is always clean and hygienic, while minimizing the impact of cleaning on production and shortening the time from processing to cleaning to product conversion.
实施例3Example 3
参照图3，数字孪生引擎搭载深度学习算法，所述数字孪生引擎还包括：优化设备保养维修单元，用于优化设备保养、维修的时间和频率。3 , the digital twin engine is equipped with a deep learning algorithm, and the digital twin engine also includes: an equipment maintenance and repair optimization unit, which is used to optimize the time and frequency of equipment maintenance and repair.
设备保养包括设备清洁、加润滑剂、加润滑油、设备固定、设备调整、零件更换和设备修理；通过传感器数字孪生模型检测设备的故障，包括摩擦增大、振动增大、不平衡、磨损、润滑性降低、发动机失灵、轴承失灵、轴失灵、零件裂纹、传感器故障和电路短路，一旦发现故障，数字孪生模型会立即发出提醒，并提供故障诊断和维修建议。Equipment maintenance includes equipment cleaning, adding lubricants, adding lubricants, equipment fixing, equipment adjustment, parts replacement and equipment repair; equipment failures are detected through the sensor digital twin model, including increased friction, increased vibration, imbalance, wear, reduced lubricity, engine failure, bearing failure, shaft failure, part cracks, sensor failure and circuit short circuit. Once a fault is detected, the digital twin model will immediately issue a reminder and provide fault diagnosis and maintenance suggestions.
本实施例中，对煮浆系统中设备关键位置布置传感器，当传感器检测到设备的摩擦增大时，通过分析传感器采集的摩擦力数据、温度变化以及润滑剂的消耗速度等参数，并与预设的正常范围进行对比，从而判断摩擦是否异常增大，数字孪生模型会建议对设备进行润滑和清洁；当检测到设备的振动增大时，利用振动传感器采集的振动频谱数据，结合快速傅里叶变换信号处理技术，分析振动频率和振幅的变化，超过预设值时，数字孪生模型会依据设备的结构特点、安装方式以及振动的频谱特征，建议对设备进行精准的平衡调整；同时数字孪生模型预设了零件的寿命和保养周期，通过深度学习算法可以计算出设备保养和维修的时间，从而确保设备的正常运行和延长设备的使用寿命，数字孪生模型通过算法分析，预测出现的故障，并提供故障诊断建议，根据实际维护情况，反馈至数字孪生模型，优化模型的准确性和预测能力，持续的改进和优化。In this embodiment, sensors are arranged at key positions of equipment in the pulp cooking system. When the sensor detects that the friction of the equipment increases, the friction data, temperature changes, lubricant consumption rate and other parameters collected by the sensor are analyzed and compared with the preset normal range to determine whether the friction has increased abnormally. The digital twin model will recommend lubrication and cleaning of the equipment. When it is detected that the vibration of the equipment increases, the vibration spectrum data collected by the vibration sensor is combined with the fast Fourier transform signal processing technology to analyze the changes in vibration frequency and amplitude. When the preset value is exceeded, the digital twin model will recommend precise balancing adjustments to the equipment based on the structural characteristics of the equipment, installation method and spectral characteristics of the vibration. At the same time, the digital twin model presets the life and maintenance cycle of parts. The deep learning algorithm can calculate the time for equipment maintenance and repair, thereby ensuring the normal operation of the equipment and extending the service life of the equipment. The digital twin model predicts faults through algorithm analysis and provides fault diagnosis suggestions. According to the actual maintenance situation, feedback is given to the digital twin model to optimize the accuracy and predictive ability of the model, and continuous improvement and optimization.
实施例4Example 4
本实施例中的数字孪生引擎具备模型校准和迭代的能力，以不断提高模型的准确性和适应性，操作人员根据产品研发经验和实际生产产品质量情况，对深度学习算法模型中的参数进行调整，提高模型下次执行操作的准确性；随着时间推移和新的数据积累，深度学习算法自动进行模型的迭代和更新，以适应产品的演变和变化。The digital twin engine in this embodiment has the ability to calibrate and iterate the model to continuously improve the accuracy and adaptability of the model. The operator adjusts the parameters in the deep learning algorithm model based on product development experience and actual production product quality to improve the accuracy of the model's next execution operation; as time goes by and new data is accumulated, the deep learning algorithm automatically iterates and updates the model to adapt to the evolution and changes of the product.
数字孪生引擎被配置为提供数据可视化和生成报告，所述生成报告包括生成设备的运行记录、故障报告、清洗记录以及保养信息，通过直观的可视化界面，操作人员可以实时查看煮浆系统的状态和性能指标。The digital twin engine is configured to provide data visualization and generate reports, including equipment operation records, fault reports, cleaning records, and maintenance information. Through an intuitive visualization interface, operators can view the status and performance indicators of the pulping system in real time.
以上均为本发明的较佳实施例，并非依此限制本发明的保护范围，故：凡依本发明的结构、形状、原理所做的等效变化，均应涵盖于本发明的保护范围之内。The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410981488.4ACN118805924B (en) | 2024-07-22 | 2024-07-22 | Full-automatic pulp boiling system based on digital twin |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410981488.4ACN118805924B (en) | 2024-07-22 | 2024-07-22 | Full-automatic pulp boiling system based on digital twin |
| Publication Number | Publication Date |
|---|---|
| CN118805924Atrue CN118805924A (en) | 2024-10-22 |
| CN118805924B CN118805924B (en) | 2025-03-07 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410981488.4AActiveCN118805924B (en) | 2024-07-22 | 2024-07-22 | Full-automatic pulp boiling system based on digital twin |
| Country | Link |
|---|---|
| CN (1) | CN118805924B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201207903Y (en)* | 2008-06-06 | 2009-03-18 | 王春成 | Unequal height continuous pulp boiling pot |
| KR20100126889A (en)* | 2009-05-25 | 2010-12-03 | 원두원 식품기계(주) | Tofu manufacturing equipment |
| CN114175044A (en)* | 2019-07-17 | 2022-03-11 | 克朗斯股份公司 | Production planning for predictive maintenance and optimization of devices in the food industry by digital twinning |
| US20220087228A1 (en)* | 2020-09-23 | 2022-03-24 | International Business Machines Corporation | Industrial livestock management leveraging digital twin computing |
| US20220164502A1 (en)* | 2020-11-23 | 2022-05-26 | Jiangsu University | Pump machine unit optimized operation regulation system and method based on digital twin |
| WO2022233596A1 (en)* | 2021-05-03 | 2022-11-10 | Basf Coatings Gmbh | Methods and systems for determining the fill level of a liquid in a container |
| EP4174597A1 (en)* | 2021-10-29 | 2023-05-03 | Tetra Laval Holdings & Finance S.A. | A method for configuring a food processing system to produce a plurality of food products |
| CN116369564A (en)* | 2023-06-06 | 2023-07-04 | 山西瑞飞机械制造有限公司 | Intelligent defoaming and deodorizing cooking pulp tank and its pulping process |
| US20230304674A1 (en)* | 2020-07-01 | 2023-09-28 | BSH Hausgeräte GmbH | Food preparation apparatus and method for food preparation |
| US20230360061A1 (en)* | 2022-05-05 | 2023-11-09 | Fiutur Information Exchange Inc. | Digital environmental claims ecosystem for physical commodity assets |
| WO2023246568A1 (en)* | 2022-06-22 | 2023-12-28 | 烽火通信科技股份有限公司 | Digital twin architecture, model evaluation system, and model running method |
| CN117408830A (en)* | 2023-11-02 | 2024-01-16 | 深圳市新纪元数据产业有限公司 | Agricultural product planting management system based on digital twinning |
| CN117475431A (en)* | 2023-12-27 | 2024-01-30 | 君华高科集团有限公司 | Food safety supervision method and system based on digital twin technology |
| US20240070527A1 (en)* | 2022-08-31 | 2024-02-29 | Bank Of America Corporation | Digital twin based evaluation, prediction, and forecasting for agricultural products |
| WO2024077271A2 (en)* | 2022-10-07 | 2024-04-11 | National Resilience, Inc. | Biosecure digital twin for cyber-physical anomaly detection and biological process modeling |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201207903Y (en)* | 2008-06-06 | 2009-03-18 | 王春成 | Unequal height continuous pulp boiling pot |
| KR20100126889A (en)* | 2009-05-25 | 2010-12-03 | 원두원 식품기계(주) | Tofu manufacturing equipment |
| CN114175044A (en)* | 2019-07-17 | 2022-03-11 | 克朗斯股份公司 | Production planning for predictive maintenance and optimization of devices in the food industry by digital twinning |
| US20230304674A1 (en)* | 2020-07-01 | 2023-09-28 | BSH Hausgeräte GmbH | Food preparation apparatus and method for food preparation |
| US20220087228A1 (en)* | 2020-09-23 | 2022-03-24 | International Business Machines Corporation | Industrial livestock management leveraging digital twin computing |
| US20220164502A1 (en)* | 2020-11-23 | 2022-05-26 | Jiangsu University | Pump machine unit optimized operation regulation system and method based on digital twin |
| WO2022233596A1 (en)* | 2021-05-03 | 2022-11-10 | Basf Coatings Gmbh | Methods and systems for determining the fill level of a liquid in a container |
| EP4174597A1 (en)* | 2021-10-29 | 2023-05-03 | Tetra Laval Holdings & Finance S.A. | A method for configuring a food processing system to produce a plurality of food products |
| CN118103785A (en)* | 2021-10-29 | 2024-05-28 | 利乐拉瓦尔集团及财务有限公司 | Method for configuring a food processing system to produce a variety of liquid food products |
| US20230360061A1 (en)* | 2022-05-05 | 2023-11-09 | Fiutur Information Exchange Inc. | Digital environmental claims ecosystem for physical commodity assets |
| WO2023246568A1 (en)* | 2022-06-22 | 2023-12-28 | 烽火通信科技股份有限公司 | Digital twin architecture, model evaluation system, and model running method |
| US20240070527A1 (en)* | 2022-08-31 | 2024-02-29 | Bank Of America Corporation | Digital twin based evaluation, prediction, and forecasting for agricultural products |
| WO2024077271A2 (en)* | 2022-10-07 | 2024-04-11 | National Resilience, Inc. | Biosecure digital twin for cyber-physical anomaly detection and biological process modeling |
| CN116369564A (en)* | 2023-06-06 | 2023-07-04 | 山西瑞飞机械制造有限公司 | Intelligent defoaming and deodorizing cooking pulp tank and its pulping process |
| CN117408830A (en)* | 2023-11-02 | 2024-01-16 | 深圳市新纪元数据产业有限公司 | Agricultural product planting management system based on digital twinning |
| CN117475431A (en)* | 2023-12-27 | 2024-01-30 | 君华高科集团有限公司 | Food safety supervision method and system based on digital twin technology |
| Publication number | Publication date |
|---|---|
| CN118805924B (en) | 2025-03-07 |
| Publication | Publication Date | Title |
|---|---|---|
| CN104414517A (en) | Food processing equipment and method for processing food by adopting food processing equipment | |
| US20050169099A1 (en) | Mixing system and process | |
| US20240122216A1 (en) | A food processing line and method for controlling a food processing line | |
| CN104873093A (en) | Full-automatic frying pan | |
| CN206612594U (en) | Cooking equipment | |
| CN118805924A (en) | A fully automatic pulp cooking system based on digital twin | |
| CN101598950A (en) | Intelligent control device and method for vacuum sugar pot temperature | |
| CN109247493A (en) | Full-automatic rice roll machine | |
| CN206609002U (en) | Automatic temperature measurement adjusts fire gas combustion range | |
| CN203065492U (en) | Automatic sugar cooking control device | |
| CN102293160B (en) | Processing equipment and method for industrially producing dried cheese | |
| CN208755685U (en) | Add water installations and steam box | |
| CN208410497U (en) | A kind of automatic liquid foam machine | |
| CN204090908U (en) | Bean curd machine | |
| Mittal | Process controls in the food industry: Problems and solutions | |
| CN103621646A (en) | Pot type pulp boiling device | |
| CN211504647U (en) | Intelligent toilet lid test machine | |
| CN118393995A (en) | Food processing production control method and system based on Internet of things technology | |
| CN207489157U (en) | One kind, which is cooked, makes and sell machine | |
| CN211983581U (en) | Full-automatic bean curd production machine | |
| CN206713962U (en) | Lab scale complete set of equipments system in a kind of process of cheese making | |
| CN205359288U (en) | Cake mix constant temperature timing stirring device | |
| CN108294113A (en) | Lab scale complete set of equipments system in a kind of process of cheese making | |
| CN202522943U (en) | Coagulated yoghourt discharging temperature control system | |
| CN104839330A (en) | Multifunctional yoghourt machine |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |