Injection molding raw material warehousing registration management system based on Internet of Things

Abstract

The invention discloses an injection molding raw material warehousing registration management system based on Internet of Things, which belongs to the technical field of warehousing registration management systems and comprises a raw material information acquisition module, a warehousing registration terminal, a data transmission and storage system, an inventory management module, a raw material tracing module, a report and statistical analysis module and a user management module. According to the injection molding raw material warehousing registration management system, by introducing the Internet of Things module, real-time monitoring and data collection of raw materials from warehousing to inventory management are achieved, and the data accuracy and timeliness are improved. The new warehousing registration terminal simplifies the process and ensures data integrity and consistency. The new inventory management module adopts a series of innovative technologies, a prediction model accurately predicts demands, a visualization function helps a manager to discover problems in time, a scheduling system optimizes configuration, a sharing cooperation mechanism promotes information sharing and cooperation, the response capability and flexibility of a supply chain are enhanced, and powerful support is provided for enterprise purchase and production.

Description

Technical Field

[0001] The present invention relates to the technical field of storage registration management systems, and in particular to a storage registration management system for injection molding raw materials based on the Internet of Things. Background Art

[0002] In the existing automobile parts production system, due to the advancement of new material technology, the cylinder head guard, intake manifold and other parts on the engine are generally injection molded. The use of injection molding to produce cylinder head guards and intake manifolds has significant advantages, including high cost-effectiveness, strong design flexibility, weight reduction, corrosion resistance, good insulation performance, fast production speed, high surface finish, high material utilization, easy integration, recyclability, color and texture diversity, and reduced maintenance. These characteristics make injection molded parts not only excellent in economy, aesthetics and environmental protection, but also improve the performance and durability of the product, while reducing production and maintenance costs, which is very suitable for the large-scale production of automobile parts.

[0003] Accordingly, in order to ensure quality and efficiency, the relevant processes of the injection molding process need to be strictly controlled. However, there are still many significant shortcomings and deficiencies in the traditional injection molding raw material warehousing registration and management technology.

[0004] First, in terms of data collection, existing technologies often rely on manual recording and manual input, which is not only inefficient but also prone to human errors, making it difficult to ensure the accuracy and completeness of data. Moreover, due to the lack of real-time monitoring methods, it is impossible to obtain the latest status and information of raw materials in a timely manner, making inventory management decisions often based on lagging and inaccurate data.

[0005] Secondly, the warehousing registration process is cumbersome and complicated. The existing warehousing registration method usually lacks convenient and efficient terminal equipment, is inconvenient to operate, does not collect comprehensive information, consumes a lot of time and manpower, and seriously affects work efficiency.

[0006] Furthermore, the inventory management module has serious defects. On the one hand, the lack of an accurate demand forecasting model makes it difficult for companies to accurately estimate inventory demand, which can easily lead to inventory backlogs or out-of-stock situations. On the other hand, the low level of inventory visualization means that managers cannot intuitively and clearly grasp the inventory status and make timely and effective decisions. In addition, inventory scheduling mostly relies on manual experience and lacks an intelligent scheduling system, resulting in unreasonable inventory configuration, low space utilization, and poor capital turnover. Moreover, there is insufficient information sharing and collaboration between departments and suppliers, and the entire supply chain is slow to respond, making it impossible to flexibly respond to market changes and emergencies. Summary of the invention

[0007] The main purpose of the present invention is to provide an injection molding raw material warehousing registration and management system based on the Internet of Things, which can effectively solve the problems mentioned in the background technology.

[0008] To achieve the above object, the technical solution adopted by the present invention is: An injection molding raw material warehousing registration management system based on the Internet of Things, the system consists of a raw material information collection module, a warehousing registration terminal, a data transmission and storage system, an inventory management module, a raw material traceability module, a report and statistical analysis module, and a user management module. Specifically: The raw material information collection module includes: IoT sensor: installed on the raw material packaging to collect basic information about the raw material packaging; RFID tag: Each batch of raw materials is equipped with a unique RFID tag to store detailed information and traceability code of the raw materials; The warehouse registration terminal includes: Scanning equipment: DWS equipment with automatic volume measurement, weighing and code scanning functions, which reads the barcode label information on the raw material packaging and obtains the weight and volume information of the raw materials in real time; The scanning equipment specifically includes: Intelligent image recognition technology: The integrated image acquisition module collects images of raw material packaging and uses image recognition algorithms to quickly and accurately identify information on the raw material packaging; Voice interaction function: equipped with voice collection, recognition and speech synthesis technology, operators can perform warehousing registration operations through voice commands; Data entry interface: manual entry when information cannot be automatically recognized, integrated into the scanning device in the form of touch screen and buttons; The data transmission and storage system includes: Wireless network: transmit the collected raw material information to the central database in real time; Central database: stores all relevant information about raw materials; Data processing center: covers the central server, database management system, data processing software and algorithms, and communication interface. It is responsible for receiving the collected inventory data, cleaning, converting, integrating and analyzing it, and storing the processing results in the central database; Data cloud processing platform: realize centralized management, integration and sharing of data through cloud computing technology; The inventory management module includes: AI-based forecasting model: Use AI technology to build an inventory forecasting model, analyze external factors and forecast raw material demand, and automatically adjust inventory strategies; Real-time inventory visualization: Real-time visual management of inventory is achieved through virtual reality (VR) or augmented reality (AR) technology; Intelligent inventory scheduling system: automatically dispatches appropriate raw materials to the production line according to production plan and inventory status; Inventory sharing and collaborative management: Using IoT technology to share inventory between departments within the enterprise and with suppliers; The raw material traceability module includes: Traceability code query: Through RFID tags, query including but not limited to the source of raw materials, production process, and transportation conditions; Quality traceability: When quality problems with raw materials are discovered, quickly trace them back to the relevant batches and suppliers and take appropriate measures; Report and statistical analysis modules include: Generate reports: Generate raw material registration and inventory reports based on demand; Statistical analysis: Conduct statistical analysis on the use of raw materials to provide data support for management decisions; The user management module includes: Permission setting: set different permissions for different users; User operation log: records user operation logs.

[0009] Preferably, the IoT sensor includes a temperature sensor, a humidity sensor, a pressure sensor and a vibration sensor, which respectively detect the temperature and humidity inside the raw material package, as well as the pressure exerted on the internal raw material package and the vibration during transportation; the above sensors are connected to the data transmission and storage system through wireless communication technology, and the collected information is transmitted to the central database of the system in real time.

[0010] Preferably, the intelligent image recognition technology is performed on an image processing unit of a processor of a scanning device, and is specifically implemented as follows: First, we choose the deep residual network (ResNet) as the core algorithm for image recognition. In the early preparation, we obtain a large-scale general image dataset from the Internet and pre-train the ResNet model. The pre-training enables the model to learn general image feature representations.

[0011] Next, we built a customized small-scale dataset based on the characteristics and requirements of injection molding raw material packaging images. This dataset covers various states, including but not limited to clear, stained, damaged, blurred, and raw material packaging images at various angles and lighting conditions, and accurately annotates the barcodes, QR codes, text, and picture information in the images.

[0012] Then, the pre-trained ResNet model was fine-tuned using a custom dataset. The model parameters were adjusted during fine-tuning to adapt it to the task of image recognition of injection molding raw material packaging. At the same time, in order to improve the recognition ability of stained, damaged or blurred images, image restoration technology based on generative adversarial networks (GAN) was introduced.

[0013] In the image acquisition process, high-resolution industrial-grade cameras are used to ensure that clear image details are captured, and adaptive fill lights are installed around the camera to ensure that clear images can be obtained under various lighting conditions.

[0014] When the collected image has stains, damage or blur, it is input into the pre-trained GAN model. GAN consists of a generator and a discriminator. The generator generates a repaired image, and the discriminator judges the authenticity of the generated image. Through continuous adversarial training, the generator learns to generate realistic repaired images. The repaired images are then input into the ResNet model for recognition; the recognized barcode, QR code, text, and picture information are encoded into a specific data format and transmitted to the data transmission and storage system in real time via a wireless network.

[0015] Preferably, the voice interaction function is implemented through the voice processing unit of the processor of the scanning device, using an end-to-end voice recognition model based on the attention mechanism, that is, a model of the Transformer architecture. The specific implementation method is: the operator speaks a voice command, and the sound signal is collected by a multi-microphone array, and is input into the voice recognition model after being processed by the beamforming technology. The model converts the voice into text and transmits it to the data transmission and storage system through the wireless network; at the same time, the feedback text generated by the system is input into the voice synthesis model, and the generated voice signal is played to the operator through the speaker.

[0016] Among them, voice collection uses a multi-microphone array, which consists of 4-8 microphones arranged in a circular or linear shape. These microphones collect sound signals simultaneously and are processed through beamforming technology. The beamforming technology enhances the target voice intensity according to the direction of the sound source and suppresses noise and interference in other directions; for speech synthesis, Tacotron 2, a neural speech synthesis technology based on deep learning, is selected.

[0017] Preferably, the establishment of the prediction model based on artificial intelligence specifically includes the following steps: ① Data collection: extensively collect historical inventory data, including the in-and-out records of different raw materials, and changes in inventory balances over time; sales data, covering the sales quantity, sales region, and sales time of each product; production plans, including product production arrangements and output expectations; market trend data, such as industry growth or recession trends, market demand fluctuations; seasonal factor data, such as changes in demand for certain raw materials in specific seasons; economic indicators, such as inflation rates, interest rates, etc.; these data are obtained through the company's internal database, sales system, production management system, and external market research and economic data platforms.

[0018] ② Feature engineering: Clean the collected data to remove duplicate, erroneous or missing values. Standardize or normalize the time series data to make it comparable, and convert the time series data into a format suitable for model input, such as converting continuous time points into equally spaced time periods. Extract key features from the data, such as converting events such as holidays, promotions, and new product launches into numerical features or Boolean features.

[0019] ③Model selection and training: Select the long short-term memory network (LSTM) as the prediction model, divide the processed data into training set, validation set and test set, use the training set to train the LSTM model, and adjust the model parameters, such as the number of neurons, number of layers, learning rate, etc., so that the model can learn the patterns and rules in the data.

[0020] ④Model evaluation and optimization: Use the validation set to evaluate the trained model and calculate the root mean square error (RMSE) and mean absolute error (MAE):

[0021]

[0022] in is the actual value, is the predicted value, is the number of samples; the smaller the value of the root mean square error, the better the prediction effect of the model; the smaller the value of the mean absolute error, the higher the prediction accuracy of the model. According to these two evaluation results, the hyperparameters of the model are adjusted by the random search method, that is, a set of hyperparameters are randomly selected from the range of hyperparameter values ​​for testing, repeated many times, and the hyperparameter combination with the best performance is selected to optimize the performance of the model.

[0023] ⑤ Real-time data update and prediction: Establish a real-time data collection mechanism to continuously input new data into the trained model. The model will make predictions based on the new data and output the forecast value of inventory demand in the future. According to the forecast results, the inventory strategy will be automatically adjusted, that is, when the predicted demand increases, replenishment will be arranged in advance, and when the predicted demand decreases, the inventory will be reduced.

[0024] Preferably, the implementation of real-time inventory visualization includes the following steps: ① Data collection and integration: Various sensors are deployed in the warehouse to collect real-time information about raw materials, such as the location, quantity, weight, etc. of the raw materials, and transmit these data to the data processing center in the data transmission and storage system through the wireless network, and integrate them with the existing data in the inventory database.

[0025] ②3D modeling and scene construction: Use professional 3D modeling software to create virtual scenes based on the actual layout and size of the warehouse, including the shape, height, number of layers of shelves, the size and color of cargo boxes, the width and direction of aisles, etc., and create unique identifiers for each item and storage location to associate with the collected data.

[0026] ③Data mapping and visualization: Map the integrated inventory data with the items and locations in the 3D model. Depending on the inventory quantity, the distribution and quantity of the inventory can be intuitively displayed through the depth of color, size of the icon or flashing frequency. For example, areas with sufficient inventory are displayed in green, and areas with tight inventory are displayed in red. At the same time, the data is updated in real time to ensure that the visualization effect reflects the latest inventory status.

[0027] ④VR / AR device integration: Select appropriate VR or AR devices, develop corresponding software interfaces, and integrate 3D inventory scenes with these devices. Users wearing VR devices can immerse themselves in a virtual warehouse environment and view and manage inventory through handles or gestures. When using AR devices, users can see superimposed virtual inventory information in the real warehouse environment.

[0028] Preferably, the intelligent inventory scheduling system is specifically embodied as follows: Production plan acquisition: Deeply integrate with the company's production management system to obtain detailed production plans in real time through API interfaces or data sharing protocols; the production plan should include accurate information such as product model, quantity, production process requirements, and expected start and end times.

[0029] Inventory data analysis: Real-time monitoring of inventory data, including current inventory quantity, storage location, remaining shelf life, inventory turnover rate, etc. of different raw materials. Analyze inventory turnover and flow trends to identify potential risk areas of inventory backlog or shortage.

[0030] Scheduling algorithm design: To design a scheduling strategy based on a genetic algorithm, first define the chromosome encoding method, represent each possible scheduling scheme as a chromosome, and then design the fitness function:

[0031] in A quantitative value indicating the quality of the body. Represents costs, including transportation costs, inventory holding costs, out-of-stock costs, etc. It represents the scheduling scheme used to punish the violation of constraint conditions, generates a new scheduling scheme population through crossover or mutation genetic operations, and continuously iterates and optimizes to find the optimal or approximately optimal scheduling scheme.

[0032] Automatic scheduling execution: According to the optimal scheduling plan generated by the scheduling algorithm, detailed scheduling instructions are automatically generated. These instructions are sent to the handling equipment in the warehouse, such as forklifts, automatic guided vehicles (AGVs), conveyor belts, etc. through the wireless network. After receiving the instructions, the handling equipment transports the appropriate amount of raw materials from the inventory location to the production line according to the specified route and sequence to ensure smooth production.

[0033] Preferably, the specific measures for inventory sharing and collaborative management are: IoT device deployment: IoT sensors, including temperature sensors, humidity sensors, pressure sensors, etc., as well as RFID readers and Bluetooth modules with communication functions, are installed in warehouses, production workshops, and suppliers. Environmental data and inventory-related data are collected in real time through these devices, and Zigbee or LoRaWAN wide area network technology is used to transmit the data to the central server in the data transmission and storage system.

[0034] Data platform construction: Use cloud computing technologies, such as Amazon AWS, Microsoft Azure, Alibaba Cloud, etc., to build a centralized data cloud processing platform. Create a database on the platform to store inventory data from various locations, and use data warehouse technology to integrate and manage data to ensure data consistency and availability.

[0035] Permission management and data sharing rule formulation: The role-based access control (RBAC) model sets different permissions for different users and departments. For example, warehouse managers can view and modify inventory quantities, production departments can view inventory to arrange production plans, and suppliers can only view the inventory of raw materials they supply. Detailed data sharing rules are formulated to specify which data can be shared between which departments and suppliers, as well as the frequency and method of sharing.

[0036] Establishment of a collaborative mechanism: When inventory is lower than the safety inventory threshold, the system automatically sends inventory warning notifications to relevant departments and suppliers; when production plans change, suppliers are notified in a timely manner to adjust supply plans; a process for handling sudden demand changes is established, and when market demand changes suddenly, departments and suppliers are notified in a timely manner and jointly negotiate solutions. Notification methods include email, text messages, and system messages.

[0037] Preferably, the specific steps of the traceability code query are: ①RFID tag reading and writing: Use a scanning device to read the RFID tag information on the raw material packaging and obtain the traceability code.

[0038] ② Database query: Use the traceability code as a keyword to query the raw material database and obtain relevant information.

[0039] ③Data display: Display the queried information to users in the form of tables, charts or reports.

[0040] Preferably, the specific steps of the quality tracing are: ① Quality problem detection: Quality problems of raw materials are found through quality testing equipment or manual testing.

[0041] ② Batch determination: Determine the relevant raw material batches based on the characteristics and scope of the quality issue.

[0042] ③ Traceability query: Use batch information to query in the raw material database and trace back to the corresponding supplier and each link in the production process.

[0043] ④ Formulation and implementation of measures: Based on the traceability results, formulate and implement corresponding treatment measures.

[0044] Compared with the prior art, the present invention has the following beneficial effects: The injection molding raw material storage registration and management system of the present invention introduces an advanced Internet of Things module, which realizes real-time monitoring and data collection of the entire process of injection molding raw materials from storage to inventory management. Through the Internet of Things technology, various sensors can accurately sense key information such as the state, quantity, and location of the raw materials, and transmit these data to the system in real time. This greatly improves the accuracy and timeliness of the data, reduces the errors that may be caused by manual intervention, and provides a reliable basis for subsequent inventory management and decision-making.

[0045] An innovative warehouse registration terminal has been designed to simplify the warehouse registration process and improve work efficiency. It can quickly and accurately collect detailed information on the incoming raw materials, including the type, specification, batch, production date, etc. At the same time, the terminal has a good user interface and is easy to operate, which reduces the operator's work difficulty and error rate. This innovative design effectively improves the speed and quality of warehouse registration and ensures the integrity and consistency of the data.

[0046] A new inventory management module has been designed, which uses a series of innovative technologies and methods. The forecasting model based on artificial intelligence can accurately predict inventory demand and provide strong support for enterprises to arrange procurement and production reasonably. The real-time inventory visualization function allows managers to understand the inventory status intuitively and clearly, find problems in time and make adjustments. The intelligent inventory scheduling system can automatically optimize inventory configuration, improve the utilization rate of inventory space and capital turnover rate. The inventory sharing and collaborative management mechanism promotes information sharing and collaboration among departments and suppliers, and enhances the responsiveness and flexibility of the entire supply chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the composition structure and flow chart of the present invention; Figure 2 It is a schematic diagram of the composition and process of the storage registration terminal of the present invention; Figure 3 It is a schematic diagram of the inventory management module composition and flow chart of the present invention. DETAILED DESCRIPTION

[0048] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0049] like Figure 1-3 As shown in the figure, an injection molding raw material warehousing registration management system based on the Internet of Things is mainly composed of a raw material information collection module, a warehousing registration terminal, a data transmission and storage system, an inventory management module, a raw material traceability module, a report and statistical analysis module, and a user management module. The following is a detailed description combined with actual application examples:

[0050] IoT sensors: Temperature sensors, humidity sensors, pressure sensors, and vibration sensors are pre-installed inside the raw material packaging. For example, the temperature sensor uses the high-precision digital temperature sensor DS18B20, which has a measurement range of -55°C to +125°C and an accuracy of ±0.5°C. The humidity sensor uses the capacitive humidity sensor HIH6130, which can measure relative humidity from 0% to 100% with an accuracy of ±3%. The pressure sensor uses the strain pressure sensor MPX5700, which has a measurement range of 0 to 700kPa and an accuracy of ±1%. The vibration sensor uses the MEMS vibration sensor ADXL345, which can detect vibrations in three directions. These sensors are connected to the data transmission and storage system through Zigbee wireless communication technology, and the collected information is transmitted to the system's central database in real time.

[0051] RFID tags: Each batch of raw materials is equipped with an ultra-high frequency RFID tag that complies with the ISO18000-6C standard and has a storage capacity of 512 bytes. The tag stores detailed information about the raw materials, such as raw material name, specification, batch, production date, shelf life, supplier information, traceability code, etc. At the same time, an RFID reader is installed on the scanning device to ensure accurate reading of the tag information.

[0052] 2. Implementation of the warehouse registration terminal Scanning equipment: DWS equipment with automatic volume measurement, weighing and code scanning functions is used. Its weighing module uses a high-precision strain gauge weighing sensor with a measurement accuracy of up to 0.1 kg. Volume measurement uses laser ranging technology with a measurement accuracy of ±1 cubic centimeter. The code scanning module integrates a high-resolution industrial-grade camera and advanced image recognition algorithm.

[0053] Intelligent image recognition technology: The image acquisition module uses an industrial-grade camera with a resolution of 5 million pixels, which can clearly capture various information on the raw material packaging. On the image processing unit, the deep residual network (ResNet) is selected as the core algorithm for image recognition. In the early preparation, the ImageNet data set containing millions of general images was obtained from the Internet to pre-train the ResNet model. Then, based on the characteristics and requirements of injection molding raw material packaging images, a custom data set containing 10,000 raw material packaging images in different states (clear, stained, damaged, blurred, and various angles and lighting conditions) was constructed, and the barcode, QR code, text, and picture information in the image were accurately annotated. Then, the pre-trained ResNet model was fine-tuned using the custom data set. At the same time, the image restoration technology based on the generative adversarial network (GAN) was introduced. When the captured image is stained, damaged or blurred, it is input into the pre-trained GAN model for restoration, and the restored image is then input into the ResNet model for recognition.

[0054] Voice interaction function: Voice collection uses a multi-microphone array with 6 microphones arranged in a ring, and processes sound signals through beamforming technology. The voice recognition model adopts an end-to-end Transformer architecture model based on the attention mechanism, which can quickly and accurately convert the operator's voice instructions into text. The speech synthesis uses Tacotron2, a neural speech synthesis technology based on deep learning, and the generated speech is clear and natural.

[0055] Data entry interface: The data entry interface is integrated into the scanning device in the form of a touch screen and buttons. The touch screen uses a capacitive touch screen, which is responsive and easy to operate. When the information cannot be automatically recognized, the operator can manually enter the relevant data through the touch screen and buttons.

[0056] Wireless network: Wi-Fi 6 wireless network technology is adopted, with a transmission rate of up to 9.6Gbps, ensuring that the collected raw material information can be transmitted to the central database in real time and stably.

[0057] Central database: A relational database based on SQL Server 2019 is selected to store all relevant information of raw materials, including data collected by sensors, RFID tag information, data collected by the warehousing registration terminal, etc. The database adopts a master-slave replication architecture to achieve data backup and high availability.

[0058] Data Processing Center: The data processing center is equipped with a high-performance central server. The database management system uses MySQL 8.0, and the data processing software and algorithms include Pandas, NumPy, Scikit-learn and other libraries in Python. The communication interface uses a network communication protocol based on TCP / IP.

[0059] Data cloud processing platform: A data cloud processing platform was built through Alibaba Cloud's cloud computing services, and its elastic computing, object storage and other services were used to achieve centralized management, integration and sharing of data.

[0060] AI-based predictive models Data collection: historical inventory data, sales data, production plan and other information are obtained from the company's internal database (such as ERP system, inventory management system), sales system (such as CRM system), and production management system. At the same time, market trend data, seasonal factor data, economic indicators, etc. are obtained from external market research websites and economic data platforms through web crawler technology. For example, the market demand trend data of injection molding raw materials in the past five years was obtained from the industry research website, and the inflation rate and interest rate data for the same period were obtained from the website of the National Bureau of Statistics.

[0061] Feature Engineering: Clean and preprocess the collected data. Use Python's Pandas library to remove duplicate, erroneous or missing values. For time series data, use the Z-Score standardization method for standardization. Convert the time series data into weekly time periods and extract event features such as holidays and promotions.

[0062] Model selection and training: Long short-term memory network (LSTM) was selected as the prediction model, and the Keras framework was used for model training. The processed data was divided into training set, validation set, and test set in a ratio of 7:2:1. During the training process, the parameters of the model were adjusted, such as the number of neurons was 128, the number of layers was 2, and the learning rate was 0.001.

[0063] Model evaluation and optimization: Use the validation set to evaluate the trained model and calculate the root mean square error (RMSE) and mean absolute error (MAE). For example, the RMSE of the validation set is 50.2 and the MAE is 35.8. Based on the evaluation results, the model's hyperparameters are adjusted through a random search method. After 50 random trials, the hyperparameter combination with the best performance is selected.

[0064] Real-time data update and prediction: A real-time data collection mechanism based on a real-time database is established, and new data is input into the trained model every hour. The model makes predictions based on the new data and outputs the inventory demand forecast for the next month. For example, it is predicted that the demand for a certain raw material will increase by 20% next week, and replenishment is arranged in advance based on the forecast results.

[0065] Real-time inventory visibility Data collection and integration: Various sensors are deployed in the warehouse, including ultrasonic sensors for measuring the position of raw materials, pressure sensors for measuring the weight of raw materials, etc. These sensors transmit data to the data processing center through the Zigbee wireless network and integrate it with the existing data in the inventory database.

[0066] 3D modeling and scene construction: Using 3dsMax professional 3D modeling software, a virtual scene was created based on the actual layout and size of the warehouse (50 meters long, 30 meters wide, and 10 meters high). The shelf is in the shape of a cuboid, 2 meters high, with 5 layers, the size of the cargo box is 50cm×50cm×50cm, the color is blue, the width of the aisle is 2 meters, and the direction is east-west. A unique identifier was created for each item and storage location, such as the identifier for the first position on the first floor of the shelf is "R1L1P1".

[0067] Data mapping and visualization: Map the integrated inventory data to the items and locations in the 3D model. Areas with sufficient inventory (more than 500 items) are displayed in green, and areas with tight inventory (less than 100 items) are displayed in red. The data is updated in real time every 5 minutes to ensure that the visualization reflects the latest inventory status.

[0068] VR / AR device integration: HTC Vive Pro virtual reality devices and Microsoft HoloLens augmented reality devices were selected. Corresponding software interfaces were developed to integrate 3D inventory scenes with these devices. Users wearing VR devices can immerse themselves in a virtual warehouse environment and view and manage inventory through handles; when using AR devices, users can see virtual inventory information superimposed in the real warehouse environment.

[0069] Intelligent inventory scheduling system Production plan acquisition: Deeply integrate with the enterprise's production management system (such as MES system) through API interface to obtain detailed production plans in real time. For example, obtain the production plan for 10,000 pieces of injection molded product A to be produced next week, and each piece of product A consumes 2 kilograms of raw material X.

[0070] Inventory data analysis: Real-time monitoring of inventory data, including the current inventory quantity of raw material X is 1,500 kg, the storage location is warehouse zone 2, row 3, column 5, the remaining shelf life is 6 months, and the inventory turnover rate is 2 times per month. By analyzing the inventory turnover rate and flow trend, it is found that raw material X is at risk of inventory backlog.

[0071] Scheduling algorithm design: Scheduling strategies are designed based on genetic algorithms. The chromosomes are encoded in binary code, and each gene bit represents a scheduling solution. These include transportation costs (50 yuan per ton), inventory holding costs (0.1 yuan per piece per day), out-of-stock costs (10 yuan per piece), etc., which are used to punish scheduling solutions that violate constraints (such as transportation time exceeding 2 hours). A new scheduling solution population is generated through genetic operations with a crossover probability of 0.8 and a mutation probability of 0.2, and it is continuously iterated and optimized. After 100 iterations, the optimal scheduling solution is found.

[0072] Automatic scheduling execution: According to the optimal scheduling plan generated by the scheduling algorithm, detailed scheduling instructions are automatically generated, such as "Move 2,000 kg of raw material X from warehouse area 2, row 3, column 5 to production line 1". These instructions are sent to the forklifts in the warehouse via the Wi-Fi wireless network. After receiving the instructions, the forklifts transport the raw materials to the production line according to the specified route and sequence.

[0073] Inventory sharing and collaborative management IoT device deployment: IoT sensors are installed in warehouses, production workshops, and suppliers. For example, temperature sensors (TMP117), humidity sensors (HDC2080), pressure sensors (MPRLS0250DPA), and RFID readers (ImpinjR2000) and Bluetooth modules (CSR8811) with communication functions are installed in warehouses. Data is transmitted to the central server through LoRaWAN wide area network technology.

[0074] Data platform construction: We used Amazon AWS cloud computing services to build a centralized data cloud processing platform. We created a PostgreSQL-based database on the platform to store inventory data from various locations. We used data warehouse technology, such as Amazon Redshift, to integrate and manage data to ensure data consistency and availability.

[0075] Permission management and data sharing rule formulation: The role-based access control (RBAC) model sets different permissions for different users and departments. For example, warehouse managers have the permission to view and modify inventory quantities, production departments can view inventory to arrange production plans, and suppliers can only view the inventory of raw materials they supply. The formulated data sharing rules stipulate that inventory quantity and location information can be shared between warehouses, production departments, and suppliers every day, and the sharing method is to send reports via email.

[0076] Establishment of a collaborative mechanism: When the inventory is lower than the safety inventory threshold (e.g., the inventory quantity of a certain raw material is lower than 100 kg), the system automatically sends an inventory warning notification to the relevant departments (e.g., the purchasing department, the production department) and the supplier. For example, a notification of "Raw material X is insufficient in stock, please purchase as soon as possible" is sent to the purchasing department via email, and a notification of "need to increase the supply of raw material X" is sent to the supplier via SMS. When the production plan changes (e.g., the production of a certain product is advanced or postponed), the supplier is notified in time to adjust the supply plan. A process for handling sudden demand changes has been established. When market demand suddenly increases (e.g., the order of a certain product increases by 50%) or decreases (e.g., the order of a certain product decreases by 30%), the departments and suppliers are notified in time through messages within the system, and a video conference is held to jointly negotiate a solution.

[0077] Traceability code query RFID tag reading and writing: Use a handheld RFID reader / writer (model: AlienALR-9900) to read the RFID tag information on the raw material packaging and obtain the traceability code.

[0078] Database query: Use the traceability code as a keyword to query the raw material database based on MySQL to obtain relevant information, such as the source of the raw materials (supplier name, address), production process (production date, production batch, production process), transportation conditions (transportation company, transportation time), etc.

[0079] Data display: The queried information is displayed on the computer screen in the form of tables and charts to facilitate user viewing and analysis.

[0080] Quality traceability Quality problem detection: Quality problems of raw materials are found through quality detection equipment (such as X-ray detectors, component analyzers) or manual detection. For example, foreign matter is detected in raw materials using X-ray detectors.

[0081] Batch determination: Determine the relevant raw material batches based on the characteristics and scope of the quality problem. For example, based on the distribution of foreign matter and the time of detection, it is determined that the raw material batch No. 202307 has quality problems.

[0082] Traceability query: Use batch information to query in the raw material database and trace back to the corresponding supplier and each link in the production process. For example, the supplier of this batch of raw materials is XX Company, and there is an abnormality in the mixing link in the production process.

[0083] Formulation and implementation of measures: Based on the traceability results, formulate corresponding treatment measures, such as recalling the batch of raw materials, requiring suppliers to improve production processes, maintaining production equipment, etc., and implement them in a timely manner.

[0084] Generate reports: Use Python's ReportLab library to generate raw material registration and inventory reports based on demand. Reports include raw material entry registration form (including entry time, raw material name, specification, batch, quantity and other information), daily inventory report (including inventory quantity, location, shelf life and other information of each raw material), monthly inventory report (including monthly in and out quantity, inventory turnover rate, inventory cost and other information), etc.

[0085] Statistical analysis: Use Python's Matplotlib library to perform statistical analysis on the use of raw materials. For example, draw a line graph to show the trend of the inventory quantity of a certain raw material over time, draw a bar graph to compare the quality of raw materials supplied by different suppliers, draw a pie chart to analyze the proportion of various raw materials in the inventory, etc. Provide data support for management decisions, such as adjusting procurement strategies and optimizing inventory structure based on statistical analysis results.

[0086] Permission settings: Set different permissions for different users. For example, the administrator has the highest permissions and can view and modify all system settings and data; the warehouse administrator can view and modify warehouse-related information; the purchaser can view purchase-related information but cannot modify inventory data; the supplier can only view the information related to the raw materials it supplies.

[0087] User operation records: Use database tables to record user operation logs, including operation time, user account, operation type (such as login, query, modification, deletion, etc.), operation object (such as raw material name, inventory quantity, etc.), operation result (success or failure), etc. Audit and analyze operation logs regularly to ensure system security and data integrity.

[0088] Through the above detailed embodiments, the injection molding raw material warehousing registration and management system based on the Internet of Things can realize efficient, accurate and intelligent injection molding raw material management, and provide strong support for the production and operation of the enterprise.

[0089] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An injection molding raw material storage registration and management system based on the Internet of Things, characterized by: The system consists of raw material information collection module, warehousing registration terminal, data transmission and storage system, inventory management module, raw material traceability module, report and statistical analysis module and user management module. Specifically: The raw material information collection module includes: IoT sensor: installed on the raw material packaging to collect basic information about the raw material packaging; RFID tag: Each batch of raw materials is equipped with a unique RFID tag to store detailed information and traceability code of the raw materials; The warehouse registration terminal includes: Scanning equipment: DWS equipment with automatic volume measurement, weighing and code scanning functions, which reads the barcode label information on the raw material packaging and obtains the weight and volume information of the raw materials in real time; The scanning equipment specifically includes: Intelligent image recognition technology: The integrated image acquisition module collects images of raw material packaging and uses image recognition algorithms to quickly and accurately identify information on the raw material packaging; Voice interaction function: equipped with voice collection, recognition and speech synthesis technology, operators can perform warehousing registration operations through voice commands; Data entry interface: manual entry when information cannot be automatically recognized, integrated into the scanning device in the form of touch screen and buttons; The data transmission and storage system includes: Wireless network: transmit the collected raw material information to the central database in real time; Central database: stores all relevant information about raw materials; Data processing center: covers the central server, database management system, data processing software and algorithms, and communication interface. It is responsible for receiving the collected inventory data, cleaning, converting, integrating and analyzing it, and storing the processing results in the central database; Data cloud processing platform: realize centralized management, integration and sharing of data through cloud computing technology; The inventory management module includes: AI-based forecasting model: Use AI technology to build an inventory forecasting model, analyze external factors and forecast raw material demand, and automatically adjust inventory strategies; Real-time inventory visualization: Real-time visual management of inventory is achieved through virtual reality (VR) or augmented reality (AR) technology; Intelligent inventory scheduling system: automatically dispatches appropriate raw materials to the production line according to production plan and inventory status; Inventory sharing and collaborative management: Using IoT technology to share inventory between departments within the enterprise and with suppliers; The raw material traceability module includes: Traceability code query: Through RFID tags, query including but not limited to the source of raw materials, production process, and transportation conditions; Quality traceability: When quality problems with raw materials are discovered, quickly trace them back to the relevant batches and suppliers and take appropriate measures; Report and statistical analysis modules include: Generate reports: Generate raw material registration and inventory reports based on demand; Statistical analysis: Conduct statistical analysis on the use of raw materials to provide data support for management decisions; The user management module includes: Permission setting: set different permissions for different users; User operation log: records user operation logs.

2. According to the Internet of Things-based injection molding raw material storage registration and management system according to claim 1, it is characterized by: The IoT sensors include temperature sensors, humidity sensors, pressure sensors and vibration sensors, which respectively detect the temperature and humidity inside the raw material packaging, as well as the pressure on the internal raw material packaging and the vibration during transportation; the above sensors are connected to the data transmission and storage system through wireless communication technology, and the collected information is transmitted to the central database of the system in real time.

3. According to the Internet of Things-based injection molding raw material storage registration and management system according to claim 1, it is characterized by: The intelligent image recognition technology is performed on the image processing unit of the processor of the scanning device, and is specifically implemented as follows: First, we selected the deep residual network (ResNet) as the core algorithm for image recognition. In the early preparation, we obtained a large-scale general image dataset from the Internet and pre-trained the ResNet model. The pre-training enabled the model to learn general image feature representations. Next, we built a customized small-scale dataset based on the characteristics and requirements of injection molding raw material packaging images. This dataset covers various states, including but not limited to clear, stained, damaged, blurred, and raw material packaging images at various angles and lighting conditions, and accurately annotates the barcodes, QR codes, text, and picture information in the images. Then, the pre-trained ResNet model was fine-tuned using a custom dataset. During the fine-tuning process, the model parameters were adjusted to adapt it to the task of image recognition of injection molding raw material packaging. At the same time, in order to improve the recognition ability of stained, damaged or blurred images, the image restoration technology based on the generative adversarial network (GAN) was introduced. In the image acquisition process, a high-resolution industrial-grade camera is used to ensure the capture of clear image details, and an adaptive fill light is equipped around the camera; When the collected image has stains, damage or blur, it is input into the pre-trained GAN model. GAN consists of a generator and a discriminator. The generator generates a repaired image, and the discriminator judges the authenticity of the generated image. Through continuous adversarial training, the generator learns to generate realistic repaired images. The repaired images are then input into the ResNet model for recognition; the recognized barcode, QR code, text, and picture information are encoded into a specific data format and transmitted to the data transmission and storage system in real time via a wireless network.

4. According to the Internet of Things-based injection molding raw material storage registration and management system according to claim 3, it is characterized by: The voice interaction function is implemented through the voice processing unit of the processor of the scanning device, using an end-to-end voice recognition model based on the attention mechanism, that is, a model of the Transformer architecture. The specific implementation method is as follows: the operator speaks a voice command, the sound signal is collected by a multi-microphone array, and is input into the voice recognition model after being processed by the beamforming technology. The model converts the voice into text and transmits it to the data transmission and storage system through the wireless network; at the same time, the feedback text generated by the system is input into the voice synthesis model, and the generated voice signal is played to the operator through the speaker; Among them, voice collection uses a multi-microphone array, which consists of 4-8 microphones arranged in a circular or linear shape. These microphones collect sound signals simultaneously and are processed through beamforming technology. The beamforming technology enhances the target voice intensity according to the direction of the sound source and suppresses noise and interference in other directions; for speech synthesis, Tacotron 2, a neural speech synthesis technology based on deep learning, is selected.

5. According to the Internet of Things-based injection molding raw material storage registration and management system as claimed in claim 1, it is characterized by: The establishment of the prediction model based on artificial intelligence specifically includes the following steps: ① Data collection: extensively collect historical inventory data, sales data, production plans, market trend data, seasonal factor data, and economic indicators; obtain these data through the company's internal database, sales system, production management system, and external market research and economic data platforms; ② Feature engineering: clean the collected data to remove duplicate, erroneous or missing values; standardize or normalize the time series data, convert the time series data into a format suitable for model input, and extract key features from the data, that is, convert events into numerical features or Boolean features; ③Model selection and training: Select the long short-term memory network (LSTM) as the prediction model, divide the processed data into training set, validation set and test set, use the training set to train the LSTM model, and adjust the parameters of the model so that the model can learn the patterns and rules in the data; ④Model evaluation and optimization: Use the validation set to evaluate the trained model and calculate the root mean square error (RMSE) and mean absolute error (MAE): ; in is the actual value, is the predicted value, is the number of samples; the smaller the value of the root mean square error, the better the prediction effect of the model; the smaller the value of the mean absolute error, the higher the prediction accuracy of the model. According to the results of these two evaluations, the hyperparameters of the model are adjusted by the random search method, that is, a set of hyperparameters are randomly selected from the range of hyperparameter values ​​for testing, repeated multiple times, and the best hyperparameter combination is selected to optimize the performance of the model; ⑤ Real-time data update and prediction: Establish a real-time data collection mechanism to continuously input new data into the trained model. The model will make predictions based on the new data and output the forecast value of inventory demand in the future. According to the forecast results, the inventory strategy will be automatically adjusted, that is, when the predicted demand increases, replenishment will be arranged in advance, and when the predicted demand decreases, the inventory will be reduced.

6. The injection molding raw material storage registration and management system based on the Internet of Things according to claim 1 is characterized by: The implementation of real-time inventory visualization includes the following steps: ① Data collection and integration: Deploy various sensors in the warehouse to collect relevant information of raw materials in real time, transmit these data to the data processing center in the data transmission and storage system through the wireless network, and integrate them with the existing data in the inventory database; ②3D modeling and scene construction: Use professional 3D modeling software to create a virtual scene based on the actual layout and size of the warehouse, and create a unique identifier for each item and storage location to associate with the collected data; ③Data mapping and visualization: Map the integrated inventory data with the items and locations in the 3D model, and intuitively display the distribution and quantity of inventory according to the different inventory quantities; at the same time, update the data in real time to ensure that the visualization effect reflects the latest inventory status; ④VR / AR device integration: Select appropriate VR or AR devices, develop corresponding software interfaces, and integrate 3D inventory scenes with these devices. Users wearing VR devices can immerse themselves in a virtual warehouse environment and view and manage inventory through handles or gestures. When using AR devices, users can see superimposed virtual inventory information in the real warehouse environment.

7. The injection molding raw material storage registration and management system based on the Internet of Things according to claim 1 is characterized by: The intelligent inventory scheduling system is specifically embodied as follows: Production plan acquisition: Deeply integrate with the enterprise's production management system to obtain detailed production plans in real time through API interfaces or data sharing protocols; Inventory data analysis: real-time monitoring of inventory data, analysis of inventory turnover and flow trends, and identification of potential risk areas of inventory overstocking or shortage; Scheduling algorithm design: To design a scheduling strategy based on a genetic algorithm, first define the chromosome encoding method, represent each possible scheduling scheme as a chromosome, and then design the fitness function: in A quantitative value indicating the quality of the body. Indicates the cost, It represents the scheduling scheme used to punish the violation of constraint conditions, generates a new scheduling scheme population through crossover or mutation genetic operations, and continuously iterates and optimizes to find the optimal or approximately optimal scheduling scheme; Automatic scheduling execution: According to the optimal scheduling plan generated by the scheduling algorithm, detailed scheduling instructions are automatically generated. These instructions are sent to the handling equipment in the warehouse through the wireless network. After receiving the instructions, the handling equipment transports the appropriate amount of raw materials from the inventory location to the production line according to the specified route and sequence.

8. The injection molding raw material storage registration and management system based on the Internet of Things according to claim 1 is characterized by: The specific measures for inventory sharing and collaborative management are: IoT device deployment: IoT sensors, RFID readers and Bluetooth modules with communication functions are installed in warehouses, production workshops, and suppliers. Environmental data and inventory-related data are collected in real time through these devices, and Zigbee or LoRaWAN wide area network technology is used to transmit the data to the central server in the data transmission and storage system; Data platform construction: Use cloud computing technology to build a centralized data cloud processing platform, create a database on the platform to store inventory data from various locations, and use data warehouse technology to integrate and manage data; Permission management and data sharing rule formulation: The role-based access control (RBAC) model sets different permissions for different users and departments, and formulates detailed data sharing rules to specify which data can be shared between which departments and suppliers, as well as the frequency and method of sharing; Establishment of a collaborative mechanism: When inventory is lower than the safety inventory threshold, the system automatically sends inventory warning notifications to relevant departments and suppliers; when production plans change, suppliers are notified in a timely manner to adjust supply plans; a process for handling sudden demand changes is established, and when market demand changes suddenly, departments and suppliers are notified in a timely manner and solutions are resolved through joint consultation.

9. The injection molding raw material storage registration and management system based on the Internet of Things according to claim 1 is characterized by: The specific steps of the traceability code query are: ①RFID tag reading and writing: Use a scanning device to read the RFID tag information on the raw material packaging and obtain the traceability code; ② Database query: Use the traceability code as a keyword to query the raw material database and obtain relevant information; ③Data display: Display the queried information to users in the form of tables, charts or reports.

10. The system for registering and managing the injection molding raw materials based on the Internet of Things according to claim 1 is characterized by: The specific steps of quality traceability are: ① Quality problem detection: Quality problems of raw materials are found through quality testing equipment or manual testing; ② Batch determination: Determine the relevant raw material batches based on the characteristics and scope of the quality problem; ③Traceability query: Use batch information to query in the raw material database and trace back to the corresponding supplier and each link in the production process; ④ Formulation and implementation of measures: Based on the traceability results, formulate and implement corresponding treatment measures.

Images