On-line measurement and timely observation system for hairy cotton seeds and lustrous cotton seeds

Abstract

The application provides an online metering and timely observation system for fuzzy cotton seeds and light cotton seeds, realizes synchronous, real-time and accurate monitoring and self-adaptive regulation of material flow and key quality parameters of fuzzy cotton seeds and light cotton seeds with significant characteristic differences by integrating online metering, state observation, intelligent data processing and man-machine interaction functions and building a closed-loop control system, fundamentally solves the core problems of metering distortion, quality detection lag and production control disconnection in the traditional mode, and significantly improves the continuity, automation level, quality of the final product and economic benefits of the cotton seed processing process.

Description

Technical Field

[0001] This invention relates to the field of agricultural product processing technology, and in particular to an online metering and real-time monitoring system for raw cottonseed and smooth cottonseed. Background Technology

[0002] Cottonseed is a byproduct of cotton production and an important source of oil and feed. During cottonseed processing, the raw cottonseed obtained from the ginning mill has a large amount of short fibers attached to its surface. It needs to be delinted to become smooth cottonseed before it can be used for subsequent oil extraction or planting. The core of this processing lies in the precise measurement of the material flow between raw and smooth cottonseed, and the real-time monitoring and control of its processing quality (such as short fiber residue, seed breakage, and impurities). This directly affects the oil yield, seed germination rate, and the economic benefits of the final product.

[0003] Traditional cottonseed processing measurement and quality inspection methods are usually independent. Measurement often uses ordinary belt scales or screw scales, while quality inspection relies on manual periodic sampling and offline analysis. However, due to the strong adhesion of raw cottonseed, it is easy for residues to accumulate on measuring equipment, leading to measurement drift and distortion; while gleaming cottonseed, due to its high fluidity and impact, is difficult for traditional measurement methods to accurately capture its instantaneous flow rate. At the same time, offline manual quality inspection has a serious lag, failing to reflect the production status in real time, and by the time quality problems are discovered, a large number of defective products have already been produced. This disconnect between measurement and quality control is difficult to adapt to the needs of continuous and automated production, becoming the biggest drawback restricting the improvement of cottonseed processing quality and efficiency. Summary of the Invention

[0004] In view of this, the present invention provides an online measurement and real-time observation system for cottonseed and spun cottonseed to address the technical deficiencies in the prior art.

[0005] Specifically, the present invention provides an online measurement and real-time monitoring system for woolly cottonseed and gleaming cottonseed, comprising:

[0006] The online metering subsystem is used to measure and output material flow data for raw cottonseed and gleaming cottonseed.

[0007] The observation subsystem is used to detect and output material status parameters online;

[0008] The intelligent data processing center is used to receive, integrate, and process material flow data and material status parameters to generate control commands;

[0009] Human-computer interaction terminal, used for parameter setting and status display;

[0010] The online metering subsystem, observation subsystem, intelligent data processing center, and human-machine interaction terminal are connected through an industrial network to form a closed-loop control system.

[0011] In some implementations, the online metering subsystem includes a raw cottonseed metering unit and a polished cottonseed metering unit. The raw cottonseed metering unit adopts an integrated design of an anti-adhesion belt scale and a self-cleaning mechanism to collect the instantaneous flow rate and cumulative yield of raw cottonseed. The polished cottonseed metering unit adopts an integrated design of impact-resistant pipeline weighing and flow rate measurement to collect the instantaneous flow rate and cumulative yield of polished cottonseed. The raw cottonseed metering unit and the polished cottonseed metering unit are set up in parallel to process two materials with different characteristics simultaneously.

[0012] In some embodiments, the cottonseed metering unit includes an ultra-high molecular weight polyethylene coated conveyor belt, a tungsten carbide scraper cleaner, a weight sensor, and a speed sensor. The tungsten carbide scraper cleaner performs periodic self-cleaning motions to remove residual short fibers. Data from the weight sensor and speed sensor are used to calculate cottonseed flow data using a dynamic weighing algorithm located locally in the cottonseed metering unit.

[0013] In some embodiments, the light cottonseed metering unit includes a pipe weighing module, an ultrasonic Doppler flow velocity sensor, and a silicon nitride ceramic guide plate. The pipe weighing module is designed to withstand impact and is used to directly measure the net weight of the material in the pipe. The ultrasonic Doppler flow velocity sensor measures the material flow rate in a non-contact manner. The data from the pipe weighing module and the ultrasonic Doppler flow velocity sensor are used to calculate the light cottonseed flow rate data through a product integral model located locally in the light cottonseed metering unit.

[0014] In some implementations, the observation subsystem includes an image acquisition unit and a state parameter identification unit. The image acquisition unit uses an industrial CMOS camera and a specific light source to capture images of the material surface. The state parameter identification unit uses a deep learning algorithm model to process the images, identify and output parameters such as short fiber coverage, breakage rate and impurity content.

[0015] In some implementations, the state parameter identification unit uses an improved U-Net semantic segmentation model to calculate the short fiber coverage rate; it uses a ResNet-50 architecture classification model to identify damaged grains and impurities, and calculates the damage rate and impurity content accordingly.

[0016] In some implementations, the intelligent data processing center includes a data receiving and preprocessing module, an intelligent judgment and control module, and a data storage and reporting module. The data receiving and preprocessing module is used to perform temperature compensation and zero-point drift correction on sensor data from the online metering subsystem. The intelligent judgment and control module is used to compare real-time data with preset thresholds and generate control commands. The data storage and reporting module is used to store all process data and generate visual reports.

[0017] In some implementations, the intelligent judgment and control module generates instructions to adjust the speed of the upstream feeding equipment when the material flow rate deviates from the set value; and triggers an audible and visual alarm and generates instructions to adjust the operating parameters of the cleaning equipment when the quality parameters exceed the standard.

[0018] In some implementations, the online metering and real-time observation system for raw cottonseed and bare cottonseed also includes an environmental sensor network and an equipment configuration file storage module. The environmental sensor network is used to monitor changes in environmental factors such as temperature and humidity, and the equipment configuration file storage module is used to store equipment parameters that affect the control effect.

[0019] The intelligent judgment and control module performs adaptive material flow rate regulation calculations to generate the feeder speed adjustment amount. The calculation formula for the feeder speed adjustment amount includes:

[0020]

[0021] Where ΔRPM is the speed adjustment of the feeding equipment; K p The proportional coefficient is obtained through the system self-tuning process; K i The integral coefficients are obtained through the system self-tuning process; Q arget A target flow rate value is set and obtained from the human-computer interaction terminal; Q actual,i The i-th actual flow rate sample value is obtained from the material flow rate data output by the online metering subsystem; N t The sampling window size is determined based on the system response time; S j The real-time value of the j-th quality parameter is obtained from the state parameters output by the state parameter identification unit; S threshold,j The threshold value for the j-th quality parameter is obtained from the human-computer interaction terminal; M is the total number of quality parameters; t0 is the control start time; and t is the current time.

[0022] In some implementations, the formula for calculating the proportionality coefficient includes:

[0023]

[0024] Where E is the sum of squared system control errors, obtained through statistical analysis of historical data; V k The k-th device parameter affecting the control effect is retrieved from the device configuration file storage module; P is the total number of device parameters; ∈ is a minimal constant to prevent the denominator from being zero; W l The weight of the l-th environmental factor is set by expert experience; ΔT l The change in the l-th environmental factor is obtained by monitoring the environmental sensor network; R is the total number of environmental factors; D m is the overshoot of the m-th regulation, extracted from historical regulation records; C is the statistical value of the number of regulation times.

[0025] At least one embodiment of the present invention integrates online metering, status observation, intelligent data processing and human-machine interaction functions, and constructs a closed-loop control system. This enables synchronous, real-time, and accurate monitoring and adaptive control of material flow and key quality parameters of cottonseed and gleaming, which have significantly different characteristics. It fundamentally solves the core problems of metering distortion, quality detection lag and production control disconnect in traditional methods, and significantly improves the continuity and automation level of cottonseed processing, as well as the quality and economic benefits of the final product. Attached Figure Description

[0026] Figure 1 This is a structural block diagram of an online measurement and real-time observation system for raw cottonseed and gleaming cottonseed provided by the present invention. Detailed Implementation

[0027] Many specific details are set forth in the following description to provide a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0028] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the one or more embodiments of this specification. The singular forms “a” and “the” as used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items. The modifications “a” and “a plurality” as used in this disclosure are illustrative and not restrictive, and those skilled in the art will understand that they should be understood as “one or more” unless the context clearly indicates otherwise.

[0029] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this specification, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0030] See Figure 1 , Figure 1A structural block diagram of an online metering and real-time monitoring system for raw and glean cottonseed, according to some embodiments of this specification, is shown. The system includes: an online metering subsystem for measuring and outputting material flow data for raw and glean cottonseed; an monitoring subsystem for online detection and outputting material status parameters; an intelligent data processing center for receiving, fusing, and processing the material flow data and material status parameters to generate control commands; and a human-machine interface terminal for parameter setting and status display. The online metering subsystem, monitoring subsystem, intelligent data processing center, and human-machine interface terminal are connected via an industrial network to form a closed-loop control system.

[0031] An online metering subsystem can refer to a combination of hardware and software used for continuous measurement of material flow rate. For example, it collects data through dedicated sensors and metering devices to provide instantaneous and cumulative material flow rate data for raw and gleaming cottonseed. Material flow rate data can refer to a time-series dataset reflecting changes in material delivery volume, such as data measured and output in real time by sensors within the metering subsystem, used to characterize the material delivery rate and cumulative throughput.

[0032] The observation subsystem can refer to a combination of devices used for real-time detection of the apparent properties of materials. For example, it can employ optical imaging and image analysis technologies to identify and output the surface state and quality parameters of materials online. Material state parameters can refer to a set of quantitative indicators describing the physical properties of materials. For example, they can be extracted from acquired images using image processing algorithms to characterize the short fiber coverage, grain breakage, and impurity content of materials.

[0033] An intelligent data processing center can refer to a core processing unit that performs data fusion and decision-making calculations. For example, it receives traffic data and status parameters, performs fusion analysis, and uses this data to generate control commands and achieve intelligent system regulation. Control commands can refer to equipment adjustment commands output by the intelligent center, such as those generated based on the comparison results of real-time data and preset thresholds, used to adjust the operating parameters of upstream equipment to maintain system stability.

[0034] Human-machine interface (HMI) refers to the user interface through which information is exchanged between the user and the system. For example, a touchscreen can be used for parameter setting and status visualization, providing access for manual intervention and real-time operational status display. Industrial network refers to the communication infrastructure connecting industrial equipment. For example, protocols such as PROFINET or EtherCAT can be used to connect various subsystems, enabling data exchange and closed-loop control between components. Closed-loop control refers to a system operating mode that automatically adjusts based on feedback signals. For example, it calculates control variables based on the deviation between real-time data and setpoints, allowing the system output to automatically approach a preset target.

[0035] The present invention will be further described below through a detailed embodiment:

[0036] An online metering and real-time monitoring system for raw and gleaming cottonseed is deployed on a cottonseed processing line with an annual output of 50,000 tons to continuously monitor and control the materials before and after delinting of raw cottonseed.

[0037] The online metering subsystem employs a parallel architecture design, processing two materials simultaneously. The raw cottonseed metering unit integrates an anti-adhesion belt scale with a self-cleaning mechanism. Its core consists of an 800mm wide conveyor belt coated with ultra-high molecular weight polyethylene and a cleaner made of tungsten carbide alloy scrapers. This cleaner performs a reciprocating motion every 90 seconds, effectively scraping away short fibers and impurities adhering to the belt surface. The unit's built-in high-precision weight sensor (range 0-500 kg, accuracy 0.5% FS) and digital speed sensor (range 0-10 m / min) acquire load and speed signals in real time. Through a dynamic weighing algorithm embedded in the local PLC (Programmable Logic Controller), it calculates and outputs the instantaneous flow rate (tons / hour) and cumulative yield (tons) of raw cottonseed in real time. The glossy cottonseed metering unit integrates impact-resistant pipe weighing and flow rate measurement. Its core is a 300mm diameter 304 stainless steel weighing pipe with embedded silicon nitride ceramic baffles to resist material erosion. An S-beam load cell (range 0-200 kg, accuracy 0.1% FS) located below the pipe directly measures the net weight of the material inside the pipe. Simultaneously, an ultrasonic flow velocity sensor (range 0.5-5 m / s) mounted above the pipe, based on the Doppler effect, measures the material flow velocity non-contactly. These two signals are fed into the unit's local signal processor, where they are fused using a product-integral model (flow rate = cross-sectional area velocity density) to ultimately output the instantaneous flow rate and cumulative yield of the cottonseed.

[0038] The observation subsystem is responsible for online inspection of the material's appearance quality. Its image acquisition unit, set up 1.5 meters behind the metering point, includes two 5-megapixel industrial CMOS (Complementary Metal-Oxide-Semiconductor) cameras, one aimed at the material flow channels for raw cottonseed and the other at smooth cottonseed. The cameras are equipped with near-infrared LED (Light Emitting Diode) light sources with a wavelength of 850 nanometers, capturing high-resolution images of the material surface at a rate of 25 frames per second. The captured image data is transmitted in real-time via gigabit Ethernet to the status parameter recognition unit, an industrial computer equipped with a GPU (Graphics Processing Unit). The computer runs an image recognition model based on deep learning algorithms. This model performs pixel-level analysis on each input frame, automatically identifying and calculating three key quality parameters: short fiber coverage (%), breakage rate (%), and impurity content (%), and outputs these results in real-time.

[0039] The intelligent data processing center, acting as the system's brain, is a high-performance server. Its data receiving and preprocessing module periodically (every 200 milliseconds) reads all raw data from the local PLCs of the two metering units and the industrial computer of the observation subsystem via the Modbus TCP / IP industrial protocol. For sensor data from the metering subsystem, this module first performs temperature compensation and zero-point drift correction algorithms to eliminate the impact of environmental temperature drift on weighing accuracy. The cleaned and standardized data is then sent to the intelligent judgment and control module. This module compares the real-time collected material flow rate (including cottonseed and bleached cottonseed) with the preset target flow rate value from the human-machine interface terminal, and simultaneously compares real-time quality parameters (short fiber coverage, breakage rate, and impurity content) with preset quality thresholds. If any parameter deviates from the set range for more than 5 seconds, the module generates a corresponding control command. For example, when the instantaneous flow rate of bleached cottonseed is lower than the set value, an instruction to increase the speed of the upstream feeder will be generated; when the short fiber coverage of bleached cottonseed exceeds the standard, an instruction to adjust the operating parameters of the delinting machine will be generated and an audible and visual alarm will be triggered. All these process data, alarm events, and control instructions are recorded in the central SQL (Structured Query Language) database by the data storage and reporting module, and can automatically generate production reports and quality trend charts for each shift and each day.

[0040] The human-machine interface terminal is a 15-inch industrial touchscreen installed in the workshop's central control room. Operators can easily set target flow rates for raw and bleached cotton seeds, upper and lower limit alarm thresholds for various quality parameters, and equipment start / stop functions through its graphical interface. The terminal's main screen displays real-time flow rate and cumulative production data from the metering subsystem, as well as real-time values ​​and historical curves of short fiber coverage, breakage rate, and impurity content from the observation subsystem, providing a clear overview of the production status.

[0041] Finally, all the aforementioned subsystems and terminals are connected through a single industrial gigabit Ethernet network. Real-time flow data generated by the online metering subsystem, real-time quality parameters generated by the observation subsystem, control commands generated by the intelligent data processing center, and parameters set by the human-machine interface terminal are all transmitted and interacted on this network in an orderly manner. This constitutes a fully automated closed-loop control system from measurement, analysis, decision-making to execution, ensuring that the cottonseed processing process remains continuously stable in its optimal state.

[0042] The beneficial effects of one of the embodiments in this specification include at least the following: by integrating online metering, status observation, intelligent data processing and human-machine interaction functions, and constructing a closed-loop control system, synchronous, real-time, and accurate monitoring and adaptive control of material flow and key quality parameters of cottonseed and gleaming with significantly different characteristics are achieved. This fundamentally solves the core problems of metering distortion, quality detection lag and production control disconnect in traditional methods, and significantly improves the continuity and automation level of the cottonseed processing process, as well as the quality and economic benefits of the final product.

[0043] In some implementations, the online metering subsystem includes a raw cottonseed metering unit and a polished cottonseed metering unit. The raw cottonseed metering unit adopts an integrated design of an anti-adhesion belt scale and a self-cleaning mechanism to collect the instantaneous flow rate and cumulative yield of raw cottonseed. The polished cottonseed metering unit adopts an integrated design of impact-resistant pipeline weighing and flow rate measurement to collect the instantaneous flow rate and cumulative yield of polished cottonseed. The raw cottonseed metering unit and the polished cottonseed metering unit are set up in parallel to process two materials with different characteristics simultaneously.

[0044] A cottonseed weighing unit can refer to a dedicated device for continuous weighing of cottonseed, such as an integrated anti-adhesion belt scale and a self-cleaning mechanism, used to collect instantaneous flow and cumulative yield data of cottonseed. An anti-adhesion belt scale can refer to a belt weighing device that uses special anti-adhesion materials, such as ultra-high molecular weight polyethylene coated conveyor belts, to prevent short cottonseed fibers from sticking together and ensure weighing accuracy. A self-cleaning mechanism can refer to a mechanical device that automatically removes residues, such as using tungsten carbide alloy scrapers for periodic cleaning, to maintain the cleanliness of the belt scale's working surface.

[0045] A cottonseed metering unit can refer to a dedicated device for measuring cottonseed flow, such as an integrated design of impact-resistant pipeline weighing and flow velocity measurement, used to collect the instantaneous flow rate and cumulative yield of cottonseed. Impact-resistant pipeline weighing refers to pipeline weighing technology resistant to material impact, achieved through reinforced sensors and buffer structures, used to directly measure the net weight of the material flowing within the pipeline. An integrated flow velocity measurement design refers to a combined structure of weighing and flow velocity detection, such as integrating ultrasonic Doppler sensors into the pipeline section to simultaneously acquire material flow velocity and weight data. Instantaneous flow rate refers to the real-time flow rate of material passing through per unit time, calculated by fusing data from weight and velocity sensors, used to reflect the instantaneous state of material transport. Cumulative yield refers to the total amount of material passing through over a period of time, obtained by integrating the instantaneous flow rate over time, used for statistical production volume and efficiency evaluation. Materials with differentiated characteristics refer to material types with significantly different physical properties; for example, the difference between woolly cottonseed containing short fibers and prone to clumping and cottonseed with good flowability requires different metering methods. Parallel setup can refer to a layout in which two metering units work independently at the same time. For example, after the materials are diverted, they enter two metering channels respectively, which is used to achieve synchronous metering and processing of two types of materials.

[0046] As a concrete example: On a cottonseed processing production line, raw cottonseed is continuously weighed using an anti-adhesion belt scale. Its ultra-high molecular weight polyethylene coated conveyor belt effectively prevents short fibers from sticking together, and the tungsten carbide alloy scraper cleaner performs a self-cleaning cycle every 5 minutes. Meanwhile, glossy cottonseed is weighed through an impact-resistant pipeline weighing unit, whose internal silicon nitride ceramic guide plate guides the material flow. The weighing module directly measures the net weight, and the ultrasonic Doppler flow velocity sensor measures the flow rate in real time. The instantaneous flow rates of the two materials (e.g., 12.5 tons / hour for raw cottonseed and 10.8 tons / hour for glossy cottonseed) are uploaded to the monitoring system in parallel via the Modbus TCP protocol. The cumulative production data is recorded every 15 minutes and stored in an SQL database, achieving synchronous and accurate weighing of materials with different characteristics.

[0047] By designing dedicated metering units for different material characteristics and setting them in parallel, the metering problems caused by the easy sticking of wool and the high impact force of smooth cotton are effectively solved, improving the system's adaptability to differentiated materials and metering accuracy, and ensuring the accuracy and real-time nature of production data.

[0048] In some embodiments, the cottonseed metering unit includes an ultra-high molecular weight polyethylene coated conveyor belt, a tungsten carbide scraper cleaner, a weight sensor, and a speed sensor. The tungsten carbide scraper cleaner performs periodic self-cleaning motions to remove residual short fibers. Data from the weight sensor and speed sensor are used to calculate cottonseed flow data using a dynamic weighing algorithm located locally in the cottonseed metering unit.

[0049] Ultra-high molecular weight polyethylene (UHMWPE) coated conveyor belts refer to conveyor belts with a surface coated with special polymer materials, such as polyethylene materials with a molecular weight exceeding one million, used to reduce the adhesion of short cottonseed lint. Tungsten carbide alloy scraper cleaners refer to cleaning devices using hard alloy materials, such as alloy scraper structures with tungsten carbide as the main component, used to effectively scrape away residual materials on the conveyor belt surface. Periodic self-cleaning motion refers to cleaning actions performed at fixed time intervals, such as a PLC controller periodically triggering a cylinder to push the scraper, used to keep the metering unit in a continuously clean working state. Residual short lint refers to fibrous material left after cottonseed processing, such as short fiber impurities adhering to the conveyor belt surface, which affects metering accuracy and needs to be removed periodically. Weight sensors refer to sensing devices that detect the weight of materials, such as resistance strain gauge load cells arranged below the weighing frame, used to collect the weight signal of materials on the conveyor belt in real time. Speed ​​sensors refer to sensing devices that measure the speed of the conveyor belt, such as incremental encoders that detect the speed of the drive rollers, used to obtain real-time operating speed data of the conveyor belt. Dynamic weighing algorithms refer to methods for calculating weight data under motion conditions. For example, they employ digital filtering and real-time integration to process sensor signals, enabling the calculation of accurate cottonseed flow rate data. Cottonseed flow rate data refers to the flow rate information during the cottonseed transport process, such as data obtained through the fusion of weight and speed sensor data, used to characterize the instantaneous flow rate and cumulative yield of cottonseed.

[0050] As a concrete example: In the cottonseed metering unit, an ultra-high molecular weight polyethylene coated conveyor belt runs at a speed of 0.5 m / s. Weight sensors collect weight signals in real time (range 0-200 kg), and speed sensors detect belt speed via an encoder (accuracy ±0.1%). A tungsten carbide scraper cleaner performs a self-cleaning motion every 10 minutes (controlled by a 24VDC signal output from a PLC to activate a cylinder). Sensor data is transmitted to the local processing unit via a 4-20mA analog signal. A dynamic weighing algorithm is used for temperature compensation and filtering, ultimately outputting the instantaneous cottonseed flow rate (e.g., 12.8 units) and cumulative yield data, which are then uploaded to the central monitoring system via the PROFIBUS-DP protocol.

[0051] By combining ultra-high molecular weight polyethylene coated conveyor belts with tungsten carbide alloy scraper cleaners, the problem of short cotton lint adhesion is effectively solved. Combined with dynamic weighing algorithm to process sensor data, the accuracy and reliability of cotton lint measurement are significantly improved, ensuring the real-time and continuous nature of production data.

[0052] In some embodiments, the light cottonseed metering unit includes a pipe weighing module, an ultrasonic Doppler flow velocity sensor, and a silicon nitride ceramic guide plate. The pipe weighing module is designed to withstand impact and is used to directly measure the net weight of the material in the pipe. The ultrasonic Doppler flow velocity sensor measures the material flow rate in a non-contact manner. The data from the pipe weighing module and the ultrasonic Doppler flow velocity sensor are used to calculate the light cottonseed flow rate data through a product integral model located locally in the light cottonseed metering unit.

[0053] Pipeline weighing modules refer to weighing devices installed on material conveying pipelines, such as those employing cantilever beam sensors and buffer structures, used to directly measure the net weight of materials flowing within the pipeline. Impact-resistant design refers to structural designs that enhance the equipment's resistance to material impacts, such as by adding buffer springs and dampers, to protect the weighing sensor from the impact of material flow. Ultrasonic Doppler velocity sensors refer to devices that measure flow velocity based on the Doppler effect, such as those that measure the flow velocity of materials within a pipeline by transmitting and receiving changes in the frequency of ultrasonic signals. Silicon nitride ceramic guide plates refer to flow guiding devices made of silicon nitride ceramic material, such as those installed at specific angles inside the pipeline, used to guide the material flow smoothly through the measurement area and avoid turbulence. Non-contact measurement refers to measurement methods that do not directly contact the material being measured, such as utilizing the propagation characteristics of ultrasound in air, to avoid interfering with the material flow state. Material flow velocity refers to the distance material travels per unit time, such as by calculating changes in ultrasonic signals using Doppler frequency shift, used to characterize the flow velocity state of materials within a pipeline. The product-integral model can refer to the mathematical processing method for flow rate calculation. For example, multiplying instantaneous weight and flow velocity data and then integrating the results can be used to calculate accurate cottonseed flow rate data. Cottonseed flow rate data can refer to the flow rate information of cottonseeds flowing in a pipeline. For example, it can be calculated by fusing weighing and flow velocity measurement data to output the instantaneous flow rate and cumulative yield of cottonseeds.

[0054] As a concrete example: In the cottonseed metering unit, a silicon nitride ceramic baffle is installed at a 45-degree angle inside a DN200 pipe. The pipe weighing module uses four cantilever beam load cells (range 0-500kg, accuracy 0.05%). An ultrasonic Doppler flow velocity sensor transmits a signal at a frequency of 1MHz, and the flow velocity is calculated by detecting the echo frequency offset. Weighing data is transmitted via the HART protocol, and flow velocity data is output via a 4-20mA analog signal. The local processor uses a product integral model to calculate the instantaneous flow rate (e.g., 15.2 units) with a 100ms sampling period, and integrates over time to obtain the cumulative yield. The data is uploaded to the monitoring system via the Modbus TCP protocol.

[0055] By combining an impact-resistant pipe weighing module with an ultrasonic Doppler flow sensor, and leveraging the flow-stabilizing effect of a silicon nitride ceramic guide plate, accurate measurement of cottonseed flow is achieved. The non-contact measurement method avoids interference with material flow, and the product integral model ensures the accuracy and reliability of flow data, providing reliable data support for the production process.

[0056] In some implementations, the observation subsystem includes an image acquisition unit and a state parameter identification unit. The image acquisition unit uses an industrial CMOS camera and a specific light source to capture images of the material surface. The state parameter identification unit uses a deep learning algorithm model to process the images, identify and output parameters such as short fiber coverage, breakage rate and impurity content.

[0057] An image acquisition unit can refer to a hardware device used to capture images of a material surface. For example, it may consist of an industrial CMOS camera and a specific light source to acquire high-quality image data of the material surface. An industrial CMOS camera can refer to an industrial camera based on complementary metal-oxide-semiconductor (CMOS) technology, such as one employing a global shutter and gigabit network interface, for high-speed acquisition of clear images of the material surface. A specific light source can refer to an illumination system specifically configured for image acquisition, such as a high color rendering index (CRI) LED strip light source arranged at a specific angle to eliminate shadows and highlight material surface features. A material surface image can refer to a digital image reflecting the apparent state of the material, such as one captured by a camera under specific lighting conditions, for subsequent image processing and analysis.

[0058] A state parameter recognition unit can refer to a software module that processes images and extracts feature parameters. For example, it can use deep learning algorithms to analyze images to identify and output the quality parameters of materials. A deep learning algorithm model can refer to a feature recognition model based on neural networks, such as using a convolutional neural network structure for training, to automatically extract useful feature information from images. Short fiber coverage can refer to a quantitative indicator of the degree of short fiber coverage on the material surface. For example, it can be calculated using image segmentation algorithms to determine the coverage area ratio, used to assess the processing quality of cottonseed. Damage rate can refer to the proportion of broken grains in a material. For example, it can be used to count the number of broken particles using classification algorithms, used to characterize the integrity of the material. Impurity content rate can refer to the proportion of impurity components in a material. For example, it can be used to identify non-grain components through image analysis, used to assess the purity level of the material.

[0059] As a concrete example: The image acquisition unit uses a 2-megapixel industrial CMOS camera (IMX264 sensor) with a 4500K color temperature LED ring light source to acquire images of the material surface at a rate of 25 frames per second, which are then transmitted to the state parameter recognition unit via the GigE Vision protocol. This unit uses an improved U-Net semantic segmentation model to calculate the short fiber coverage rate and a ResNet-50 architecture classification model to identify damaged grains and impurities. Finally, it outputs parameters such as short fiber coverage rate (e.g., 85.2%), damage rate (e.g., 2.1%), and impurity content (e.g., 1.8%), which are then uploaded to the intelligent data processing center via the OPC UA protocol.

[0060] By acquiring high-quality images using an industrial CMOS camera and a specific light source, and combining this with a deep learning algorithm model to automatically identify material state parameters, accurate quantitative assessments of short fiber coverage, breakage rate, and impurity content have been achieved, providing reliable detection methods and data support for product quality control.

[0061] In some implementations, the state parameter identification unit uses an improved U-Net semantic segmentation model to calculate the short fiber coverage rate; it uses a ResNet-50 architecture classification model to identify damaged grains and impurities, and calculates the damage rate and impurity content accordingly.

[0062] The improved U-Net semantic segmentation model can refer to a structurally optimized image segmentation neural network, such as improved by adding attention mechanisms and depthwise separable convolutions, for accurately segmenting short-fiber regions in images and calculating coverage. The ResNet-50 architecture classification model can refer to a classification model employing a 50-layer deep residual network, such as using skip connections to solve the gradient vanishing problem, for accurately identifying broken cottonseeds and impurity particles. Broken cottonseeds can refer to cottonseed particles that break during processing, for example, identified by extracting texture and shape features using convolutional neural networks, for statistically assessing breakage rates and processing quality. Impurities can refer to non-cottonseed components mixed into materials, for example, identifying discolored and irregularly shaped particles using deep learning models, for calculating impurity content and assessing material purity.

[0063] As a concrete example: the state parameter recognition unit employs an improved U-Net semantic segmentation model. Its encoder uses a VGG16 backbone, and the decoder incorporates an attention gate mechanism. It performs short-fiber region segmentation on the input 512x512 pixel image, outputting a pixel-level mask and calculating the short-fiber coverage rate. Simultaneously, it uses a ResNet-50 model pre-trained on ImageNet, fine-tuning the last fully connected layer through transfer learning to classify and recognize the cropped seed images, accurately distinguishing between intact seeds, broken seeds, and impurities. Finally, it outputs the breakage rate and impurity content parameters. All inference processes are accelerated using TensorRT, and real-time recognition services are provided via the gRPC interface.

[0064] The improved U-Net model achieves accurate segmentation of short-fiber regions, and the ResNet-50 deep network accurately identifies damaged grains and impurities, improving the accuracy and reliability of state parameter identification and providing a scientific basis for product quality assessment. At the same time, model optimization ensures real-time processing performance.

[0065] In some implementations, the intelligent data processing center includes a data receiving and preprocessing module, an intelligent judgment and control module, and a data storage and reporting module. The data receiving and preprocessing module is used to perform temperature compensation and zero-point drift correction on sensor data from the online metering subsystem. The intelligent judgment and control module is used to compare real-time data with preset thresholds and generate control commands. The data storage and reporting module is used to store all process data and generate visual reports.

[0066] The data receiving and preprocessing module refers to the software unit responsible for data acquisition and preliminary processing. For example, it receives sensor data via the OPC UA protocol and performs format conversion and outlier filtering on the raw data. Temperature compensation refers to signal processing techniques that eliminate the influence of temperature, such as using a PT100 temperature sensor to monitor ambient temperature and correct measurement errors caused by temperature changes. Zero-point drift correction refers to calibration methods that eliminate sensor zero-point offset, such as recording the offset through a periodic automatic zeroing procedure to ensure the baseline accuracy of the measurement data. The intelligent judgment and control module refers to the software unit that performs logical judgments and control decisions, such as using rule engines and fuzzy logic algorithms to generate equipment control commands based on real-time data. The data storage and reporting module refers to the software unit responsible for data persistence and report generation, such as using a time-series database to store process data for generating production reports and historical data queries. Visualized reports refer to data reports displayed graphically, such as displaying trend curves and statistical charts through a web interface to intuitively present production status and quality analysis results.

[0067] As a concrete example: the data receiving and preprocessing module acquires raw data from the weight and speed sensors every second via the OPC UA protocol, monitors the ambient temperature using a PT100 temperature sensor, performs temperature compensation using a polynomial fitting algorithm, and performs automatic zero-point calibration every 30 minutes. The intelligent judgment and control module compares real-time flow data with preset thresholds (e.g., a target flow rate of 12 tons / hour), and sends adjustment commands to the feeding equipment via the PROFINET protocol when the deviation exceeds 5%. The data storage and reporting module uses the InfluxDB time-series database to store all process data and uses Grafana to generate visual reports that include flow trend graphs and quality parameter statistics.

[0068] By employing professional data preprocessing techniques to ensure the accuracy of measurement data, utilizing intelligent judgment algorithms to achieve precise process control, and combining data storage and visualization reporting functions, the system provides complete data support and decision-making basis for production process monitoring and quality analysis, thereby improving the system's automation level and operational reliability.

[0069] In some implementations, the intelligent judgment and control module generates instructions to adjust the speed of the upstream feeding equipment when the material flow rate deviates from the set value; and triggers an audible and visual alarm and generates instructions to adjust the operating parameters of the cleaning equipment when the quality parameters exceed the standard.

[0070] Material flow rate deviation from set value refers to the state where the actual flow rate deviates from the target value. For example, comparing the absolute difference between the real-time flow rate and the set value can trigger a flow regulation control mechanism. Upstream feeding equipment speed refers to the operating speed of the feeding equipment located at the beginning of the process, such as the motor speed parameter controlled by a frequency converter, used to adjust the material supply to maintain a stable flow rate. Quality parameter exceeding standard refers to the state where material quality indicators exceed the allowable range. For example, when the impurity content or breakage rate exceeds the threshold, it can trigger a quality anomaly handling process. Audible and visual alarm refers to an alarm method that uses both sound and light signals, such as a combination of a buzzer and LED indicator lights, used to promptly alert operators to abnormal conditions. Cleaning equipment operating parameters refer to the working status settings of the material cleaning machinery, such as adjusting the screen vibration frequency and air volume, used to optimize the cleaning effect and improve material quality.

[0071] As a specific example: when the intelligent judgment and control module detects that the instantaneous cottonseed flow rate (11.8 tons / hour) deviates from the set value (12.0 tons / hour) by more than the allowable deviation (±0.3 tons / hour), it sends a speed adjustment command to the frequency converter of the feeding equipment via the PROFIBUS-DP protocol, adjusting the motor speed from 1450 RPM to 1480 RPM. Simultaneously, when the detected impurity content (2.3%) exceeds the threshold (2.0%), the audible and visual alarm device (24VDC buzzer and red warning light) is triggered, and the frequency of the cleaning equipment's vibration motor is increased from 50Hz to 55Hz, and the fan speed is increased from 980 RPM to 1050 RPM, via the Modbus TCP protocol.

[0072] By monitoring changes in material flow and quality parameters in real time, the operating status of upstream equipment can be adjusted in a timely manner, effectively maintaining the stability of the production process and the consistency of product quality. At the same time, the audible and visual alarm mechanism ensures that abnormal situations are handled in a timely manner, thereby improving the system's level of automation control and production reliability.

[0073] In some implementations, the online metering and real-time monitoring system for raw and smooth cottonseed also includes an environmental sensor network and a device configuration file storage module. The environmental sensor network monitors changes in environmental factors such as temperature and humidity, while the device configuration file storage module stores equipment parameters that affect the control effect. The intelligent judgment and control module performs adaptive material flow control calculations to generate the feeder speed adjustment amount. The formula for calculating the feeder speed adjustment amount includes:

[0074]

[0075] Where ΔRPM is the speed adjustment of the feeding equipment; K p The proportional coefficient is obtained through the system self-tuning process; K i The integral coefficients are obtained through the system self-tuning process; Q arget A target flow rate value is set and obtained from the human-computer interaction terminal; Q actual,i The i-th actual flow rate sample value is obtained from the material flow rate data output by the online metering subsystem; N t The sampling window size is determined based on the system response time; S j The real-time value of the j-th quality parameter is obtained from the state parameters output by the state parameter identification unit; S threshold,j The threshold value for the j-th quality parameter is obtained from the human-computer interaction terminal; M is the total number of quality parameters; t0 is the control start time; and t is the current time.

[0076] Environmental sensor networks refer to distributed sensing systems that monitor environmental parameters. For example, a network composed of temperature, humidity, and barometric pressure sensors can be used to collect real-time status data of the production environment. Temperature environmental factor changes refer to the degree of change in ambient temperature relative to a reference value. For example, data collected using a PT100 temperature sensor can be used to assess the impact of temperature changes on measurement accuracy. Humidity environmental factor changes refer to the degree of change in ambient humidity relative to a reference value. For example, a capacitive humidity sensor can be used to monitor humidity and analyze its impact on material flow characteristics. Equipment configuration file storage modules refer to system modules that store equipment parameters. For example, non-volatile memory can be used to store equipment characteristic data, providing the configuration parameters required for equipment operation. Equipment parameters affecting control effectiveness refer to equipment characteristic parameters that affect control performance, such as motor response time and sensor accuracy parameters, used to optimize the adjustment effect of control algorithms. Adaptive material flow control calculations refer to control algorithms that automatically adjust based on real-time data. For example, an adaptive PID control algorithm can be used to achieve precise and stable flow control. Feeding equipment speed adjustment refers to the change in the speed of the feeding equipment that needs to be adjusted, such as the speed increment calculated by the control algorithm, used to precisely adjust the material supply. The system self-tuning process refers to the process by which the control system automatically adjusts its parameters, such as using the Ziegler-Nichols method to tune parameters and optimize the controller's response characteristics. The proportional gain parameter in the control algorithm, such as a control parameter obtained through system identification, is used to adjust the system's response strength to deviations. The integral gain parameter in the control algorithm, such as the optimal parameter determined through experimental debugging, is used to eliminate the system's steady-state error. The target flow rate setpoint refers to the desired material flow rate target, such as a flow rate reference value set through a human-machine interface, used as the control system's adjustment target.

[0077] Actual flow rate sampling value refers to the real-time measured value of material flow rate, such as instantaneous flow data collected by sensors, used as the actual state input in feedback control. Sampling window size refers to the number of sampling points used for data processing, such as the number of samples contained in a sliding window, used for statistical analysis and filtering. System response time refers to the system's reaction speed to input changes, such as the time interval from input change to output stabilization, used to determine the adjustment cycle of control parameters. Real-time quality parameter value refers to the current value of the quality index, such as impurity content data obtained through online detection, used for quality monitoring and control decisions. Quality parameter threshold refers to the permissible limit of the quality index, such as the maximum impurity content set according to process requirements, used as a benchmark for determining whether quality is qualified. Control start time refers to the time point at which control adjustment begins, such as the time marker when the deviation exceeds the dead zone, used as the time benchmark for calculating integral action.

[0078] By monitoring environmental changes in real time through an environmental sensor network and combining this with equipment parameter configuration files, an adaptive control algorithm is used to achieve precise control of material flow. This improves the system's adaptability to environmental changes and control accuracy, ensuring the stability of the production process and the consistency of product quality.

[0079] In some implementations, the formula for calculating the scaling factor includes:

[0080]

[0081] Where E is the sum of squared system control errors, obtained through statistical analysis of historical data; V k The k-th device parameter affecting the control effect is retrieved from the device configuration file storage module; P is the total number of device parameters; ∈ is a minimal constant to prevent the denominator from being zero; W l The weight of the l-th environmental factor is set by expert experience; ΔT l The change in the l-th environmental factor is obtained by monitoring the environmental sensor network; R is the total number of environmental factors; D m is the overshoot of the m-th regulation, extracted from historical regulation records; C is the statistical value of the number of regulation times.

[0082] Multi-parameter optimization calculation refers to optimization algorithms that consider multiple variables simultaneously, such as using gradient descent to solve for optimal parameters, used to determine the optimal proportional gain of the control system. The sum of squares of system control errors refers to the cumulative square of the control deviations, such as those calculated from historical data, used to evaluate the overall performance of the control system. Equipment parameters affecting control effectiveness refer to parameters in equipment characteristics that influence control performance, such as sensor accuracy and actuator response time, used to optimize the adaptability of the control algorithm. Minimal constants refer to the smallest values ​​set to prevent division by zero errors, such as a tiny positive number added to the denominator, used to ensure the stability of mathematical calculations. Environmental factor weights refer to the importance coefficients of different environmental parameters, such as temperature and humidity influence coefficients set by expert experience, used to weight the impact of environmental factors. Environmental factor changes refer to the magnitude of change of environmental parameters relative to a baseline value, such as the difference between the current temperature and the standard temperature, used to quantify the degree of change in environmental conditions. Expert experience setting refers to methods for determining parameters based on professional knowledge and experience, such as having domain experts assess the importance of each factor to set reasonable weight coefficients. Historical control records refer to data records of past control operations, such as control parameters and results stored in a database, used for analysis, learning, and optimization of control strategies. Control overshoot refers to the overshoot magnitude in the control system response, such as the difference between the maximum deviation and the steady-state value, used to evaluate the dynamic performance of the control system. Control frequency statistics refer to the count of the number of control adjustments performed, such as the number of control actions within a certain time period, used for statistical analysis and performance evaluation.

[0083] As a specific example: The intelligent judgment and control module performs multi-parameter optimization calculations, queries the most recent 100 control records from the MySQL database, extracts overshoot data, obtains temperature change ΔT = 2.5℃ and humidity change ΔH = 5% through the environmental sensor network, combines the expert-set temperature weight of 0.6 and humidity weight of 0.4, uses the gradient descent algorithm to calculate the proportional coefficient Kp = 0.85, the integral coefficient Ki = 0.12, and sets the minimum constant ε to 1e-6 to prevent division by zero. The optimized parameters are updated to the real-time control system via the OPC UA protocol.

[0084] By comprehensively considering equipment characteristics, environmental factors, and historical control effects through multi-parameter optimization calculations, control parameters are dynamically adjusted, improving the system's adaptability and control accuracy, reducing overshoot, and ensuring the stability and control quality of the production process. At the same time, by combining expert experience with real-time data, the overall system performance is optimized.

[0085] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this invention. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. An online measurement and real-time observation system for raw cottonseed and gleaming cottonseed, characterized in that, include: The online metering subsystem is used to measure and output material flow data for raw cottonseed and gleaming cottonseed. The observation subsystem is used to detect and output material status parameters online; The intelligent data processing center is used to receive and integrate the material flow data and the material status parameters to generate control commands; Human-computer interaction terminal, used for parameter setting and status display; The online metering subsystem, the observation subsystem, the intelligent data processing center, and the human-machine interaction terminal are connected through an industrial network and form a closed-loop control system.

2. The system according to claim 1, characterized in that, The online metering subsystem includes a raw cottonseed metering unit and a polished cottonseed metering unit. The raw cottonseed metering unit adopts an integrated design of anti-adhesion belt scale and self-cleaning mechanism to collect the instantaneous flow rate and cumulative yield of raw cottonseed. The polished cottonseed metering unit adopts an integrated design of impact-resistant pipeline weighing and flow velocity measurement to collect the instantaneous flow rate and cumulative yield of polished cottonseed. The raw cottonseed metering unit and the polished cottonseed metering unit are set up in parallel to process two materials with different characteristics simultaneously.

3. The system according to claim 2, characterized in that, The cottonseed metering unit includes an ultra-high molecular weight polyethylene coated conveyor belt, a tungsten carbide alloy scraper cleaner, a weight sensor, and a speed sensor. The tungsten carbide alloy scraper cleaner performs periodic self-cleaning motion to remove residual short fibers. The data from the weight sensor and the speed sensor are calculated using a dynamic weighing algorithm located locally in the cottonseed metering unit to obtain cottonseed flow data.

4. The system according to claim 2, characterized in that, The light cottonseed metering unit includes a pipe weighing module, an ultrasonic Doppler flow velocity sensor, and a silicon nitride ceramic guide plate. The pipe weighing module is designed to withstand impact and is used to directly measure the net weight of the material in the pipe. The ultrasonic Doppler flow velocity sensor measures the material flow rate in a non-contact manner. The data from the pipe weighing module and the ultrasonic Doppler flow velocity sensor are calculated using a product integral model located locally in the light cottonseed metering unit to obtain the light cottonseed flow rate data.

5. The system according to claim 1, characterized in that, The observation subsystem includes an image acquisition unit and a state parameter identification unit. The image acquisition unit uses an industrial CMOS camera and a specific light source to capture images of the material surface. The state parameter identification unit uses a deep learning algorithm model to process the images, identify and output parameters such as short fiber coverage, damage rate and impurity content.

6. The system according to claim 5, characterized in that, The state parameter identification unit uses an improved U-Net semantic segmentation model to calculate the short fiber coverage rate; it uses a ResNet-50 architecture classification model to identify damaged grains and impurities, and calculates the damage rate and impurity content accordingly.

7. The system according to claim 1, characterized in that, The intelligent data processing center includes a data receiving and preprocessing module, an intelligent judgment and control module, and a data storage and reporting module. The data receiving and preprocessing module is used to perform temperature compensation and zero-point drift correction on sensor data from the online metering subsystem. The intelligent judgment and control module is used to compare real-time data with preset thresholds and generate control commands. The data storage and reporting module is used to store all process data and generate visual reports.

8. The system according to claim 7, characterized in that, The intelligent judgment and control module generates instructions to adjust the speed of the upstream feeding equipment when the material flow rate deviates from the set value; and triggers an audible and visual alarm and generates instructions to adjust the operating parameters of the cleaning equipment when the quality parameters exceed the standard.

9. The system according to claim 7, characterized in that, It also includes an environmental sensor network and a device configuration file storage module. The environmental sensor network is used to monitor changes in environmental factors such as temperature and humidity, and the device configuration file storage module is used to store device parameters that affect the control effect. The intelligent judgment and control module performs adaptive material flow rate regulation calculation to generate the feeder speed adjustment amount, wherein the calculation formula for the feeder speed adjustment amount includes: Where ΔRPM is the speed adjustment of the feeding equipment; K p The proportional coefficient is obtained through the system self-tuning process; K i The integral coefficients are obtained through the system self-tuning process; Q arget A target flow rate value is set and obtained from the human-computer interaction terminal; Q actual,i The i-th actual flow rate sample value is obtained from the material flow rate data output by the online metering subsystem; N t The sampling window size is determined based on the system response time; S j The real-time value of the j-th quality parameter is obtained from the state parameters output by the state parameter identification unit; S threshold,j The threshold value for the j-th quality parameter is obtained from the human-computer interaction terminal; M is the total number of quality parameters; t0 is the control start time; and t is the current time.

10. The system according to claim 9, characterized in that, The formula for calculating the proportionality coefficient includes: Where E is the sum of squared system control errors, obtained through statistical analysis of historical data; V k The k-th device parameter affecting the control effect is retrieved from the device configuration file storage module; P is the total number of device parameters; ∈ is a minimal constant to prevent the denominator from being zero; W l The weight of the l-th environmental factor is set by expert experience; ΔT l The change in the l-th environmental factor is obtained by monitoring the environmental sensor network; R is the total number of environmental factors; D m is the overshoot of the m-th regulation, extracted from historical regulation records; C is the statistical value of the number of regulation times.

Images