Iron phosphate preparation full-process control system with multi-source data fusion monitoring function

Abstract

The invention belongs to the technical field of iron phosphate preparation, and discloses an iron phosphate preparation full-process control system with multi-source data fusion monitoring, and the system is characterized in that a feeding control module uses a variable frequency screw pump to convey raw materials as required, and based on a feeding flow PID closed-loop control model, an electric flow control valve can accurately adjust the feeding flow; raw materials enter a stirring tank of the reaction control module, the stirring speed is accurately regulated and controlled through a servo-driven stirring motor, and a heating jacket controls the reaction temperature based on a predicted temperature correction model; after the raw material reaction is completed, a centrifugal separator of the separation control module adjusts the separation speed through frequency conversion, a separation liquid level sensor monitors the liquid level in real time, an automatic slag discharge valve controls slag discharge according to a threshold value, and iron phosphate solid and liquid are separated; and after the drying control module receives the separated wet materials, the hot air dryer accurately controls the temperature based on a drying temperature PID control model, and a temperature and humidity sensor feeds back data in real time.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of iron phosphate preparation, and particularly relates to an iron phosphate preparation full-process control system with multi-source data fusion monitoring. BACKGROUND

[0002] As a key precursor of lithium ion battery positive materials, the quality of iron phosphate directly affects the core performances such as energy density and cycle life of the battery, so that the preparation process has strict requirements on product quality and production efficiency. However, the traditional iron phosphate preparation control system has significant defects. In the system architecture level, each functional module runs relatively independently, lacks data interaction and coordination mechanism, and cannot form an efficient full-process control system.

[0003] In terms of data processing, the traditional system only collects a few key process parameters, which is difficult to fully reflect the complex production process, and lacks comprehensive analysis and deep fusion ability of multi-source data, which cannot provide effective support for precise control. In the process control link, when feeding, due to the lack of real-time working condition perception and dynamic adjustment mechanism, the feeding flow cannot be accurately adjusted according to the reaction requirements, which often leads to unbalanced raw material ratio; in the reaction process, the control of temperature and stirring rate depends on the simple preset program, which is difficult to cope with the complex reaction kinetics changes, resulting in insufficient and uneven reaction; in the separation and drying stage, it is difficult to accurately grasp the solid-liquid separation process and material drying state, and problems such as incomplete solid-liquid separation and product moisture exceeding the standard are prone to occur. These deficiencies make the traditional preparation system have large product quality fluctuation, low production efficiency, and high energy and material consumption, which is difficult to meet the growing demand for high-quality iron phosphate production. SUMMARY

[0004] The purpose of the present application is to provide an iron phosphate preparation full-process control system with multi-source data fusion monitoring to solve the problems raised in the background.

[0005] In order to achieve the above purpose, the present application provides the following technical scheme: an iron phosphate preparation full-process control system with multi-source data fusion monitoring, which comprises: The feeding control module is the starting module of the iron phosphate preparation, which stably supplies raw materials through the raw material storage tank, transports the raw materials on demand by using the variable frequency screw pump, and makes the electric flow control valve accurately adjust the feeding flow based on the feeding flow PID closed loop control model, so as to ensure accurate raw material input for the subsequent reaction. The model of the variable frequency screw pump is G30-1, the flow range is 0-50 L / min, and the accuracy of the electric flow control valve is ±0.5% FS. Reaction control module: After receiving the raw materials from the feed control module, the reaction control module uses a stirred tank as the reaction vessel. The stirring speed is precisely controlled by a servo-driven stirring motor. The reaction temperature is controlled based on a predictive temperature correction model using a heating jacket. This ensures that the raw materials are fully reacted. The stable operation of the reaction control module directly affects the reaction efficiency and product quality of the iron phosphate; Separation control module: After the reaction is complete, the separation control module is put into operation. The centrifugal separator adjusts the separation speed through frequency conversion. The liquid level sensor monitors the liquid level in real time. The automatic deslagging valve controls deslagging based on a threshold value. Efficient solid-liquid separation is achieved, separating the iron phosphate solids from the liquid. This prepares the material for the drying process. Drying control module: The wet material after separation enters the drying control module. The hot air dryer precisely controls the temperature based on a drying temperature PID control model. The temperature and humidity sensor provides real-time feedback data. The exhaust flow valve adjusts the exhaust speed. The moisture content of the material is removed to obtain the finished product. Precise control ensures that the product moisture meets the standards. The heating power of the hot air dryer is 10-50 kW. The model of the temperature and humidity sensor is, for example, SHT30. Multi-source data acquisition module: To achieve precise control of the entire process, the multi-source data acquisition module is distributed in each process module. It collects key data such as feed flow and reaction temperature in real time. The time synchronization module synchronizes the time sequence. The sampling frequency is self-adaptive. It provides comprehensive, accurate, and real-time data support for subsequent data processing. Data fusion processing module: Based on the collected data, the data fusion processing module plays a role. The data anomaly elimination module selects reliable data. The missing data correction module improves the data. The dynamic weighted fusion module fuses the data according to importance. The comprehensive data after deep processing provides a high-quality data foundation for intelligent prediction and control, improves data usability, and updates the weight every 100 samples. Intelligent prediction and fuzzy control module: With high-quality data after processing, the intelligent prediction and fuzzy control module starts working. The sliding time window prediction unit analyzes historical data to predict process trends. The sliding time window length is the last 30 minutes of historical data. The fuzzy control unit generates adjustment instructions based on prediction and current data, completing precise prediction and control of the entire process. Process execution control module: After receiving the adjustment instructions, the process execution control module connects each process module through PLC, precisely controls the operating parameters of devices such as flow valves and stirring motors, completes closed-loop control of the entire process, ensures that each process link operates according to optimal parameters, converts control instructions into actual production adjustments, and ensures stable system operation. Human-computer interaction and communication module: as the bridge between the system and the operator, the human-computer interaction and communication module displays process data, curves and alarm information in real time, adopts double-ring redundant communication structure to ensure stable communication, and uses EtherCAT protocol for double-ring redundant communication, supports remote monitoring and encrypted access, facilitates personnel operation and management, helps to discover and handle problems in time, and improves system operation and management efficiency.

[0006] In view of the dynamic change of material properties, flow, temperature and stirring rate with time during the preparation of iron phosphate, a sampling frequency self-adaptive adjustment strategy based on machine learning algorithm is first introduced, which can realize real-time sensing of process state fluctuation degree, dynamically optimize the sampling frequency of each link, significantly improve the real-time performance and control accuracy of key data capture, and reduce data redundancy and system response delay.

[0007] In the multi-source data fusion processing, a dynamic weight adjustment method based on sensor historical fluctuation, real-time abnormality elimination and online credibility correction is adopted, combined with a sliding time window prediction model, to output the process parameter data with the highest comprehensive credibility in real time; compared with the existing single weighting or simple filtering method, this fusion method can significantly improve the accuracy and response speed of process parameters under complex working conditions, and belongs to a non-obvious technical progress.

[0008] Innovatively, the LSTM prediction algorithm in the sliding time window is cooperatively integrated with the fuzzy control unit, the historical process data are deeply learned and compared with the current real-time data, the process trend change is predicted in advance, the optimal control instruction is dynamically generated, and the adaptability to complex reaction kinetics and material property fluctuation is significantly improved; the control system can realize closed-loop intelligent regulation and control of the whole process of feeding, reaction, separation and drying, and has intelligent optimization effect beyond the existing multi-stage preset control.

[0009] The double-ring redundant industrial Ethernet, 4G wireless and VPN remote encrypted link are deeply integrated to build a security protection system with millisecond-level link switching and AES high-strength encryption, which supports remote precise regulation and control and intelligent diagnosis; compared with the existing iron phosphate preparation control system which only supports single-link wired communication, the production continuity, remote control convenience and industrial data security level are significantly improved.

[0010] Preferably, the feeding control module comprises: (1) Raw material storage guarantee: the feeding control module comprises a raw material storage tank. The raw material storage tank is equipped with an intelligent monitoring device, which can monitor the raw material inventory in real time. Once the inventory is insufficient, the early warning mechanism is triggered immediately to supplement the raw material in time, thereby providing continuous and stable raw material supply for the whole iron phosphate preparation process and avoiding production interruption due to raw material shortage; (2) Precise flow control: The frequency conversion screw pump and the electric flow control valve cooperate with each other to realize the precise control of the raw material conveying flow. The frequency conversion screw pump flexibly adjusts the rotating speed according to the requirements of different process stages, and ensures that the raw material is conveyed to the reaction link at a suitable flow rate. The electric flow control valve is based on the feed flow PID closed-loop control model, driven by a stepping motor, continuously monitors the feed flow, and quickly adjusts the valve opening according to the deviation between the actual and set flow, to ensure the stability of the feed flow.

[0011] The expression of the feed flow PID closed-loop control model is: ; In the formula: is the feed flow corresponding to the current opening of the feed flow valve (L / min); is the target feed flow calculated by the intelligent prediction and control unit (L / min) is the feed flow deviation, defined as is the current actual feed flow (L / min), which is collected in real time by the feed flow sensor; 、 is the proportional, integral, and differential coefficients of the feed flow PID controller, and the engineering setting method of Kd is, for example, the critical proportional degree method.

[0012] This formula directly controls the opening of the feed electric valve to ensure that the actual feed flow is dynamically consistent with the predicted flow, which belongs to the conventional application in the process site and is fully disclosed.

[0013] Preferably, the reaction control module comprises: (1) Efficient stirring and mixing: The reaction control module uses a stirred tank as the reaction site, and is equipped with a servo-driven stirring motor to create good conditions for raw material reaction. The stirring motor can flexibly adjust the stirring speed according to the requirements of different reaction stages, automatically switch the stirring mode through a preset program, make the raw materials fully mixed during the reaction process, enhance the collision and contact between molecules, effectively improve the uniformity and efficiency of the reaction, and ensure that the reaction proceeds fully; (2) Intelligent temperature regulation: The heating jacket uses a predictive temperature correction model to comprehensively analyze reaction historical data, current reaction status, and environmental factors, and predict the temperature change trend in advance. When the reaction temperature fluctuation is detected, the system quickly responds to automatically adjust the heating power of the heating jacket, timely compensate for the hysteresis of temperature change, and quickly stabilize the reaction temperature at the set value, to provide an ideal temperature environment for the synthesis of iron phosphate and ensure the stable and reliable product quality.

[0014] The expression of the predictive temperature correction model is: ; In the formula: is the target temperature of the heating jacket (°C); Actual reaction temperature (°C) Temperature correction amount output by prediction module (exponential smoothing prediction based on historical temperature data).

[0015] This formula realizes synchronous correction of predicted temperature and actual temperature, dynamically adjusts heating power, belongs to intelligent temperature control technology path, and is fully disclosed.

[0016] Preferably, the separation control module comprises: (1) High-efficiency solid-liquid separation: The separation control module relies on a variable-frequency centrifugal separator, flexibly adjusts the separation speed according to the characteristics of the material after the reaction, realizes high-efficiency solid-liquid separation, and realizes real-time monitoring of the running state of the equipment through the built-in separation liquid level sensor of the centrifugal separator, thereby ensuring the safety and stability of the separation process. The centrifugal force generated by high-speed rotation quickly separates the iron phosphate solid from the liquid, improves the separation efficiency, and facilitates subsequent drying work; (2) Automatic residue discharge management: The separation liquid level sensor monitors the liquid level in the variable-frequency centrifugal separator in real time, and feeds back the data to the control system in real time. When the liquid level reaches the pre-set threshold, the automatic residue discharge valve rapidly acts through the pneumatic actuator, and timely discharges the separated liquid and solid residue. The system can flexibly control the residue discharge period according to the actual situation, avoid liquid overflow and solid blockage, ensure the continuous and stable operation of the separation process, and reduce material residue and waste.

[0017] Preferably, the drying control module comprises: (1) Precise temperature and humidity control: The drying control module comprises a hot air dryer and a temperature and humidity sensor, and combines a drying temperature PID control model to precisely adjust the drying temperature. The temperature and humidity sensor monitors the humidity of the material and the ambient temperature during the drying process in real time, and feeds back the data to the control system. When the actual humidity deviates from the set humidity, the system automatically adjusts the heating power of the hot air dryer according to the deviation, ensures uniform drying of the material at a suitable temperature, and improves the drying effect and quality; Drying temperature PID control model expression: ; In the formula: Current temperature set value of the dryer; Predicted or process preset temperature; ; Proportional adjustment coefficient; (2) Intelligent exhaust management: The exhaust flow valve automatically adjusts the exhaust speed according to the generation of moisture during the drying process. The system timely discharges the moisture generated during the drying process based on the data fed back by the temperature and humidity sensor, and maintains the stability of the drying environment. By precisely controlling the drying temperature and exhaust speed, the drying efficiency is effectively improved, the moisture content of the product is ensured to meet the standard, and the product quality is improved.

[0018] Preferably, the multi-source data acquisition module comprises: (1) Comprehensive data acquisition: The multi-source data acquisition module is distributed in each key link of the preparation process, covering modules such as feeding, reaction, separation, and drying, and real-time acquisition of various key process data such as feed flow, reaction temperature, and stirring speed. High-precision sensors are used to ensure the accuracy of the collected data and provide rich and comprehensive data sources for the system to achieve precise control; (2) Intelligent timing optimization: The multi-source data acquisition module includes a time synchronization module that processes the data collected by each module in a unified timing manner to ensure the time consistency of the data. Based on an adaptive adjustment model, the sampling frequency is intelligently adjusted according to the dynamic changes of the process. When the process is stable, the sampling frequency is appropriately reduced to reduce data processing; when the process changes dramatically, the sampling frequency is promptly increased to ensure the capture of changes in key process parameters.

[0019] Sampling frequency adaptive adjustment model expression: ; In the formula: Current sampling frequency (Hz); Basic sampling frequency (Hz); Fluctuation sensitivity coefficient; Fusion parameter change rate.

[0020] The sampling frequency is adjusted in real time according to the process fluctuations to meet the process stability requirements, and the formula is fully disclosed.

[0021] Preferably, the data fusion processing module comprises: (1) Abnormal data processing: The data fusion processing module includes an abnormal data elimination module that uses a pre-set threshold to strictly filter the collected data, real-time identify and eliminate abnormal data caused by sensor failure, external interference, etc., effectively ensuring the reliability of the data and avoiding misleading production decisions by false data; (2) Data deep fusion: The data fusion processing module also includes a missing data correction module and a dynamic weighted fusion module. The missing data correction module reasonably corrects missing data caused by various reasons to ensure the integrity of the data. Based on the historical error rate of the sensor, the dynamic weighted fusion module uses a multi-source data weighted fusion model to assign appropriate weights to each source data according to its importance and reliability in the process, and deeply fuses the processed data to generate more accurate and more accurate comprehensive data reflecting the actual state of the process.

[0022] Multi-source data weighted fusion model expression: ; In the formula: Fused process parameters; The Real-time sensor data; The Weight satisfies .

[0023] The formula discloses the data fusion path and dynamic weight adjustment logic, and meets the sufficient disclosure requirement of real-time data fusion.

[0024] Preferably, the intelligent prediction and fuzzy control module comprises: (1) Accurate trend prediction: the intelligent prediction unit is based on continuous process historical data, deeply mines the internal relationship and change law between process parameters, accurately predicts the future process change trend, and perceives the process trend in advance through learning and analysis of historical data, so that the system can make preparations in advance; (2) Flexible intelligent regulation and control: the fuzzy control unit combines the prediction results with the current real-time data, and generates accurate adjustment instructions by using the fuzzy control algorithm. In the face of complex and changeable process, the fuzzy control algorithm can simulate the thinking mode of human beings, quickly and flexibly adjust the control strategy according to different process states and prediction results, accurately regulate and control the key parameters such as feed flow, stirring speed, heating power, etc., and make the system stably run in the best process state.

[0025] Preferably, the process execution control module comprises: (1) Accurate instruction execution: the process execution control module takes PLC as the core control unit, connects with the devices of each process unit such as feeding, reaction, separation and drying through multiple communication protocols, quickly and accurately converts the adjustment instructions output by the intelligent prediction and fuzzy control module into the actual actions of the devices by virtue of the powerful output capacity and accurate control characteristics of PLC, accurately controls the running parameters of each device, realizes real-time regulation and optimization of the whole process, and realizes real-time regulation and optimization of the whole process; (2) Safe and stable operation: the system is built-in with perfect fault diagnosis and safety interlocking mechanism, which can monitor the running state of each device in real time. Once the device is detected to be abnormal or the control instruction is out of limit, the safety shutdown program is started immediately, and the operator is notified in time through sound and light alarm and other ways. At the same time, the PLC online programming and remote debugging function is supported, which can facilitate the operator to flexibly adjust the control strategy according to the actual production demand, and ensure the safe, stable and efficient operation of the production process.

[0026] Preferably, the man-machine interaction and communication module comprises: (1)Convenient operation interface: adopt industrial-grade touch screen, design simple and intuitive graphical operation interface, integrated process parameter real-time display, historical curve query, alarm information processing, equipment remote control and other functions. The operator can easily realize process parameter setting, production report generation and other operations through simple touch operation, greatly improving the convenience and management efficiency of operation, and facilitating real-time grasp of production status; (2) Stable and safe communication: adopt industrial Ethernet double ring redundant communication structure, with high reliability and stability, support fast switching of main and standby links, ensure uninterrupted communication. At the same time, support 4G wireless communication and VPN remote encrypted access, use AES encryption algorithm to encrypt communication data, ensure the safety and confidentiality of data in the transmission process, and management personnel can remotely monitor the system running state and adjust parameters no matter where they are, realize intelligent remote management; Main and standby link switching delay formula: ; In the formula: Total link switching delay; Main link abnormality detection time; Standby link recovery communication time; Comply with the tolerance of industrial Ethernet link switching, which is an open industrial standard and can be directly used in the application.

[0027] The beneficial effects of the application are as follows: 1. The application can accurately predict the process change trend, generate adjustment instructions in advance, and quickly respond and adjust the running parameters of each process unit through the multi-source data acquisition module to collect comprehensive process data in real time, the data fusion processing module to process the data, and the intelligent prediction and fuzzy control module.

[0028] 2. The application can accurately control the feed flow of raw materials based on the feed flow PID closed-loop control model of the feed control module, avoid waste of raw materials, and improve reaction efficiency and reduce energy consumption through the prediction temperature correction model and servo drive of the reaction control module. The separation control module and drying control module improve the solid-liquid separation effect and drying efficiency through precise control, reduce material loss and energy cost. At the same time, the intelligent control of the system reduces manual intervention and labor cost, thereby effectively reducing the preparation cost of ferric phosphate.

[0029] 3. This invention ensures that each process step operates under optimal conditions through precise control throughout the entire process. The reaction control module precisely controls the reaction temperature and stirring rate, ensuring the sufficiency and uniformity of the reaction; the drying control module strictly controls the drying temperature and humidity, ensuring that the product's moisture content meets standards; the intelligent prediction and fuzzy control module adjusts process parameters in a timely manner, effectively avoiding product quality problems caused by process fluctuations; through the fusion analysis and intelligent control of multi-source data, product quality can be stabilized, product consistency and pass rate can be improved, and the company's market competitiveness can be enhanced. Attached Figure Description

[0030] Figure 1 This is a flowchart of the whole-process control system for the preparation of iron phosphate with multi-source data fusion monitoring, as described in this invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] like Figure 1 As shown, this embodiment of the invention provides a control system for the entire process of iron phosphate preparation with multi-source data fusion monitoring. The system includes: Feed control module: As the starting module for iron phosphate preparation, the feed control module provides stable feed through the raw material storage tank and delivers raw materials on demand using a variable frequency screw pump. Based on the feed flow rate PID closed-loop control model, the electric flow control valve can accurately adjust the feed flow rate to ensure accurate raw material input for subsequent reactions. Reaction control module: After receiving the raw materials from the feed control module, the reaction control module uses the stirred tank as the reaction vessel and precisely controls the stirring rate through the servo-driven stirring motor. Combined with the heating jacket, the reaction temperature is controlled based on the predicted temperature correction model to ensure that the raw materials react fully. Its stable operation directly affects the reaction effect of ferric phosphate and the product quality. Separation control module: After the reaction is completed, the separation control module is put into operation. The centrifuge adjusts the separation speed through frequency conversion, the separation liquid level sensor monitors the liquid level in real time, and the automatic slag discharge valve controls the slag discharge according to the threshold, so as to achieve efficient solid-liquid separation, separate iron phosphate solid from liquid, and prepare materials for the drying process. Drying control module: The separated wet material enters the drying control module. The hot air dryer accurately controls the temperature based on the drying temperature PID control model. Temperature and humidity sensors provide real-time feedback data, and the exhaust flow valve adjusts the exhaust speed to remove the moisture from the material and obtain the finished product. Its precise control ensures that the product moisture content meets the standards. Multi-source data acquisition module: To realize precise control of the whole process, the multi-source data acquisition module is distributed in each process module, which collects key data such as feed flow and reaction temperature in real time, synchronizes the time sequence through the time synchronization module, and adjusts the sampling frequency adaptively to provide comprehensive, accurate and real-time data support for subsequent data processing; Data fusion processing module: Based on the collected data, the data fusion processing module plays a role, its data anomaly elimination module filters reliable data, the missing data correction module perfects the data, and the dynamic weighted fusion module fuses the data according to importance. The comprehensive data after deep processing provides a high-quality data basis for intelligent prediction and control, and improves data availability; Intelligent prediction and fuzzy control module: With high-quality data after processing, the intelligent prediction and fuzzy control module starts to work, the sliding time window prediction unit analyzes historical data to predict process trends, and the fuzzy control unit generates adjustment instructions based on prediction and current data, realizing precise prediction and control of the whole process; Process execution control module: After receiving the adjustment instructions, the process execution control module connects each process module through PLC, precisely controls the operating parameters of devices such as flow valve and stirring motor, realizes closed-loop control of the whole process, ensures each process link to run according to the optimal parameters, converts control instructions into actual production adjustment, and ensures stable operation of the system; Human-computer interaction and communication module: As an interactive bridge between the system and the operator, the human-computer interaction and communication module displays process data, curves and alarm information in real time, adopts double-ring redundant communication structure to ensure stable communication, supports remote monitoring and encrypted access, facilitates personnel operation and management, helps to discover and handle problems in time, and improves system operation and management efficiency.

[0033] Through systematic technological innovation, it shows significant advantages in improving production efficiency, reducing production cost and improving product quality. In terms of production efficiency, the multi-source data acquisition module comprehensively captures process parameters, the data fusion processing module deeply purifies them, the intelligent prediction and fuzzy control module predicts the process trend in advance and generates instructions, and the process execution control module responds quickly. Precise closed-loop control makes the production process always in the best condition, greatly reduces downtime and adjustment time caused by process fluctuations, and significantly improves production efficiency.

[0034] In terms of cost control, the feed control module accurately regulates the flow based on PID closed loop to avoid raw material waste; the reaction control module improves reaction efficiency and reduces energy consumption through intelligent temperature control and stirring adjustment; the separation and drying module accurately operates to reduce material loss and energy consumption. At the same time, intelligent control reduces manual intervention and reduces labor costs, effectively reducing the cost of iron phosphate preparation.

[0035] In terms of product quality assurance, precise control throughout the entire process ensures that each link is in the best condition. The reaction stage precisely controls temperature and stirring to ensure that the reaction is fully and evenly distributed. The drying link strictly controls temperature and humidity to ensure that the product moisture meets the standards. The intelligent prediction and fuzzy control module adjusts the parameters in real time to avoid process fluctuation risks. Multi-source data fusion analysis and intelligent control stabilize product quality, improve consistency and pass rate, and enhance the market competitiveness of enterprises The raw material storage tank of the feed control module is made of double-layer composite corrosion-resistant material. The inner layer is made of acid and alkali resistant special stainless steel, and the outer layer is wrapped with high strength fiber reinforced plastic. It can resist the chemical corrosion of raw materials such as phosphoric acid and iron salt, and can withstand the pressure changes of long-term storage. The intelligent monitoring device integrated at the top of the storage tank combines ultrasonic liquid level sensor and radar liquid level meter double detection technology, which can avoid misjudgment caused by single sensor failure through redundant design, and can accurately sense the subtle changes of the raw material liquid level in the storage tank.

[0036] The monitoring system has a built-in machine learning algorithm that can dynamically adjust the raw material inventory warning threshold based on past production cycles, order plans, and other data. When the liquid level approaches the warning line, in addition to triggering local audible and visual alarms, it will also send warning information to the procurement and production departments through the enterprise internal management system and automatically generate a procurement application form.

[0037] The frequency conversion screw pump is equipped with a high-performance frequency converter that uses magnetic flux vector control technology to complete speed adjustment in a very short time. Whether it is low-speed stable conveying of high-viscosity raw materials or rapid filling of low-viscosity raw materials, it can be accurately adapted. The electric flow control valve is based on a PID closed-loop control model, and its built-in high-precision pressure and flow sensors can collect data dozens of times per second, monitoring the flow rate and pressure of the raw material in the pipeline in real time. When the actual flow deviates from the set value, the PID controller quickly analyzes the deviation characteristics and adjusts the valve opening accurately in micro-steps through a stepper motor, and cooperates with the dynamic pressure compensation mechanism in the pipeline to ensure that the raw material flow remains stable under complex working conditions.

[0038] The stirring tank of the reaction control module adopts a double-layer cross-flow structure design, the inner tank body surface is treated by special polishing to reduce raw material adhesion and residue, and the outer layer is provided with a flow guide groove to optimize the fluid path. The matched servo-driven stirring motor is equipped with a high-resolution encoder and a precision planetary reducer, which can realize sub-trans level accurate adjustment of stirring speed.

[0039] Its built-in multi-mode control program can automatically switch between dispersion stirring, turbulent stirring, and laminar stirring modes according to the reaction progress: high-speed dispersion stirring is used in the early stage to quickly break up the raw material agglomeration; turbulent stirring is used in the middle stage to enhance molecular collision; laminar stirring is used in the later stage to avoid excessive shear. At the same time, the stirring blade is designed with a bionics curved surface, and the variable frequency torque compensation technology is used to ensure uniform mixing even for high-viscosity reaction systems.

[0040] The heating jacket adopts a composite temperature control mode of "electric heating + heat conduction oil circulation". The predictive temperature correction model deeply integrates reaction kinetics equations and machine learning algorithms. By correlating and analyzing multi-dimensional data such as temperature curves of historical batches, environmental temperature and humidity, and initial temperature of raw materials, the reaction temperature change trend can be predicted in advance.

[0041] The high-precision platinum resistance temperature sensor distributed in the jacket collects temperature data in real time at a frequency of 20 times per second. When temperature fluctuations are detected, the system immediately controls the electric heating power through the PID adjustment algorithm and synchronously adjusts the flow rate of the heat conduction oil circulating pump, so that the heat transfer efficiency of the heating jacket dynamically matches the reaction heat demand. At the same time, the infrared temperature measurement device on the tank wall surface is used for real-time monitoring of the temperature field to correct local temperature difference and ensure that the reaction temperature is always stable within the set range.

[0042] For example: in the feeding stage, the flow set value is 20 L / min, the PID parameter Kp is 1.2, and the actual flow fluctuation is ≤±0.3 L / min. In the reaction stage, the temperature set value is 80℃, the predictive correction amount ΔTpred is ±2℃, and the actual temperature is stable at 80±1℃.

[0043] Among them, the frequency conversion centrifugal separator in the separation control module adopts a double-cone drum structure design. The drum surface is treated by special laser etching to form micron-level anti-skid texture, which enhances the grabbing force of the iron phosphate solid and reduces liquid residue. The vector variable frequency drive system mounted on it can realize stepless adjustment of the separation speed in a very short time, adapting to the viscosity and density differences of different post-reaction materials.

[0044] The frequency conversion centrifugal separator is equipped with multi-axis vibration sensors and temperature monitoring devices to collect vibration frequency, bearing temperature and other parameters in real time during equipment operation. Through intelligent diagnostic algorithm analysis of equipment operation state, once abnormality is detected, early warning is triggered and operation parameters are automatically adjusted. The centrifugal field generated by high-speed rotation cooperates with the guide vane design inside the drum to guide the rapid separation of solid-liquid two-phase and improve the separation efficiency.

[0045] The separation liquid level sensor adopts non-contact radar measurement technology, which can penetrate complex working conditions and monitor the liquid level in the separator in real time with millimeter-level precision. The monitoring data is transmitted to the control system through redundant communication lines, and is processed through digital filtering and trend analysis to ensure accurate and reliable data. When the liquid level reaches the preset threshold, the control system immediately sends instructions to the pneumatic actuator to drive the automatic deslagging valve to open quickly.

[0046] The slag valve adopts a wear-resistant ceramic sealing structure, cooperates with a high-frequency response electromagnetic valve group, and can realize millisecond-level rapid opening and closing. The system supports manual and automatic dual-mode slag discharge control. In automatic mode, the slag discharge time and frequency can be dynamically adjusted according to the material characteristics, separation cycle and other parameters to avoid liquid overflow or solid blockage due to delayed slag discharge, ensure the continuous and stable operation of the separation process, and minimize material residue.

[0047] In the drying control module, the hot air dryer adopts a double-circulation air duct structure design, with a honeycomb ceramic heat accumulator and a spiral air curtain device inside. The ceramic heat accumulator can efficiently store and uniformly release heat, and the spiral air curtain can avoid hot air short-circuiting to ensure uniform heating of the material in the drying chamber. Combined with the drying temperature PID control model, the system collects data at a frequency of 15 times per second through multiple high-precision temperature and humidity sensors distributed at the air inlet, air outlet and material contact area of the dryer.

[0048] When a deviation between the actual humidity and the set value is detected, the PID controller quickly analyzes the humidity change trend and adjusts the heating pipe power, hot air circulation fan speed and fresh air supply to accurately fine-tune the drying temperature and ensure uniform drying of the material at the appropriate temperature.

[0049] The exhaust flow valve adopts a combination structure of a straight-travel electric push rod drive and a butterfly valve, with fast response and precise adjustment characteristics. Based on the feedback data of the temperature and humidity sensors, the system dynamically adjusts the exhaust speed through a fuzzy control algorithm. In the initial drying stage, the material moisture rapidly evaporates, and the exhaust system automatically increases the exhaust speed to timely remove a large amount of moisture. As the drying process progresses, the system gradually reduces the exhaust speed according to the real-time humidity changes, ensuring moisture removal while reducing heat loss.

[0050] The exhaust duct also has a pressure sensor and a dust concentration monitoring device. When an abnormal duct pressure or excessive dust concentration is detected, the system automatically adjusts the exhaust strategy to avoid the risk of reduced drying efficiency or dust accumulation due to poor exhaust.

[0051] The multi-source data acquisition module adopts a distributed deployment architecture and installs multiple types of industrial-grade sensors such as pressure, temperature, flow and liquid level at key points such as the feed pipe, reaction kettle, separation equipment and drying chamber. These sensors are strictly calibrated and protected, and some critical points are redundantly configured to ensure stable data collection with sub-second response speed under complex conditions such as high temperature, high humidity and strong corrosion.

[0052] The time synchronization module is built based on IEEE 1588 precision clock protocol, and achieves nanosecond-level time synchronization of each data acquisition node through fiber cascade networking mode, ensuring the time sequence consistency of cross-module data. The adaptive sampling frequency adjustment mechanism combines dynamic threshold detection and machine learning algorithm to continuously analyze the trend of process parameters. When the system detects that the process is in a steady state, it automatically reduces the sampling frequency to reduce data redundancy; while in the dynamic stage of reaction start, working condition switching, etc., the Kalman filter algorithm is used to strengthen data capture, and the sampling frequency is increased to meet the key parameter change monitoring requirements, which not only ensures the integrity of the data, but also effectively reduces the data transmission and storage pressure.

[0053] Among them, the data fusion processing module converts multi-source acquisition data into high-value information through the dual mechanisms of abnormal data processing and data deep fusion, and the abnormal data elimination module builds multiple levels of screening defense lines. First, based on the statistical 3σ principle, the basic threshold of key parameters such as feed flow and reaction temperature is set to quickly filter out data that deviates from the normal range; second, the isolation forest algorithm is introduced to identify abnormal points from the data distribution feature level in response to occasional transient interference or drift of the sensor. The module also has intelligent diagnosis function. When abnormal data is found, it will automatically trace back the data acquisition link, combine the sensor running state log and historical data to judge whether the abnormality is caused by equipment failure or environmental interference, and then accurately eliminate the wrong data.

[0054] In the data deep fusion link, the missing data correction module uses a hybrid interpolation strategy. For data missing for a short time, the effective data at the previous and next time points is used to fill in the missing data through the Lagrange interpolation method; for data missing for a long time, a prediction model based on LSTM neural network is called to learn the correlation and variation of parameters under similar working conditions, and a reasonable estimated value is generated. The parameter configuration of the LSTM network is 2 layers of hidden layers, 128 nodes per layer, and the training data comes from the first 100 batches of production data.

[0055] The dynamic weighted fusion module is based on D-S evidence theory and builds an adaptive weight distribution system. The system dynamically evaluates the credibility of each source data according to the calibration record of the sensor, the historical data fluctuation amplitude, the data update frequency and other factors, and gives appropriate weights to different types of data such as feed flow and stirring speed. Finally, through weighted average algorithm, the processed data is deeply fused to output comprehensive data that can accurately reflect the real-time state of the process.

[0056] The intelligent prediction and fuzzy control module realizes dynamic optimization of the whole process by the organic combination of trend accurate prediction and flexible intelligent regulation. In terms of trend accurate prediction, the intelligent prediction unit adopts a deep learning architecture that combines long short-term memory (LSTM) and attention mechanism, which can extract time series features from continuous process historical data. By analyzing the coupling relationship of multiple parameters such as feed flow, reaction temperature, stirring speed, etc., a nonlinear prediction model is established to accurately capture the trend of process parameters. The model also has online learning capability, which can dynamically update parameters based on real-time production data, continuously optimize prediction accuracy, and perceive the process trend in advance.

[0057] The fuzzy control unit constructs a multi-input multi-output fuzzy rule base covering key parameters such as temperature deviation, flow fluctuation, and speed change. The unit fuzzy processes the prediction results and real-time data, performs inference operations based on preset fuzzy rules, and generates accurate adjustment instructions through defuzzification. In the face of complex conditions such as temperature mutation and raw material characteristic fluctuation during the reaction process, the fuzzy control algorithm can simulate the decision logic of experienced operators, quickly adjust parameters such as feed flow, stirring speed, and heating power, and maintain the best operating state of the system in dynamic changes.

[0058] In terms of accurate execution of instructions, the process execution control module is based on high-performance PLC, integrates multiple industrial communication protocols such as PROFINET and Modbus-TCP, and realizes seamless connection with process unit devices such as feed and reaction. The PLC uses a multi-core processor architecture with high-speed data processing capability, which can analyze the adjustment instructions issued by the intelligent prediction and fuzzy control module at millisecond level. Through the modular digital and analog output unit, the device parameters such as variable frequency screw pump speed, stirring motor torque, and heating jacket power are accurately controlled, so that the instructions are converted into precise device actions, realizing real-time dynamic adjustment of process parameters in the whole process.

[0059] The system constructs a multi-level fault diagnosis system, uses monitoring devices such as vibration sensors and current transformers to collect real-time operation data of each unit, and combines preset fault feature models to quickly identify device abnormalities. Once an overload, temperature overrun, or control instruction exceeding the safety threshold is detected, the safety interlocking mechanism starts immediately, triggering the emergency shutdown program, and reminding the operator through sound and light alarm and pop-up window of the central control system.

[0060] In addition, the PLC supports online programming and remote debugging functions, allowing engineers to remotely access the system through encrypted networks, modify control logic flexibly according to actual production needs, and ensure safe, stable, and efficient production process.

[0061] Among them, the man-machine interaction and communication module constructs an efficient and reliable interaction management system through the two core functions of convenient operation interface and stable and safe communication, selects an industrial-grade high-resolution touch screen with anti-oil, scratch-resistant and wear-resistant characteristics to adapt to the complex operation environment of the workshop. Its graphical interface is designed based on a layered architecture, and the first page directly presents the real-time data of core process parameters such as feed flow and reaction temperature; the secondary page can expand the historical curve query function, support custom time span, generate smooth curves through data interpolation algorithm, and facilitate operators to analyze process trends. The alarm information processing module adopts hierarchical and classified display, reminds in different colors and sound-light modes according to the emergency level, and the operator can quickly locate the faulty equipment.

[0062] The industrial Ethernet dual-ring redundant communication structure adopts RSTP fast spanning tree protocol, and the main and standby link switching time is controlled within milliseconds, effectively dealing with network sudden failures. The 4G wireless communication module is equipped with an intelligent signal enhancement antenna, supports multi-operator network automatic switching, and guarantees the communication stability in remote areas. The VPN remote encrypted access adopts IPSec protocol to build a secure tunnel, combines with AES-256 high-strength encryption algorithm, and encrypts the transmission data in the whole process to prevent data theft and tampering. The system also has a communication state real-time monitoring function, which immediately triggers an early warning and automatically attempts link repair once a link exception is detected.

[0063] It should be noted that, in this text, relational terms such as first and second are used merely to distinguish one entity or action from another, without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variant thereof are intended to cover non-exclusive inclusions, so that a process, method, article, or apparatus that includes a list of elements does not only include those elements, but also includes other elements not explicitly listed, or inherent to such a process, method, article, or apparatus.

[0064] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A phosphorus iron preparation whole-process control system with multi-source data fusion monitoring, characterized in that: The system comprises: ​ The feed control module utilizes a variable frequency screw pump to deliver raw materials on demand, and based on a feed flow PID closed loop control model, enables an electric flow control valve to accurately adjust the feed flow; the raw materials enter a stirred tank of the reaction control module, and the stirring speed is accurately controlled by a servo-driven stirring motor; the heating jacket controls the reaction temperature based on a predictive temperature correction model; after the raw materials are reacted, a centrifugal separator of the separation control module adjusts the separation speed by frequency conversion, a liquid level sensor monitors the liquid level in real time, and a automatic slag discharge valve discharges slag according to a threshold value, thereby separating the iron phosphate solid from the liquid; after receiving the separated wet material, the drying control module accurately controls the temperature of the hot air dryer based on a drying temperature PID control model, and the temperature and humidity sensor feeds back data in real time, and the exhaust flow valve adjusts the exhaust speed; The multi-source data acquisition module is distributed in each process module, and real-time acquisition of feed flow, reaction temperature key data is performed, and the time synchronization module is unified in time; after receiving the collected data, the data exception elimination module of the data fusion processing module screens reliable data, the missing data correction module perfects the data, and the dynamic weighted fusion module fuses the data according to importance; after receiving the processed data, the intelligent prediction and fuzzy control module analyzes historical data to predict process trends through a sliding time window prediction unit, and a fuzzy control unit generates adjustment instructions according to the prediction and current data; after receiving the adjustment instructions, the process execution control module connects each process module through PLC, accurately controls the flow valve and stirring motor equipment operating parameters, and completes the full-process closed loop control; the man-machine interaction and communication module displays process data, curves and alarm information in real time, adopts a double-ring redundant communication structure to ensure stable communication, and supports remote monitoring and encrypted access. 2.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The feed control module comprises: (1) Raw material storage guarantee: the raw material storage tank is equipped with an intelligent monitoring device, and the raw material inventory is monitored in real time; once the inventory is insufficient, the early warning mechanism is triggered immediately, and the raw material is replenished in time; (2) Accurate flow control: the variable frequency screw pump flexibly adjusts the speed according to the needs of different process stages, and based on the feed flow PID closed loop control model, the electric flow control valve is driven by a stepping motor, the feed flow is continuously monitored, and the valve opening is quickly adjusted according to the deviation between the actual and set flow; The feed flow PID closed loop control model expression is: ； In the formula: is the feed flow rate corresponding to the current opening of the feed flow rate valve; The target feed flow rate calculated by the intelligent prediction and control unit is the feed flow rate deviation, defined as is the current actual feed flow rate; , is the feed flow rate PID controller proportional, integral, and derivative coefficient, Kp=0.8, Ki=0.2, and Kd=0.

1. 3.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The reaction control module comprises: (1) Efficient stirring and mixing: the stirring motor flexibly adjusts the stirring speed according to the needs of different reaction stages, and automatically switches the stirring mode through a preset program, so that the raw materials are fully mixed during the reaction process; (2) Intelligent temperature adjustment: the working of the heating jacket is based on the predictive temperature correction model; when the reaction temperature fluctuation is detected, the heating power of the heating jacket is automatically adjusted to compensate for the hysteresis of temperature change in time, so that the reaction temperature is quickly and stably set to the set value; The predictive temperature correction model expression is: ； In the formulae: Heating jacket target temperature Reaction actual temperature Temperature correction amount output by the prediction module.

4. The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The separation control module comprises: (1) Efficient solid-liquid separation: the separation control module comprises a variable frequency centrifugal separator, and the centrifugal separator is provided with a separation liquid level sensor to monitor the equipment operating condition in real time, and the centrifugal force generated by high-speed rotation separates the iron phosphate solid from the liquid quickly; (2) Automatic deslagging management: The liquid level sensor monitors the liquid level in the centrifugal separator in real time. When the liquid level reaches the pre-set threshold, the automatic deslagging valve quickly acts through the pneumatic actuator to timely discharge the separated liquid and solid residue. 5.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The drying control module comprises: (1) Precise temperature and humidity control: The drying control module comprises a hot air dryer and a temperature and humidity sensor. The drying temperature is precisely adjusted by combining the drying temperature PID control model. The temperature and humidity sensor monitors the humidity of the material and the ambient temperature in the drying process in real time. When the actual humidity deviates from the set humidity, the system automatically adjusts the heating power of the hot air dryer according to the deviation; Drying temperature PID control model expression: ； In the formula: Dryer current temperature set value; Predicted or process preset temperature; ; Proportional regulation coefficient; (2) Intelligent exhaust management: Based on the data feedback of the temperature and humidity sensor, the moisture generated in the drying process is discharged in time to maintain the stability of the drying environment. The drying efficiency is improved by precisely controlling the drying temperature and the exhaust speed. 6.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The multi-source data acquisition module comprises: (1) Comprehensive data acquisition: The multi-source data acquisition module is distributed in each key link of the preparation process to acquire real-time data of feed flow, reaction temperature and stirring speed; (2) Intelligent time sequence optimization: The multi-source data acquisition module includes a time synchronization module for unified time sequence processing of the data collected by each module. Based on the adaptive adjustment model, the sampling frequency is intelligently adjusted according to the dynamic changes of the process. Sampling frequency adaptive adjustment model expression: ； In the formula: current sampling frequency; base sampling frequency; fluctuation sensitivity coefficient; fusion parameter change rate. 7.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The data fusion processing module comprises: (1) Abnormal data processing: The data fusion processing module includes an abnormal data elimination module for strictly screening the collected data, identifying and eliminating abnormal data caused by sensor failure or external interference in real time; (2) Deep data fusion: The data fusion processing module also includes a missing data correction module and a dynamic weighted fusion module. The missing data correction module reasonably corrects the missing data caused by various reasons. The dynamic weighted fusion module uses a multi-source data weighted fusion model to assign appropriate weights to each source data, deeply fuses the processed data, and generates more accurate and more comprehensive data reflecting the actual process state. Multi-source data weighted fusion model expression: ； In the formula: Fused process parameters; The first Sensor real-time data; The first Weight, satisfying . 8.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The intelligent prediction and fuzzy control module comprises: (1) Accurate trend prediction: Based on continuous process historical data, the internal relationship and change law between process parameters are deeply mined to accurately predict future process trends; (2) Flexible intelligent control: The prediction results are combined with real-time data to generate accurate adjustment instructions using fuzzy control algorithms. According to different process states and prediction results, the control strategy is quickly and flexibly adjusted. 9.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The process execution control module comprises: (1) Accurate instruction execution: Based on the powerful output capability and precise control characteristics of PLC, the adjustment instructions output by the intelligent prediction and fuzzy control module are quickly and accurately converted into actual actions of the equipment to accurately control the operating parameters of each device. (2) Safe and stable operation: Once the equipment is detected to be abnormal or the control instruction is out of limit, the safety shutdown program is started immediately, and the operator is notified in time through sound and light alarm. Online programming and remote debugging functions of PLC are supported. 10.The iron phosphate preparation full-process control system with multi-source data fusion monitoring according to claim 1, characterized in that: The human-computer interaction and communication module comprises: (1) Convenient operation interface: Industrial-grade touch screen is adopted, and simple and intuitive graphical operation interface is designed, which integrates real-time display of process parameters, historical curve query, alarm information processing, and remote control function of equipment; (2) Stable and safe communication: Industrial Ethernet double-ring redundant communication structure is adopted, which has high reliability and stability, supports fast switching of main and standby links, ensures uninterrupted communication, supports 4G wireless communication and VPN remote encrypted access, and uses AES encryption algorithm to encrypt communication data. The specific process of main and standby link switching is: detecting main link failure → starting standby link → synchronizing data → switching complete (≤ 50 ms); Main and standby link switching delay formula: ； In the formula: Total link switching delay; Master link exception detection time, the master link exception detection mode is periodic heartbeat packet detection; Standby link recovery communication time.

Images