High-precision full-automatic titration system and method based on fusion of spectrum sensor and PLC

Abstract

The invention belongs to the technical field of physical and chemical analysis, and particularly discloses a high-precision full-automatic titration system and method based on fusion of a spectrum sensor and a PLC, a multi-channel spectrum sensor is adopted to capture fine spectrum characteristics of solution color change in real time, and the flow rate of titration liquid is dynamically adjusted through a PLC embedded PID control algorithm, so that the titration accuracy is improved. The method realizes submilliliter-level precision control, constructs a real-time compensation mechanism in combination with temperature, humidity, pH and other environmental parameters, improves the accuracy and robustness of titration end point judgment, automatically updates an end point judgment threshold and control parameters in combination with an AI self-learning module, realizes self-adaptive adjustment and sustainable evolution of the system, and improves the accuracy and robustness of titration end point judgment. The device has the characteristics of modular structure, high detection precision, high control response and high operation stability.

Description

Technical Field

[0001] The present invention belongs to the technical field of physical and chemical analysis, and particularly relates to a high-precision full-automatic titration system and method based on the integration of a spectral sensor and a PLC. Background Art

[0002] In routine laboratory chemical analysis, titration has always been an important means for determining the concentration of certain substances in a solution due to its simple operation and wide applicability. However, the application of traditional titration methods in the actual chemical production environment has many limitations, mainly reflected in the following aspects: (1) Strong dependence on manual operation and low automation. Traditional titration relies on manual operation, including adding liquid, observing color changes, and recording data. This not only has a large labor intensity and low efficiency but also has a large subjective error. It is difficult to be linked with modern chemical production lines (such as PLC and DCS systems) and lacks the ability of automatic feedback and control. (2) Unable to meet the requirements of real-time on-line monitoring. Most chemical production processes are continuous, requiring real-time detection and rapid response. Traditional titration requires off-line sampling, sending for inspection and analysis, resulting in delayed feedback, which may cause product fluctuations or even quality accidents. In addition, changes in environmental conditions during the sampling process are likely to introduce errors and cannot reflect the dynamic changes of the actual working conditions. (3) High complexity of samples and difficult endpoint recognition. Chemical samples may contain components with deep colors, high viscosities, suspended particles, or strong corrosiveness, which increase the difficulty of colorimetric judgment. Traditional visual colorimetry or single-band photoelectric detection methods perform poorly in such complex media, with inaccurate endpoint judgment and poor repeatability. (4) Harsh industrial site environment. Titration equipment needs to cope with adverse environments such as high temperature, high humidity, high pressure, vibration, and strong corrosive gases at the site. Traditional glass titration devices are easily damaged and pose a safety hazard to operators. External environmental fluctuations also affect the stability of the spectral sensor and the endpoint judgment. (5) Precision and repeatability cannot be guaranteed. The method of manually judging the endpoint and adjusting the titration rate is likely to introduce uncertainties, and traditional photoelectric endpoint judgment often uses fixed RGB thresholds, ignoring the process of color gradual change in the reaction system and unable to meet the control requirements at the sub-milliliter level.

[0003] Therefore, there is an urgent need for an on-line titration system integrating high-sensitivity color detection, environmental adaptability, intelligent control algorithms, and automation functions to improve the analysis precision, operation efficiency, and system reliability of chemical production and provide support for the next generation of industrial intelligence. Summary of the Invention

[0004] The present invention aims to solve the problems existing in traditional titration methods in industrial applications, such as low precision, cumbersome operation, poor environmental adaptability, and low automation. It constructs a high-precision fully automated online titration system and method based on the collaborative control of a spectral sensor and a PLC to achieve continuous, high-precision, low-error, and real-time quantitative analysis of key components in chemical solutions. This system is suitable for online sampling and real-time titration reaction monitoring in the chemical production process and can be widely applied to the automatic quantitative analysis of key components (such as ions, acid-base radicals, etc.) in chemical raw materials or products; continuous titration operations in high-frequency detection scenarios; and automated analysis tasks with extremely high requirements for endpoint judgment accuracy in fields such as environmental monitoring, water treatment, pharmaceuticals, and fine chemicals.

[0005] The system of the present invention achieves high-precision endpoint determination through multi-spectral color recognition, closed-loop precision control, and multi-parameter compensation, breaking through the limitations of traditional manual titration or low-intelligence devices in industrial applications and having significant engineering adaptability and industrial practical value.

[0006] The present invention provides the following technical solutions: a high-precision fully automated titration system based on the integration of a spectral sensor and a PLC, including a multi-channel color detection module, a PLC control module, an environmental compensation module, an intelligent liquid management system, an AI self-learning module, and a remote communication and data management module. The multi-channel color detection module uses an AS7341 spectral sensor to obtain multi-band spectral data of the titrant; the PLC control module collects spectral data and executes a PID closed-loop adjustment algorithm to dynamically control the injection rate of the titrant; the environmental compensation module integrates temperature, humidity, and pH sensors and uses a non-linear regression algorithm to correct the color monitoring error in real time; the intelligent liquid management system includes an automatic cleaning unit, a liquid level detection and replenishment unit, and an electric valve / micropump control device; the AI self-learning module uses historical titration error data to train a model, automatically optimizes the endpoint judgment threshold, and dynamically adjusts control parameters; the remote communication and data management module has functions such as data log recording, abnormal alarm, and remote parameter configuration.

[0007] Furthermore, the multi-channel color detection module can obtain multi-band spectral data in the range of 350 nm to 1000 nm to identify the color change of the titration solution and accurately judge the endpoint; the remote communication and data management module supports Modbus, OPCUA, Web, and APP remote control.

[0008] Furthermore, the multi-channel color detection module uses a high-resolution spectral sensor such as AS7341, supports 14-channel spectral acquisition, and cooperates with a filtering algorithm to remove background interference and monitor the color change process in real time.

[0009] Further, the PLC closed-loop control module incorporates a PID algorithm to precisely regulate the injection rate of the titrant, automatically decelerate titration according to the color change trend to avoid exceeding the end point, and improve accuracy and repeatability.

[0010] Further, the environmental parameter compensation module includes the input of sensor data for temperature, humidity, and pH, and dynamically corrects the color curve through a non-linear / regression fitting algorithm to enhance the adaptability to multiple working conditions.

[0011] Further, the intelligent liquid management system realizes the functions of automatic liquid replenishment, liquid level monitoring, and cleaning of the reaction vessel, and can switch multiple reagent channels in cooperation with electric valves / micropumps, suitable for complex titration processes.

[0012] Further, the AI self-learning module uses historical titration error data to train a model, fits and predicts the color change slope, and realizes the self-learning and parameter optimization update of the end point recognition model. The historical titration error data training model is a time series model such as LSTM.

[0013] Further, the remote communication and monitoring module supports industrial protocols such as Modbus and OPC UA, can be connected to the SCADA system, and is equipped with Web and mobile platforms such as APP to remotely configure parameters, view data, and receive warnings.

[0014] Further, the PLC control module can dynamically adjust the parameters of the PID controller according to the output of the AI self-learning module to achieve adaptive fluid titration adjustment.

[0015] Further, the multi-channel color detection module includes a 14-channel spectral sensor to obtain spectral data in the range of 350 nm to 1000 nm, which is used to identify the color change of the titration solution and accurately judge the end point, and has an anti-environmental interference light source filtering algorithm to improve the robustness of end point recognition.

[0016] Further, the AI self-learning module adopts a time series deep learning algorithm to extract end point features from the color change curve during the titration process and iteratively optimize the judgment criteria.

[0017] Further, the AI self-learning module adopts the LSTM or RNN algorithm.

[0018] Further, the intelligent liquid management system includes an electric valve / micropump control device and an automatic cleaning module; the electric valve / micropump control device is a peristaltic pump or a solenoid valve for automatically switching and quantitatively injecting various titration reagents; the automatic cleaning module supports periodic or result-triggered cleaning of the reaction vessel and the liquid delivery pipeline.

[0019] Furthermore, the remote communication and data management module supports the industrial platform to take over the titration process in real time and analyze historical trends through SCADA / ModbusTCP / IP, and remote data visualization and alarm response can be carried out through the Web interface and the mobile APP.

[0020] Furthermore, the high precision reaches sub-milliliter control precision.

[0021] A high-precision fully automatic titration method based on the integration of a spectral sensor and a PLC. The high-precision fully automatic titration method of the system includes the following steps: Step 1, sample introduction and system initialization: (1) Start the system, and the PLC controls the electric valve / micropump control device to automatically extract the sample liquid; (2) The detection system automatically calibrates the sensor, initializes the spectral sensor, temperature, humidity, and pH modules, and cleans the reaction vessel; (3) Turn on the light source of the titration container and prepare for the titration process; Step 2, start the titration process and collect data in real time: (1) Start titration, and use the AS7341 spectral sensor to collect data on the color change of the reaction solution in real time; (2) Upload the color change data to the PLC system to form a color change curve; (3) Synchronously record the temperature, humidity, and pH environmental parameters for dynamic compensation; Step 3, PID dynamic control of the titration speed (1) The PLC controls the start-stop frequency or rotation speed of the electric valve / micropump control device; (2) When approaching the reaction end point, gradually reduce the titration speed according to the rate and slope trend of the color change curve; Step 4, end point judgment and system feedback: (1) The system integrates the color change slope, color stability, and pH mutation to judge the reaction end point; (2) If the color signal meets the judgment criteria preset by the AI self-learning module, the system immediately stops titration; (3) Automatically record the process parameters. The process parameters include the end volume of the titrant, reaction time, temperature, humidity, and pH value; calculate the concentration value of the target ion or component; Step 5, result output and cloud upload: (1) Real-time display the titration result on the local screen; (2) At the same time, pack and upload the data to the cloud server. The data includes the detection time, sample number, and original curve; (3) Provide remote viewing functions on the Web side / APP side for easy traceability and data analysis; Step 6, automatic cleaning and preparation for the next round of detection: (1) The PLC automatically controls the cleaning liquid to be introduced into the reaction tank to efficiently clean the channels; (2) After the cleaning is completed, the system resets and waits for the next titration instruction; (3) If samples continuously enter, the system can automatically perform the next round of titration to form a fully automatic cyclic titration task flow.

[0022] The automatic titration system designed by the present invention achieves the following technical objectives: (1) Achieve high-precision recognition of multi-spectral colors. Using a high-channel resolution spectral sensor (such as AS7341), the color change data can be spectral data, and the color change trend of the solution in the range of 350nm - 1000nm can be recognized to accurately judge the titration end point. (2) Construct a closed-loop PID control mechanism. The dynamic adjustment of the titration liquid is realized through the PLC, and it adaptively decelerates when approaching the end point to ensure that the end point is not disturbed by "over-titration", achieving a sub-milliliter-level control accuracy (<0.1mL). (3) Integrate an environmental parameter compensation system. Temperature, humidity, and pH sensors are introduced to compensate the color data in real time to improve the stability of the system under complex working conditions such as high temperature and high humidity. (4) Introduce an AI self-learning module. Combining machine learning methods, the system can optimize the color end point discrimination logic based on historical titration error data, dynamically correct the threshold, and adjust the control strategy. (5) Achieve an automatic maintenance function. The system has functions of automatic cleaning, reagent replenishment, data recording, and remote control, meeting the requirements of continuous operation and reducing manual intervention.

[0023] This technical solution proposes an online fully automatic titration system for chemical production based on the deep integration of spectral sensors and PLC intelligent control, successfully breaking through multiple technical bottlenecks existing in the industrial application of traditional titration methods: (1) High-precision endpoint recognition ability. Adopting a 14-channel multi-band spectral detector, far superior to traditional RGB / single-band recognition; capable of sensitively capturing the color gradient and stability at the reaction endpoint, and achieving accurate determination in complex systems. (2) Closed-loop dynamic control and multi-parameter adaptation. Based on the closed-loop fluid control mechanism of the PID algorithm, the titration speed is adaptively adjusted; combined with sensor data such as temperature, humidity, and pH for non-linear compensation, effectively coping with on-site fluctuations and interference. (3) AI-driven continuous learning and model optimization. Introducing time series models such as LSTM, the system can dynamically correct the judgment criteria according to historical errors; with the increase in usage time, the accuracy and stability of endpoint recognition continue to improve, possessing the intelligent characteristic of "getting more accurate with use". (4) Industrial-level intelligence and automation level. The system has complete functions such as automatic liquid replenishment, cleaning, self-calibration, remote communication, and data uploading; supports industrial protocols such as Modbus / OPC / SCADA, and is suitable for the industrial 4.0 intelligent factory environment. (5) Wide applicability and engineering feasibility. It can adapt to various titration requirements in the fields of chemical engineering, environment, water quality, medicine, etc.; the modular structure is convenient for secondary development and on-site integration; it has the ability to operate stably for a long time and supports unattended scenarios.

[0024] The fully automatic titration system described in the present invention not only has significant advantages in multiple aspects such as measurement accuracy, operation efficiency, adaptation to the environment, and intelligent control, but also has broad application prospects and industrialization value in future digital factories and intelligent analysis platforms.

[0025] The beneficial effects of the present invention are as follows: Aiming at the deficiencies of traditional titration systems in industrial applications, a set of fully automated online titration systems and methods with the following significant technological innovations are proposed, having the advantages of a multi-channel spectral endpoint recognition mechanism, achieving a closed-loop PID micro-control accuracy up to the sub-milliliter level, integrating a multi-parameter real-time compensation algorithm, implementing an AI self-learning module, a fully automatic closed-loop mechanism for liquid management, and integrating remote visualization + industrial integration capabilities.

[0026] Compared with the prior art, the present invention includes a multi-channel spectral endpoint recognition mechanism. The system adopts high-resolution spectral sensors such as AS7341, with the ability to recognize 14 channels of colors in the range of 350nm to 1000nm. Compared with traditional RGB three-channel or single-band photoelectric sensors, this solution can more sensitively capture the changing trend of the solution color during titration; improve the resolution and anti-interference ability of endpoint determination; and adapt to chemical reaction systems with complex colors and multi-band changes.

[0027] Compared with the prior art, the present invention achieves a closed-loop PID micro-control accuracy up to the sub-milliliter level. The PLC system incorporates a closed-loop PID control algorithm, adjusts the rotation speed of the titration pump through feedback, and controls the dynamic change of the titration liquid flow rate. The achieved advantages include slowing down the dropping speed near the end point to avoid over-dropping; the control accuracy reaching less than 0.1 mL, far superior to manual operation; and being applicable to micro-analysis and systems with a small concentration gradient.

[0028] Compared with the prior art, the present invention integrates a multi-parameter real-time compensation algorithm. The system integrates environmental sensor data such as temperature, humidity, and pH value, and compensates the spectral sensor data through a non-linear fitting regression model to achieve consistency in end-point judgment under different environmental conditions; reduces misjudgment and drift caused by environmental fluctuations; and enhances the environmental robustness of the system in the industrial field.

[0029] Compared with the prior art, the present invention realizes end-point optimization with the AI self-learning module. The system introduces a deep learning model (such as LSTM), automatically trains a color change prediction model based on the titration historical data, and realizes dynamic adjustment of the titration end-point recognition standard. The longer the system runs, the more stable the model; improves the end-point judgment accuracy and model adaptability during long-term operation.

[0030] Compared with the prior art, the present invention realizes a fully automatic closed-loop mechanism for liquid management. The solution includes a complete reagent replenishment, liquid level monitoring, and automatic cleaning mechanism. With the cooperation of electric valve and micro-pump control, it can achieve regular cleaning and pollution prevention of the titration reaction channel, automatic switching and replenishment of multiple titration reagents; reduces the frequency of manual maintenance and adapts to unattended environments.

[0031] Compared with the prior art, the present invention integrates remote visualization + industrial integration capabilities. The system has complete industrial communication interfaces (ModbusTCP / IP, OPCUA), can be docked with SCADA, and realizes data synchronization, historical records, and trend analysis on the Web and APP sides; remote start, alarm notification, and parameter adjustment; meets the full-process digital requirements of "data + control + visualization" in the industrial 4.0 environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Hereinafter, the present invention will be described in more detail based on embodiments with reference to the drawings.

[0033] Figure 1 It is a structural diagram of an automatic titration system.

[0034] Figure 2 It is a top view structural schematic diagram of the automatic titration system.

[0035] Figure 3 It is a front view structural schematic diagram of the automatic titration system.

[0036] Figure 4 Schematic diagram of the connection of the PLC controller, peristaltic pump, and titration cell in the automatic titration device.

[0037] Figure 5 Closed-loop PID control structure diagram.

[0038] Figure 6 Structure diagram of the endpoint learning and prediction model of the AI self-learning module.

[0039] Figure 7 SCADA / Modbus and Web / APP data interaction architecture diagram.

[0040] Figure 8 Automatic titration curve generated by the titration system shown in Example 3. Detailed implementation manners

[0041] Next, the technical solutions in this embodiment will be clearly and completely described in conjunction with the accompanying drawings in this embodiment. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0042] Example 1 High-precision fully automatic titration system based on the integration of a spectral sensor and a PLC

[0043] The system of the present invention includes a multi-channel color detection module, a PLC control module, an environmental compensation module, an intelligent liquid management system, an AI self-learning module, and a remote communication and data management module; the multi-channel color detection module uses a spectral sensor such as AS7341 to obtain multi-band spectral data in the range of 350 nm to 1000 nm for identifying the color change of the titration solution and accurately judging the endpoint; the PLC control module collects the data of the spectral sensor and executes the PID closed-loop adjustment algorithm to realize the dynamic control of the injection rate of the titrant; the environmental compensation module integrates temperature, humidity, and pH sensors and uses an algorithm fitting to correct the color measurement error in real time; the intelligent liquid management system includes an automatic cleaning unit, a liquid level detection and replenishment module, and an electric valve / micropump control device to ensure the stable and continuous operation of the system; the AI self-learning module uses historical titration error data to train the model, automatically optimizes the endpoint judgment threshold, and dynamically adjusts the control parameters; the remote communication and data management module supports Modbus, OPCUA, Web, and APP remote control and has functions such as data logging, abnormal alarm, and remote parameter configuration.

[0044] The system uses a multi-channel spectral sensor (such as AS7341) to capture the subtle spectral characteristics of the color change of the solution in real time, and dynamically adjusts the titrant liquid flow rate through the PID control algorithm embedded in the PLC to achieve precise control at the sub-milliliter level. The system combines environmental parameters such as temperature, humidity, and pH to construct a real-time compensation mechanism to improve the accuracy and robustness of titration endpoint judgment. The present invention also introduces an AI self-learning module, trains and optimizes the model based on historical data, automatically updates the endpoint judgment threshold and control parameters, and realizes the adaptive adjustment and continuous evolution of the system.

[0045] Reference Figures 1-7 , Figure 1 illustrates the connection relationship between the functional modules of the system; the module includes a PLC controller, which can receive the information feedback from the color sensor module and the environmental compensation module to realize the automatic control of titration in the reaction vessel. The module also includes an AI self-learning module and a PID controller, including a remote monitoring module; the PID controller controls the colorimetric cell and the peristaltic pump. The colorimetric cell is used to detect the color change during titration, and the peristaltic pump is used to control the speed of the titrant; Figure 1 indicates the logical order of physical connection and signal transmission of each module. Figure 2 is a top view of the spectral sensor and the titration cell structure. Among them, the titration cell is made of a transparent corrosion-resistant material (such as quartz glass or PMMA); the light source is a highly stable LED light source (white light or a specific wavelength) aligned with the titration container to emit light; the AS7341 spectral sensor is located at the receiving end to collect the transmitted / reflected spectrum; the pH electrode and the temperature sensor detect the pH and temperature of the solution in the titration cell, and the indicator 1 port and the indicator 2 port are used to indicate the color change; the titration 1 and titration 2 ports are used to inject the titrant solution, and the pure water port and the drain port are the inlet and outlet of the solution entering the titration cell respectively. Figure 3 is a front view of the spectral sensor and the titration cell structure, which shows that during the magnetic stirring and heating process, the temperature and pH during the reaction process are automatically monitored. At the same time, under the irradiation of the LED light source, the spectral sensor collects the spectral changes during the solution titration process to achieve full-automatic precise control; Figure 4 shows the relationship between the titration pump (multi-head peristaltic pump) and the PLC controller and the components such as sensors, titration ports, sensors, and light sources on the titration cell. The PLC controller connects and controls the light source, sensors, and multi-head peristaltic pump; connects the pH electrode and the temperature sensor; the peristaltic pump controls the injection rate of the titrant. Figure 5 is the process of controlling titration through the PLC control module and the PID control algorithm; the color sampling is located in the reaction vessel and is used to online detect the PID feedback loop. The process includes the sensor collecting the color change → the controller calculating the error in real time to obtain the feedback → adjusting the action of the peristaltic pump controller to adjust the flow rate; terminating the titration when the color value (such as ΔRGB, ΔAbsorbance) approaches the stable interval; the system has the functions of anticipatory control and soft landing to avoid "over-titration".Figure 6 It is described that the model input is a historical titration data sample library. Combining the color change curve with environmental variables (such as Temperature / pH / humidity), a real-time feature extraction module is used to obtain the RGB change rate and slope; the output is the predicted end time point or color stable section; the model structure adopts the LSTM / RNN framework and is iteratively trained through historical errors. Figure 7 It is described that this system accesses the industrial control platform SCADA through Modbus, connects to the PLC network control unit, and realizes the lloT communication gateway through MOTT or HTTP; describes data upload, remote control, alarm feedback, and historical trend query; at the same time, it supports cloud, Web management UI platform and mobile APP visualization interfaces.

[0046] This system supports functions such as automatic calibration, automatic cleaning, and reagent replenishment, and can be connected to the SCADA platform through industrial communication protocols (such as Modbus, OPCUA) to achieve remote monitoring and industrial integration; at the same time, it provides Web and mobile data visualization interfaces to meet the requirements for intelligent chemical analysis equipment under the background of Industry 4.0.

[0047] This system has the characteristics of modular structure, high-precision detection, high-response control, and high-stability operation, and is especially suitable for continuous and unmanned industrial analysis scenarios, providing reliable technical support for chemical industrial intelligent control systems.

[0048] Example 2 A method for realizing fully automated titration operation at a chemical production site Adopt a high-precision fully automated titration system based on the integration of a spectral sensor and a PLC. The specific working process or operation method steps are as follows: Step 1, sample introduction and system initialization; (1) After starting the system, the PLC controls the peristaltic pump / solenoid valve to automatically extract the sample liquid; (2) The detection system automatically calibrates the sensor, initializes the spectral sensor, temperature / humidity / pH module, and cleans the reaction vessel; (3) Turn on the light source of the titration vessel and prepare for the titration process; Step 2, start the titration process and collect data in real time; (1) Start titration, and use the AS7341 spectral sensor to collect spectral data on the color change of the reaction solution in real time; (2) Upload the color change data to the PLC system to form a color change curve; (3) Synchronously record environmental parameters such as temperature, humidity, and pH for dynamic compensation; Step 3, PID dynamic control of the titration speed; (1) The PLC controls the start-stop frequency or rotation speed of the peristaltic pump / solenoid valve; (2) When approaching the reaction end point, gradually reduce the titration speed according to the rate and slope trend of the color change curve; (3) Implement the fine control logic of "fast start, slow approach" during the titration process; the control accuracy can be stabilized at <0.1 mL; Step 4, end point judgment and system feedback; (1) The system integrates multiple factors such as the color change slope, color stability, and pH mutation to obtain the judgment criteria for the reaction end point; (2) If the color signal meets the judgment criteria preset by the AI self-learning module, the system immediately stops titration; the AI self-learning module uses historical titration error data to train and optimize the model, fits and predicts the color change slope, and realizes the self-learning and parameter optimization update of the end point recognition model; the historical titration error data training model is preferably a time series model such as LSTM.

[0049] (3) Automatically record process parameters such as the end point volume of the titrant, reaction time, temperature / humidity / pH value, etc.; the end point volume is used to calculate the concentration value of the target ion or component.

[0050] Step 5, result output and cloud upload; (1) Real-time display the titration results (including , , etc. concentrations) on the local screen; (2) At the same time, pack and upload the data to the cloud server, including the detection time, sample number, original curve, etc.; (3) Provide remote viewing functions on the Web side / APP side for easy traceability and data analysis.

[0051] Step 6, automatic cleaning and preparation for the next round of detection; (1) The PLC automatically controls the cleaning liquid to be introduced into the reaction cell to efficiently clean the channels; (2) After the cleaning is completed, the system resets and waits for the next titration instruction; (3) If samples continuously enter, the system can automatically perform the next round of titration to form an unattended cyclic titration task flow.

[0052] This process can be fine-tuned according to different samples (such as titration, detection or iron ion complexometric titration), and has good adaptability and scalability.

[0053] Example 3 An automatic titration for acid radicals ( ) and chloride ions ( ) in compound fertilizers Using the historical titration error data of sulfate and chloride ions in the laboratory to train the LSTM time series model, fitting and predicting the slope of the titration color change, realizing the self-learning and parameter optimization and update of the end-point recognition model, and obtaining the optimized model of the AI self-learning module for acid root ions ( ), and chloride ions ( ).

[0054] Based on the multi-channel color detection module, PLC control module, environmental compensation module, intelligent liquid management system, AI self-learning module, remote communication and data management module, this system can realize the full-automatic on-line titration determination of sulfate and chloride ion concentrations in compound fertilizer samples. The operation process is as follows: (1) Sample preparation: The PLC controls the electronic balance to weigh 5 g of compound fertilizer particles and records the mass M; Start the peristaltic pump to add 100 mL of pure water (record the volume V1), and at the same time turn on the magnetic stirrer; Heat to 95 °C to accelerate dissolution, and measure the final volume V2 after cooling; The system calculates the concentration of the original solution according to M, V1 and V2.

[0055] (2) Titration of sulfate ions ( ): The PLC controls the peristaltic pump to extract 10 mL of the sample into the reaction vessel; Automatically add 2 drops of methyl red-methylene blue mixed indicator, and the solution turns purple; The system automatically collects the temperature, humidity, and pH value to adjust the titration parameters and the color determination threshold R1; Gradually add the NaOH standard solution, and the color changes from purple-red to gray-green; The AS7341 spectral sensor collects the spectrum in real time, and the AI self-learning module identifies the end point and stops immediately when the stable threshold is reached; The PLC records the titration volume V3 and calculates of the concentration value.

[0056] (3) Titration of chloride ions ( ): After titrating , automatically add 2 mL of potassium chromate (K2CrO4) as the second indicator; The peristaltic pump controls the addition of the AgNO3 titrant, and the color changes from light yellow to brick red (the formation of Ag2CrO4 precipitate); The spectral sensor collects the RGB / spectrum change and judges the end point (threshold R2); Record the titration volume V4 and calculate the chloride ion concentration.

[0057] (4)Automatic output and cleaning: The titration results are displayed on the touch screen and simultaneously uploaded to the cloud database; The system automatically controls the cleaning liquid to clean the reaction vessel and pipelines; Prepare for the next round of sample detection.

[0058] This embodiment verifies the reliability and adaptability of the system under complex backgrounds, complex reactions, and multi-step judgment conditions, and demonstrates the broad feasibility of the high-precision titration system in industrial automation analysis. Refer to Figure 8 as shown; in the early stage of titration, the PLC controls the stepper motor peristaltic pump to drip the NaOH standard solution at a speed of 1 mL / min. When approaching the titration end point, the PLC controls the stepper motor peristaltic pump to drip at a speed of 0.5 mL / min until the titration end point. During the dripping process, the AS7343 spectral sensor continuously detects the spectral data of the liquid in the titration container. There are a total of sixteen channels. The AS7343 spectral sensor uploads the data to the single-chip microcomputer, uploads the spectral data to the host computer for storage and data processing, and draws the spectral curve during titration to visually display the titration process. Figure 8 The 16 titration curves in correspond to the data collected by the sixteen-channel signal acquisition. The color of the solution before titration is purple. During the titration process, the change in the solution color is not obvious in the early stage. When approaching the titration end point, at about 580 seconds, the color changes suddenly from purple to green, corresponding to a drastic change in the curve of the spectral data collected by the 4th channel at 505 - 525 nm. The AS7343 sensor captures the spectral mutation at a specific wavelength and calculates the volume of the NaOH standard solution added based on this, and finally calculates the sulfuric acid content.

[0059] The fully automatic titration system integrates a multi-channel color detection module, a PLC control module, an environmental compensation module, an intelligent liquid management system, an AI self-learning module, and a remote communication and data management module. Due to its high precision, high robustness, high intelligence, and industrial interconnection capabilities, it can be widely applied to the following typical industries and application scenarios: (1) On-line quality monitoring of chemical production lines. It is applied to the ion concentration detection of intermediates and finished products in continuous chemical plants; it realizes the real-time identification and feedback of the end points of acid-base neutralization, complexation reactions, etc.; it can be integrated into the DCS system to realize closed-loop process control. (2) High-throughput titration analysis platform in laboratories. It is suitable for automated titration experiments in universities, research institutes, and testing institutions; it can complete batch quantitative analysis of multiple groups of samples, improve work efficiency; reduce human operation errors, and improve the repeatability and accuracy of experimental data. (3) Water quality monitoring and environmental protection industries. It is used for on-line monitoring of heavy metal ions / inorganic salt concentrations in rivers, lakes, and drinking water sources; typical indicators such as iron ions ( ), chloride ions ( ), nitrate, sulfate, etc.; it can be deployed in fixed monitoring stations to achieve remote unattended operation. (4) Monitoring of pharmaceutical and fine chemical synthesis processes. Titration detection of pH and intermediate concentration during drug synthesis; automatically feedback endpoint information to assist the automatic end of the synthesis reaction or control temperature / reactant addition; achieve quality consistency control under GMP-compliant production. (5) Quality control in the food and beverage industry. Used for the ratio control of food additives and acidity regulators; achieve quantitative detection of acidity / sodium salt concentration in products such as fruit juices, beverages, and flavoring liquids; combined with the data traceability mechanism, improve the level of food safety management. (6) Intelligent manufacturing and Industry 4.0 integrated platform. This system supports SCADA, Modbus, Web remote visualization and big data analysis; it can be deployed as an intelligent chemical analysis terminal in digital factories; supports the expansion of industrial Internet (lloT) nodes and the optimized operation and maintenance of edge AI modules. The modules of this system have high scalability and can be customized and combined with sensors, reaction pools, and pump valve systems according to actual scenarios, so as to cover a wider range of chemical and chemical detection fields.

[0060] Although the present invention has been described with reference to the preferred embodiments, various improvements can be made to it and components therein can be replaced with equivalents without departing from the scope of the present invention. In particular, as long as there is no structural conflict, the technical features mentioned in each embodiment can be combined in any way. The present invention is not limited to the specific embodiments disclosed in the text, but includes all technical solutions falling within the scope of the claims.

Claims

1. A high-precision fully automatic titration system based on the integration of a spectral sensor and a PLC, characterized in that, Including: A multi-channel color detection module, a PLC control module, an environmental compensation module, an intelligent liquid management system, an AI self-learning module, and a remote communication and data management module. The multi-channel color detection module uses an AS7341 spectral sensor to obtain multi-band spectral data of the titrant. The PLC control module collects the spectral data and executes a PID closed-loop regulation algorithm to dynamically control the injection rate of the titrant. The environmental compensation module integrates temperature, humidity, and pH sensors and uses a non-linear regression algorithm to correct the color monitoring error in real time. The intelligent liquid management system includes an automatic cleaning unit, a liquid level detection and replenishment unit, and an electric valve / micropump control device. The AI self-learning module uses historical titration error data to train and optimize the model, automatically optimizes the end-point judgment threshold, and dynamically adjusts the control parameters. The remote communication and data management module has functions of data logging, abnormal alarm, and remote parameter configuration.

2. The high-precision full-automatic titration system based on the fusion of a spectral sensor and a PLC according to claim 1, wherein The PLC control module can dynamically adjust the parameters of the PID controller according to the output of the AI self-learning module to achieve adaptive fluid titration regulation.

3. The high-precision full-automatic titration system based on the fusion of a spectral sensor and a PLC according to claim 1, characterized in that, The multi-channel color detection module includes a 14-channel spectral sensor, which obtains spectral data in the range of 350nm to 1000nm, is used to identify the color change of the titration solution and accurately judge the end point, and has an anti-environmental interference light source filtering algorithm to improve the robustness of end-point recognition.

4. The high-precision fully automatic titration system based on the fusion of a spectral sensor and a PLC according to claim 1, wherein, The AI self-learning module uses a time series deep learning algorithm to extract end-point features from the color change curve of the titration process and iteratively optimize the judgment criteria.

5. The high-precision full-automatic titration system based on the fusion of a spectral sensor and a PLC according to claim 1, wherein, The AI self-learning module uses the LSTM or RNN algorithm.

6. The high-precision full-automatic titration system based on the fusion of a spectral sensor and a PLC according to claim 1, wherein The intelligent liquid management system includes an automatic cleaning unit; the electric valve / micropump control device is a peristaltic pump or a solenoid valve, which is used to automatically switch and quantitatively inject various titration reagents; the automatic cleaning unit supports periodic or result-triggered cleaning of the reaction vessel and the liquid delivery pipeline.

7. The high-precision full-automatic titration system based on the fusion of a spectral sensor and a PLC according to claim 1, characterized in that, The remote communication and data management module supports the industrial platform to take over the titration process in real time and analyze the historical trend through SCADA or Modbus or TCP / IP, and perform remote data visualization and alarm response through a Web interface and a mobile APP.

8. The high-precision full-automatic titration system based on the fusion of a spectral sensor and a PLC according to claim 1, wherein The high precision is sub-milliliter-level control precision.

9. The high-precision fully automatic titration system based on the fusion of a spectral sensor and a PLC according to any one of claims 1-8, characterized in that, The high-precision full-automatic titration method of the system includes the following steps: Step 1, sample introduction and system initialization: (1) Start the system, and the PLC controls the electric valve / micropump control device to automatically extract the sample liquid. (2) The detection system automatically calibrates the sensors, initializes the spectral sensor, temperature, humidity, and pH modules, and cleans the reaction vessel. (3) Turn on the light source of the titration container and prepare for the titration process. Step 2, start the titration process and collect data in real time: (1) Start the titration, and use the AS7341 spectral sensor to collect the color change of the reaction solution in real time to obtain color change data. (2) Upload the color change data to the PLC system to form a color change curve. (3) Synchronously record the temperature, humidity, and pH environmental parameters for dynamic compensation. Step 3, PID dynamically control the titration speed (1) The PLC controls the start-stop frequency or rotational speed of the electric valve / micropump control device; (2) Near the reaction end point, gradually reduce the titration speed according to the rate and slope trend of the color change curve; Step 4, End point judgment and system feedback: (1) The system integrates the color change slope, color stability, and pH mutation to obtain the determination criteria for the reaction end point; (2) If the color signal meets the determination criteria preset by the AI self-learning module, the system immediately stops titration; (3) Automatically record the process parameters, which include the end volume of the titrant, reaction time, temperature, humidity, pH value; calculate the concentration value of the target ion or component; Step 5, Result output and cloud upload: (1) Real-time display the titration result on the local screen; (2) At the same time, pack and upload the data to the cloud server. The data includes the detection time, sample number, and original curve; (3) Provide remote viewing functions on the Web / APP side for easy traceability and data analysis; Step 6, Automatic cleaning and preparation for the next round of detection: (1) The PLC automatically controls the cleaning liquid to enter the reaction cell to efficiently clean the channels; (2) After cleaning, the system resets and waits for the next titration instruction; (3) If samples continuously enter, the system can automatically perform the next round of titration to form a fully automatic cyclic titration task flow.

Images