Metal ion online detection method and system

Abstract

The invention relates to a metal ion on-line detection method and system, and the metal ion on-line detection method comprises the following steps: (1) leading out a sample to be detected from a system to be detected, forming an independent continuous flow type sample flow path, and carrying out real-time detection; (2) collecting a detection signal, and sequentially carrying out preprocessing, interference correction processing and feature extraction to obtain input data of inversion calculation; and (3) carrying out inversion calculation and outputting a detection result in real time. Compared with a traditional off-line detection mode, the real-time monitoring technology can achieve process visualization and quick response, and theoretical and technical supports are provided for building an intelligent and green rare earth separation production system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent detection and control in rare earth extraction and purification process, and particularly relates to a metal ion online detection method and system. BACKGROUND

[0002] Currently, in important technical fields such as science and technology, military and national defense, rare earth materials play an irreplaceable role. Rare earth elements (REE) refer to the total of 17 elements from 57th element lanthanum (La) to 71st element lutetium (Lu) in the periodic table of chemical elements, and scandium (Sc) and yttrium (Y). Ion-type rare earth ore has the advantages of shallow burial, convenient mining, high enrichment of rare earth elements and low radioactivity, and becomes the main source of heavy rare earth resources. For ion-type rare earth ore, its unique occurrence form determines the different mining and extraction process from traditional rare earth ore. The rare earth elements in ion-type rare earth ore are mainly adsorbed on the surface of clay minerals in the weathering crust in exchangeable state, so neutral or weak acid solution is often used for ion exchange extraction, which has the advantages of low energy consumption and simple process. However, the traditional leaching process has the problems of serious environmental pollution, destruction of vegetation and soil, and serious water and soil loss. Therefore, green mining of ion-type rare earth ore has become the core direction and technical bottleneck of current rare earth industry development, and the green extraction technology system represented by multi-stage extraction has become an important means to improve resource utilization efficiency and reduce environmental burden, because of its advantages of high selectivity, high recovery rate and process controllability. Multi-stage extraction technology can realize efficient separation and enrichment of rare earth elements by step-by-step control of extractant concentration, pH value and phase ratio, which reduces the co-extraction of impurities and the generation of harmful waste liquid. The energy efficiency improvement of multi-stage extraction technology is limited by the lag of traditional detection technology. The production line generally lacks real-time monitoring mechanism for key parameters, especially the concentration of rare earth ions in extraction and stripping phases, which restricts the implementation of process optimization and precise control to some extent.

[0003] Current industrial production mainly relies on offline chemical analysis methods such as ICP-OES and spectrophotometry, which not only have lagging response and complex operation, but also cannot meet the real-time data demand of continuous and automatic production.

[0004] Therefore, it is particularly important to develop real-time dynamic detection technology for ion concentration in rare earth ion purification and extraction process, which not only helps to realize precise monitoring and adjustment of key parameters in extraction process, but also effectively improves extraction efficiency and product purity. Then, corresponding monitoring and control strategies and feedback adjustment mechanisms can be planned for different production process lines to realize closed-loop management from data acquisition, process analysis to process optimization. SUMMARY

[0005] To solve the above technical problems, the present application realizes real-time and accurate detection of metal ion concentration in a rare earth extraction system by constructing a continuous flow sample online detection system, combining a multi-step signal processing and an intelligent inversion calculation integrated detection system, solving the problems of strong lag, complex operation and inability to meet the needs of continuous production of traditional offline detection, and providing key technical support for precise regulation and energy efficiency improvement of multi-stage extraction process.

[0006] To achieve this purpose, the present application adopts the following technical solutions:

[0007] In a first aspect, the present application provides a metal ion online detection method comprising:

[0008] (1) The sample to be detected is introduced from the system to be detected and forms an independent continuous flow sample flow path for real-time detection;

[0009] (2) Collecting detection signals for pre-processing, interference correction processing and feature extraction to obtain input data for inversion calculation;

[0010] (3) Inversion calculation is performed to output detection results in real time.

[0011] The present application solves the technical problem that traditional offline detection cannot realize continuous real-time monitoring by constructing a continuous flow sample flow path, signal multi-stage layer processing and real-time inversion calculation process, and realizes dynamic online detection of metal ion concentration. Through the design of continuous flow sample flow path, the real-time update of the sample to be detected is ensured and the original system is not disturbed, the detection signals are pre-processed, interference corrected and feature extracted in sequence, the quality of the input data is greatly improved, and the inversion calculation is performed based on the processed data, which breaks through the lag limitation of offline detection and can real-time feedback the change of metal ion concentration, providing data support for real-time regulation of extraction process.

[0012] The following is a preferred technical solution of the present application, but not as a limitation of the technical solutions provided by the present application, through the following preferred technical solutions, the purpose and beneficial effects of the present application can be better achieved and realized.

[0013] As a preferred technical solution of the present application, the detection of step (1) includes spectral detection.

[0014] Preferably, the spectral detection includes absorption spectrum detection.

[0015] Preferably, the absorption spectrum detection wavelength covers 340-1100 nm.

[0016] Preferably, the system to be detected includes a rare earth extraction system.

[0017] Preferably, the sample to be tested includes an aqueous phase liquid and / or an organic phase liquid of a rare earth extraction system.

[0018] As a preferred technical solution of the present invention, the preprocessing in step (2) includes baseline correction, denoising and normalization performed sequentially.

[0019] Preferably, the preprocessing can be performed using models commonly used in the art, and no further limitations are imposed here.

[0020] Preferably, the interference correction process in step (2) includes dynamic compensation for changes in environmental and device parameters.

[0021] Preferably, changes in environmental and device parameters include temperature drift, humidity drift, spectral baseline drift, light source attenuation, and environmental vibration.

[0022] Preferably, the dynamic compensation is performed by combining models such as Idle noise subspace learning, CCA common mode suppression, adaptive threshold, regular intensity selection, EWMA / Kalman drift compensation, and RPCA low-rank-sparse separation.

[0023] This invention solves the technical problem of decreased accuracy of detection signals due to interference from environmental and device parameter changes by optimizing the signal preprocessing process and adopting a dynamic compensation mechanism combining multiple models. It significantly improves the stability of data quality. Baseline correction, denoising, and normalization successively eliminate signal background shift, random noise, and magnitude differences, providing clean raw data for subsequent processing. For complex interferences such as temperature drift and light source attenuation, multiple models work together to accurately cancel out multi-source interference, effectively resisting the influence of environmental and device fluctuations on the signal, providing high-quality input for inversion calculations, and further ensuring the accuracy of detection results.

[0024] As a preferred technical solution of the present invention, the inversion calculation in step (3) includes inputting the input data in step (2) into the inversion model for calculation.

[0025] Preferably, the inversion model is trained by using machine learning and statistical regression algorithms to calculate input data from multiple sets of solutions with known concentrations.

[0026] Preferably, the machine learning algorithm includes any one of support vector regression, random forest regression, gradient boosting regression tree, and artificial neural network regression.

[0027] Preferably, the statistical regression algorithm includes any one of multiple linear regression, ridge regression, and principal component regression.

[0028] As a preferred embodiment of the present invention, the method further includes self-learning updates of the inversion model.

[0029] Preferably, the self-learning update comprises verifying calculation of the step (3) detection result with the true concentration, and when the determination coefficient R 2 optimizing the inversion model when the value is lower than a preset threshold.

[0030] Preferably, the true concentration is manually entered into the system after ICP-OES detection in a specific period, and the preset threshold can be adjusted according to the detection accuracy requirement, which is not limited here.

[0031] Preferably, the optimization of the inversion model comprises repeatedly performing the inversion model training on the original inversion model.

[0032] The present application realizes efficient inversion of metal ion concentration by constructing an inversion model based on machine learning and statistical regression. The model is trained using signal data of multiple groups of known concentration samples, so that the model learns the internal correlation between signal characteristics and ion concentration. The self-learning update mechanism of the inversion model is introduced to solve the technical problem of the decrease of model accuracy with the change of environment in long-term operation, and to ensure the long-term stability of the detection system. The ICP-OES detection result in a specific period is taken as the benchmark, and the R 2 value is used to evaluate the current accuracy of the model; when the accuracy is lower than the preset threshold, the model is retrained using new true concentration data, so that the model adapts to long-term factors such as system aging and sample characteristic changes.

[0033] Preferably, the metal ion online detection method further comprises linkage control of process parameters of the to-be-detected system according to the detection result.

[0034] Preferably, the process parameters include but are not limited to stirring speed, feed ratio, dosing amount of extractant and / or stripping agent, etc.

[0035] In a second aspect, the present application provides a metal ion online detection system for the metal ion online detection method of the first aspect, which comprises a sample leading-out module, a detection module, an environmental parameter monitoring and control module, a signal acquisition and processing module, a concentration inversion and intelligent identification module, and a man-machine interaction module.

[0036] The present application solves the technical problem of lack of integrated equipment for metal ion online detection by integrating multiple functional modules to build a complete online detection system, realizes full-process automation from sample collection to result output, and greatly improves the detection efficiency and real-time performance.

[0037] As a preferred technical solution of the present application, the sample leading-out module comprises a sample leading-out pipeline, a micro pump, a detection window, and a backflow pipeline.

[0038] The micro pump drives the sample to be tested to enter the detection window through the outlet pipeline, and returns to the original system through the return pipeline after detection, forming a closed loop flow path; this design not only ensures real-time updating of the sample, but also avoids disturbance of sample loss to the original system.

[0039] As a preferred technical solution of the application, the detection module comprises a spectrum detector, which comprises a wide-spectrum light source, an optical lens, a light splitting component and a spectrum detector.

[0040] As a preferred technical solution of the application, the environmental parameter monitoring and control module comprises a sensor and a controller.

[0041] Preferably, the sensor comprises a temperature sensor and / or a humidity sensor.

[0042] Preferably, the controller comprises an actuator and a control drive.

[0043] Preferably, the actuator comprises a cooling fan.

[0044] Preferably, the control drive drives the actuator according to real-time parameters collected by the sensor.

[0045] The application solves the technical problem of environmental temperature and humidity fluctuation affecting detection accuracy by designing an environmental parameter monitoring and control module, and provides stable environmental conditions for the detection process, which effectively offsets the interference of environmental changes on the detection equipment and sample characteristics, and ensures the repeatability and reliability of the detection results.

[0046] As a preferred technical solution of the application, the signal acquisition and processing module comprises a main control platform, a serial communication interface, a display interface and an Ethernet communication module, the main control platform acquires detection data of the detection module in an Ethernet communication mode through the communication interface, processes the detection data, and inputs the concentration inversion and intelligent identification module through the communication interface.

[0047] Preferably, the concentration inversion and intelligent identification module comprises a concentration inversion model, which is used for real-time prediction of the concentration of rare earth ions and output of results.

[0048] Preferably, the man-machine interaction and system display module comprises a graphical user interface, which is connected with the signal acquisition and processing module and the concentration inversion and intelligent identification module, and is used for real-time display of a spectrum curve, a concentration change trend, an operation state and alarm information.

[0049] Preferably, a user inputs or adjusts system operation parameters through the graphical user interface.

[0050] The present application solves the technical problems of low data processing efficiency and inconvenient human-computer interaction by optimizing signal processing, inversion calculation and function design of human-computer interaction, improves the intelligence and ease of use of the system, improves the data processing efficiency and operation convenience of the system, and meets the real-time monitoring and flexible regulation and control requirements in industrial scenes.

[0051] Compared with the prior art, the present application has at least the following beneficial effects:

[0052] (1) The metal ion online detection method of the present application realizes dynamic online detection of metal ion concentration by constructing a continuous flow sample flow path, combining signal preprocessing, interference correction and real-time inversion calculation, and provides key data support for immediate regulation and control of extraction process;

[0053] (2) The present application further optimizes the interference correction of multiple model cooperation and the inversion model based on machine learning and statistical regression, effectively resists environmental and device interference, improves the detection accuracy and calculation efficiency, and the self-learning update mechanism of the inversion model ensures the stability of long-term operation of the system, avoiding the problem of accuracy decline with time;

[0054] (3) The detection system matched with the present application realizes full-process automation from sample collection to result output through the cooperation of multiple modules, combined with environmental control and intelligent interaction design, greatly improves the reliability and ease of use of the system, and can be directly applied to online monitoring of rare earth extraction process, improving the extraction efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 is a flowchart of the metal ion online detection method provided by the present application embodiment 1;

[0056] Figure 2 is a module connection logic diagram of the metal ion online detection system provided by the present application embodiment 1;

[0057] Figure 3 is a flowchart of step S104 of the metal ion online detection method provided by the present application embodiment 1;

[0058] Figure 4 is a flowchart of step S106 of the online concentration identification method provided by the present application embodiment 1;

[0059] Figure 5 is a constituent diagram of the online concentration identification system provided by the present application embodiment 1;

[0060] Figure 6 is a front-end operation interface schematic diagram of the online concentration identification system provided by the present application embodiment 1;

[0061] Figure 7 is a flowchart of the online concentration identification method provided by the present application embodiment 1;3+ ion concentration versus relative light intensity plot;

[0062] Figure 8 Pr is provided by the embodiment of the application 3+ ion concentration versus relative light intensity plot;

[0063] Figure 9 Dy is provided by the embodiment of the application 3+ ion concentration versus relative light intensity plot;

[0064] Figure 10 Tb is provided by the embodiment of the application 3+ ion concentration versus relative light intensity plot;

[0065] Figure 11 Fe is provided by the embodiment of the application 3+ ion concentration versus relative light intensity plot. DETAILED DESCRIPTION

[0066] In order to facilitate the understanding of the present application, the present application lists the following embodiments. It should be understood by those skilled in the art that the embodiments are only to help understand the present application, and should not be regarded as specific limitations of the present application.

[0067] Embodiment 1

[0068] The embodiment provides a metal ion online detection method as shown in the figure, comprising the following steps: Figure 1

[0069] S101, sample export, a water phase sample diversion path is arranged at the water phase extraction outlet of the last mixed settling tank in the multi-stage mixed settling tank in the rare earth extraction and purification process, a micro peristaltic pump and an acid-resistant capillary are used to guide part of the water phase feed liquid to a detection channel, and a continuous flow detection loop independent of the main extraction and purification process is constructed;

[0070] A detection window is arranged in the detection channel, the detection window has a leak-proof design, and the detection section in the detection loop is composed of high-transparency quartz on both sides, so as to ensure the real-time performance of spectral measurement and the stability of sample output, and avoid interference with the normal production process;

[0071] S102, spectral acquisition, the water phase sample entering the detection window is subjected to original spectral data acquisition by a spectral detection device, the spectral detection device comprises a wide-spectrum excitation light source, an LED array composed of LED lamp beads with different light-emitting wavelength ranges, a collimating lens, a spectral single slit, a linear array CCD and a temperature control component;

[0072] ​S103, signal preprocessing, the acquired original spectral data is denoised, baseline corrected and normalized, time sliding window is used for dynamic processing of spectral time series data, wavelet transform and Fourier transform are combined to decouple time-frequency characteristics, finally dynamic mode decomposition is used to extract main dynamic characteristic variables, specifically multiple DMD modes with different frequencies, decay rates and amplitudes, which are used as input data of the inversion model;

[0073] S104, feature extraction and concentration inversion, having offline and online processes, as shown in Figure 3 , this step includes offline and online processes:

[0074] The offline process is to build a standard database, a large number of rare earth ion standard samples with known concentrations are collected, and the actual concentrations are obtained by ICP-OES;

[0075] After the original spectral data is processed as described above, the features are extracted, and a concentration inversion model is established using statistical regression algorithms and machine learning algorithms, and R 2 , etc. are selected for evaluation, when R 2 ≥ 0.95, the model training is considered qualified, the meta-heuristic algorithm Bayesian optimization is used to optimize the performance of the model structure and parameters, and finally the trained concentration inversion model is output;

[0076] The online process is that the real-time collected spectral data is input into the above model after preprocessing, and the concentration recognition result is obtained, which is transmitted to the extraction process control system through Ethernet or serial communication interface at a frequency of 1~10 Hz, as a feedback variable to realize the linkage control with the process operation parameters.

[0077] S105, dynamic compensation and anti-interference processing, by setting temperature, humidity sensors and light source state monitoring circuit in the detection device, the environmental parameters and light source state are obtained in real time; a drift compensation model is established, the feature signal is normalized according to the drift trend, and the signal stability and concentration recognition accuracy of the system under complex working conditions are improved;

[0078] S106, algorithm deployment and data output, as shown in Figure 4As shown, the concentration inversion model is deployed in the embedded processing device Raspberry Pi 5, and the feature extraction, model reasoning and concentration calculation are completed locally; the concentration data is transmitted to the production line automation control system through RS232, Ethernet and other communication methods, the control system compares and judges the identified concentration value with the target value, when the concentration deviation in the extraction water phase exceeds ±3%, the linkage control operation is triggered, the stirring rate, feed ratio, control of the amount of stripping agent and the speed of the peristaltic pump in the extraction process are adjusted, when the phase separation performance is good, the system will automatically reduce the stirring paddle speed and increase the peristaltic pump speed to reduce the time cost and energy consumption; when the phase separation performance is poor, the system will increase the stirring paddle speed and reduce the peristaltic pump speed, while reducing the amount of extractant and feed liquid, so as to ensure the stability of each extraction process, improve the extraction quality and efficiency, verify the robustness and universality of the rare earth ion dynamic spectrum detection technology and its equipment by deploying the online monitoring system in the production line, and combine the energy-saving regulation feedback mechanism to significantly improve the precision and efficiency of the rare earth element extraction process, reduce energy consumption and production cost, and improve the extraction quality and yield;

[0079] S107, model self-learning and performance updating, the system sets a concentration identification accuracy threshold, when the R 2 < 0.99, trigger the model updating mechanism, automatically download the latest model parameters or trigger local retraining, complete the update of model structure and weight, and ensure the accuracy and robustness of the system under long-period operation;

[0080] As shown in Figure 7 The real-time rare earth ion concentration of the rare earth extraction system water phase obtained by the method is compared with the actual concentration tested by artificial ICP-OES, as shown in the figure, the detection result shows high accuracy and strong stability, and the correlation coefficient R 2 ≥ 0.997 in long-time test;

[0081] The spectrometer, temperature sensor and other devices are logically connected according to Figure 2 ;

[0082] The method is carried out by the metal ion online detection system, including the following modules:

[0083] S201, sample guiding and detection channel module

[0084] This module is used to shunt and construct an independent detection circuit from the liquid sample of the rare earth extraction and purification process. Its composition includes a water phase shunt pipeline, a micro peristaltic pump, a detection window and a reflux pipeline, through stable flow rate and corrosion-resistant design, the optical flux of the flowing sample entering the spectral detection window is ensured to be stable, and the detection accuracy and equipment service life are improved;

[0085] S202, spectrum acquisition module

[0086] The module is composed of a wide-spectrum light source, an optical collimating lens, a light splitting component, and a high-sensitivity detector, and is integrated into a miniature spectral detector. The module covers a wavelength band of 340-1100 nm, is suitable for detecting characteristic absorption peaks of various rare earth ions, and has the characteristics of rapid scanning, high resolution, and high signal-to-noise ratio. The sample can be scanned in the full spectrum of ultraviolet-visible-near infrared, and raw absorption spectrum data is output;

[0087] S203, embedded signal acquisition and processing module

[0088] The module is composed of an embedded main control platform, a serial communication interface, a display interface, and an Ethernet communication module. It is used for real-time acquisition, preprocessing, and concentration recognition calculation of spectral data, and uploads the concentration results to the control system through the communication interface. The concentration inversion algorithm program is embedded, and the reasoning process can be completed independently;

[0089] S204, concentration inversion and intelligent recognition module

[0090] This module deploys the concentration inversion model trained in advance and supports online model calling and updating mechanism. The model structure is constructed based on support vector regression, random forest, gradient boosting tree, and artificial neural network machine learning algorithm, further combined with Bayesian optimization for hyperparameter tuning, and cross-validation is used to evaluate the stability of the model to prevent overfitting, Figures 7-11 The system can be used for different concentrations of Nd 3+ , Pr 3+ , Dy 3+ , Tb 3+ , and impurity Fe 3+ ion samples. The collected specific light intensity absorption curve (ΔLight strength vs λ) shows that the concentration change has a significant correlation with the absorption peak intensity, verifying the technical reliability of the inversion model and the stability of the inversion algorithm;

[0091] S205, anti-interference and dynamic compensation module

[0092] For the interference factors such as light source drift, environmental temperature fluctuation, baseline offset, and coexisting ion disturbance, this module introduces humidity sensors, light source detection circuits, and other subunits for real-time monitoring, and combines wavelet transform and feature engineering algorithm for dynamic compensation processing, thereby improving the overall signal-to-noise ratio and recognition robustness;

[0093] S206, human-computer interaction and system display module

[0094] Through the mini HDMI interface external display or embedded graphical UI interface, the user can real-time view the spectral curve, rare earth ion concentration trend, device running state, and system alarm information,Figure 6 The right area shows typical control functions of the interactive interface, such as exposure adjustment, temperature and humidity display, spectrum preservation, etc.

[0095] To sum up, the application realizes real-time online accurate detection of multiple metal ions in the rare earth extraction and purification process by constructing a continuous flow detection loop independent of the main process, integrating wide-spectrum spectrum acquisition and multi-algorithm fusion signal processing system, carrying machine learning concentration inversion model, combining dynamic compensation and anti-interference mechanism, model self-learning update and process linkage control strategy, significantly improving the response speed of the extraction process, product quality stability and production efficiency, while reducing energy consumption and production cost, ensuring the robustness and universality of the system in long-period operation.

[0096] The applicant declares that the above is only a specific embodiment of the application, but the protection scope of the application is not limited thereto, and those skilled in the art should understand that any changes or replacements within the technical scope disclosed by the application can be easily thought of by any person skilled in the art in the technical field, and all fall within the protection scope and disclosure scope of the application.

Claims

1. A method for on-line detection of metal ions, characterized by, The metal ion online detection method comprises: (1) introducing the sample to be detected from the system to be detected and forming an independent continuous flow sample flow path for real-time detection; (2) collecting and detecting signals for pretreatment, interference correction processing and feature extraction to obtain input data for inversion calculation; (3) performing inversion calculation to output detection results in real time.

2. The method for on-line detection of metal ions according to claim 1, wherein, The detection in step (1) comprises spectral detection; Preferably, the spectral detection comprises absorption spectrum detection; Preferably, the system to be detected comprises a rare earth extraction system; Preferably, the sample to be detected comprises aqueous phase liquid and / or organic phase liquid of the rare earth extraction system.

3. The method for on-line detection of metal ions according to claim 1 or 2, characterized in that, The pretreatment in step (2) comprises baseline correction, denoising and normalization in sequence; Preferably, the interference correction processing in step (2) comprises dynamic compensation for changes in environmental and device parameters; Preferably, the changes in environmental and device parameters include temperature drift, humidity drift, spectral baseline drift, light source attenuation and environmental vibration.

4. The method for on-line detection of metal ions according to any one of claims 1 to 3, characterized in that, The inversion calculation in step (3) comprises inputting the input data in step (2) into an inversion model for calculation; Preferably, the inversion model is trained by machine learning algorithm and statistical regression algorithm calculation on input data of multiple groups of known concentration solutions; Preferably, the machine learning algorithm comprises any one of support vector regression, random forest regression, gradient boosting regression tree and artificial neural network regression; Preferably, the statistical regression algorithm comprises any one of multiple linear regression, ridge regression and principal component regression.

5. The method for on-line detection of metal ions according to any one of claims 1 to 4, characterized in that, The method further comprises self-learning update of the inversion model; Preferably, the self-learning updating comprises verifying the result of step (3) with the real concentration to calculate the decision coefficient R 2 when the value is lower than a preset threshold, optimizing the inversion model. Preferably, the optimization of the inversion model comprises repeatedly performing the inversion model training on the original inversion model.

6. A metal ion on-line detecting system for use in the metal ion on-line detecting method according to any one of claims 1 to 5, characterized by The metal ion online detection system comprises a sample introduction module, a detection module, an environmental parameter monitoring and control module, a signal acquisition and processing module, a concentration inversion and intelligent identification module and a man-machine interaction module.

7. The metal ion on-line monitoring system according to claim 6, wherein The sample introduction module comprises a sample introduction pipeline, a micro pump, a detection window and a reflux pipeline.

8. The metal ion on-line detection system according to claim 6 or 7, characterized in that, The detection module comprises a spectral detector, which comprises a wide-spectrum light source, an optical lens, a light splitting assembly and a spectral detector.

9. The metal ion on-line monitoring system according to any one of claims 6 to 8, wherein the metal ion on-line monitoring system is characterized by, The environmental parameter monitoring and control module comprises a sensor and a controller; Preferably, the sensor comprises a temperature sensor and / or a humidity sensor; Preferably, the controller comprises an actuator and a control drive; Preferably, the actuator comprises a cooling fan; Preferably, the control drive drives the actuator according to the real-time parameters collected by the sensor.

10. The metal ion on-line monitoring system according to any one of claims 6 to 9, wherein the metal ion on-line monitoring system is characterized by, The signal acquisition and processing module comprises a main control platform, a serial communication interface, a display interface and an Ethernet communication module, the main control platform collects detection data of the detection module in an Ethernet communication mode through the communication interface, processes the detection data and inputs the processed detection data into the concentration inversion and intelligent identification module through the communication interface; Preferably, the concentration inversion and intelligent identification module comprises a concentration inversion model, which is used for real-time prediction of rare earth ion concentration and output of results. Preferably, the human-computer interaction and system display module comprises a graphical user interface connected with the signal acquisition and processing module and the concentration inversion and intelligent identification module, and displays the spectral curve, concentration change trend, running state and alarm information in real time. Preferably, the user inputs or adjusts the system running parameters through the graphical user interface.

Images