General-purpose online real-time detection method, system and application for material dewatering process

By using an in-situ multi-parameter real-time acquisition system and a mathematical decoupling algorithm, real-time online detection of three types of moisture in chemical materials was achieved, solving the problems of detection lag and misjudgment in existing technologies and improving the accuracy and efficiency of the drying process.

CN123120084APending Publication Date: 2026-11-17SUZHOU TA&A ULTRA CLEAN TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611300133.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies cannot accurately distinguish and detect changes in the content of three types of moisture (adsorbed water, crystallization water, and reaction-generated water) in chemical materials in real time, leading to misjudgment of the drying process endpoint and energy waste.

Method used

An in-situ multi-parameter real-time acquisition system is adopted, which combines reference samples of pure adsorbed water, crystallized water and reaction-generated water for multi-parameter acquisition and calibration. Through multi-dimensional characteristic parameter threshold ranges and mathematical decoupling algorithms, real-time online detection and separation of three types of moisture in chemical materials are realized.

Benefits of technology

It enables real-time and accurate detection of three types of moisture in chemical materials, automatically determines the endpoint of the drying process, improves the versatility and accuracy of detection, and reduces energy waste and product defect rate.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123120084A_ABST
    Figure CN123120084A_ABST
Patent Text Reader

Abstract

This invention provides a universal method, system, and application for real-time online detection of material dehydration processes. The method includes: constructing an in-situ multi-parameter real-time acquisition system; using pure adsorbed water reference samples, crystallization water reference samples, and reaction product water reference samples for multi-parameter acquisition and calibration to obtain threshold ranges for multi-dimensional characteristic parameters; acquiring data from the material to be tested to obtain target multi-dimensional data, and decoupling and separating the target multi-dimensional data based on the threshold ranges of the multi-dimensional characteristic parameters to obtain the adsorbed water content, crystallization water content, and reaction product water content. This invention decomposes the total moisture in the system into three independent components: adsorbed water, crystallization water, and reaction product water. It utilizes three types of standard reference materials to complete universal database calibration, achieving cross-material applicability and overcoming the shortcomings of existing technologies, such as inability to separate three types of moisture, poor versatility, detection lag, and endpoint misjudgment, thus standardizing the detection of the drying process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of chemical material drying and physicochemical online detection technology, specifically involving a general-purpose material dehydration process real-time online detection method, system and application. In particular, it relates to a method for real-time monitoring and endpoint determination of the simultaneous separation of three types of water—adsorbed free water, lattice crystallization water and chemically generated water—in inorganic chemical raw materials, lithium battery raw materials, mineral salts, and solid-phase reaction systems under drying, decrystallization, and solid-phase chemical reaction conditions. It is applicable to online process quality control of industrial drying / reaction equipment such as batch reactors, rotary kilns, and fluidized beds. Background Technology

[0002] Moisture in chemical system materials is classified into three completely different categories according to its form of existence: Category 1: Adsorbed free water, which is physically attached to the surface and pores of powder particles without chemical bonds. Typical material: dry sodium chloride (NaCl), which contains only adsorbed water and is removed by volatilization upon heating at low temperature. The second type is lattice water of crystallization, in which water molecules are bound to the crystal lattice by coordination / ionic bonds. It can only be removed by reaching a specific pyrolysis temperature. Typical material: sodium sulfide nonahydrate Na2S·9H2O. Drying is divided into two stages: low-temperature desorption of adsorbed water and high-temperature segmented decrystalline water removal. The third type: water generated by chemical reactions. Newly generated combined water during the combination and metathesis reactions of solid / liquid solid materials is produced continuously with the reaction, and is simultaneously removed from the original adsorbed water and crystal water in the system. The three types of water coexist in the same system.

[0003] In industrial production, three common operating conditions are encountered: ① conventional drying of pure adsorbed water materials; ② decrystallization of hydrated salts; ③ simultaneous water production during reaction and dehydration of existing materials. The production end needs to monitor in real time: which type of water is being removed, the remaining proportion of each type of water, and whether the dehydration / reaction has reached its endpoint.

[0004] Traditional offline detection methods, including oven drying loss method, Karl Fischer titration, and TG-DSC thermogravimetric analysis, require stopping the machine for sampling and are conducted intermittently offline. They cannot continuously track the drying process in real time, have strong lag, and cannot be used for closed-loop control in continuous industrial production. Existing online detection methods (such as near-infrared, microwave dielectric, and exhaust gas humidity detection) can only obtain the total water content of the system and cannot perform tri-component separation of the total water, nor can they distinguish whether the water comes from adsorption / desorption, lattice dissociation, or chemical reaction. In scenarios where the total water content remains unchanged but crystal water continues to be released and water is produced simultaneously in the reaction, existing instruments will completely misjudge. At the same time, in existing online detection methods, a set of calibration parameters is only suitable for a single raw material. Changing the material (from NaCl to sodium hydrate sulfide, or reaction raw materials) requires recalibrating the entire process. There is no universal detection logic, and it is impossible to form a unified judgment standard. It is also impossible to decouple and separately determine the "end point of adsorbed water drying, the end point of complete removal of crystal water, and the end point of chemical reaction transformation", which can easily lead to over-drying of materials, wasted energy, or substandard drying with excessive water content. Summary of the Invention

[0005] The main objective of this invention is to provide a general-purpose method, system, and application for real-time online detection of material dehydration process, in order to overcome the shortcomings of the prior art.

[0006] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a general-purpose online real-time detection method for the dehydration process of materials that can distinguish between adsorbed water, crystallization water, and reaction-generated water, comprising: An in-situ multi-parameter real-time acquisition system is constructed. The multi-dimensional data acquired by the in-situ multi-parameter real-time acquisition system includes internal temperature, exhaust gas water vapor partial pressure, material dielectric parameters or near-infrared characteristic spectral parameters, and real-time weight loss rate. Multi-parameter acquisition and calibration were performed using pure adsorbed water reference samples, crystallization water reference samples, and reaction product water reference samples to obtain the threshold range of multi-dimensional characteristic parameters. Data is collected from the material to be tested to obtain target multidimensional data. Based on the threshold range of the multidimensional feature parameters, the target multidimensional data is decoupled and separated to obtain the adsorbed water content, crystallization water content, and reaction product water content. The target multidimensional feature obtained by weighted fitting of the adsorbed water content, crystallization water content, and reaction product water content matches the threshold range of the multidimensional feature parameters.

[0007] Secondly, the present invention also provides a universal material dehydration process real-time online detection device for distinguishing adsorbed water, crystallization water, and reaction-generated water, used to implement the above-mentioned universal material dehydration process real-time online detection method, comprising: The in-situ multi-parameter real-time acquisition system acquires multi-dimensional data including internal temperature, exhaust gas water vapor partial pressure, material dielectric parameters or near-infrared characteristic spectral parameters, and real-time weight loss rate. And, a data processing system, used for: Perform multi-parameter acquisition and calibration to obtain the threshold range of multi-dimensional feature parameters; The test material is subjected to data acquisition to obtain target multidimensional data. Based on the threshold range of the multidimensional feature parameters, the target multidimensional data is decoupled and separated to obtain the adsorbed water content, crystallization water content and reaction product water content. The target multidimensional feature obtained by weighted fitting of the adsorbed water content, crystallization water content and reaction product water content matches the threshold range of the multidimensional feature parameters.

[0008] Thirdly, the present invention also provides the application of the above-mentioned general material dehydration process real-time online detection method in industrial inorganic raw material drying processes.

[0009] Compared with the prior art, the beneficial effects of the present invention include at least the following: The technical solution provided by this invention relies on in-situ synchronous acquisition of multiple physical parameters, pre-calibration of three types of moisture characteristic databases, and mathematical decoupling algorithms to decompose the total moisture of the system into three independent components: adsorbed water, crystallization water, and reaction-generated water. The content changes of each component are quantified in real time, and their respective dehydration progress curves are plotted. The process endpoint is automatically determined segment by segment, and the database is universally calibrated using three types of standard reference materials. This method is universal across materials, overcoming the shortcomings of existing technologies such as inability to separate the three types of moisture, poor universality, detection lag, and endpoint misjudgment. It provides a universal detection method that is not limited to any type of material, decomposes the content and desorption rate of adsorbed water, crystallization water, and reaction-generated water in real time in situ, outputs dehydration progress curves for each of the three types of substances, and independently determines the endpoint of each of the three processes, thus achieving standardization of the drying process detection.

[0010] The above description is merely an overview of the technical solution of the present invention. In order to enable those skilled in the art to better understand the technical means of this application and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described below in conjunction with detailed drawings. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1This is a schematic diagram of the overall process of a general-purpose online detection method for material dehydration process provided in a typical embodiment of the present invention. Detailed Implementation

[0013] In view of the shortcomings of the prior art, the inventors of this invention, through long-term research and extensive practice, have proposed the technical solution of this invention. The following will further explain and illustrate the technical solution, its implementation process, and its principles.

[0014] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0015] This invention first provides a general method for real-time online detection of material dehydration process that distinguishes between adsorbed water, crystallization water, and reaction-generated water, comprising the following steps: An in-situ multi-parameter real-time acquisition system is constructed. The multi-dimensional data acquired by the in-situ multi-parameter real-time acquisition system includes internal temperature, exhaust gas water vapor partial pressure, material dielectric parameters or near-infrared characteristic spectral parameters, and real-time weight loss rate. Multi-parameter acquisition and calibration were performed using pure adsorbed water reference samples, crystallization water reference samples, and reaction product water reference samples to obtain the threshold range of multi-dimensional characteristic parameters. Data is collected from the material to be tested to obtain target multidimensional data. Based on the threshold range of the multidimensional feature parameters, the target multidimensional data is decoupled and separated to obtain the adsorbed water content, crystallization water content, and reaction product water content. The target multidimensional feature obtained by weighted fitting of the adsorbed water content, crystallization water content, and reaction product water content matches the threshold range of the multidimensional feature parameters.

[0016] In some implementations, the threshold range of the multidimensional characteristic parameters includes the weightlessness temperature window, weightlessness rate, weightlessness curve morphology, water vapor partial pressure change rate, dielectric parameter change rate, or near-infrared characteristic change rate.

[0017] In some implementations, the method specifically includes the following steps: Calculate the degree of multidimensional matching between the target multidimensional data and the adsorbed water content, crystallization water content, and reaction product water content, respectively. Based on the multidimensional matching degree, establish a set of mapping equations between the adsorbed water content, crystallization water content, and reaction product water content and the internal temperature, tail gas water vapor partial pressure, material dielectric parameters, and / or near-infrared characteristic spectral parameters, respectively. Solve the system of mapping equations to determine the adsorbed water content, crystallization water content, and reaction product water content.

[0018] In some implementations, the calculation process for the multidimensional matching degree is expressed as follows: M ad (t)=α1*f T (T)+α2*f P (P)+α3*f ε (ε)+α4*f W (W) M cr (t)=β1*g T (T)+β2*g P (P)+β3*g ε (ε)+β4*g W (W) M rxn (t)=γ1*h T (T)+γ2*h P (P)+γ3*h ε (ε)+γ4*h W (W) Among them, M ad M represents the degree of adsorption water matching. cr M represents the degree of water binding compatibility. rxn The reaction produces water with a matching degree of t, where t represents time, T represents the rate of temperature change, and f represents the water production matching degree. T ,g T ,h T The membership function represents the temperature range, P represents the rate of change of water vapor partial pressure, and f represents the membership function. P ,g P ,h P The membership function represents the rate of change of water vapor partial pressure, ε represents the rate of change of dielectric parameters or near-infrared characteristic rate, and f ε ,g ε ,h ε The membership function represents the rate of change of dielectric parameters or the rate of change of near-infrared characteristic; W represents the rate of weight loss; f W ,g W ,h W Membership function of weightlessness rate, α i ,β i ,γ i This represents the weight coefficients calibrated based on the threshold interval of the multidimensional feature parameters, satisfying ∑α i =∑β i =∑γ i =1, where i=1 / 2 / 3 / 4.

[0019] In this invention, the membership function preferably adopts the trapezoidal membership function, which is well known in the art, but it is not limited to this. Other types of membership functions with similar functions can also achieve the same purpose. The four parameters [a, b, c, d] of the trapezoidal membership function are determined based on the threshold range of the measured multidimensional characteristic parameters of the calibrated samples in the S2 reference library: parameters a and d are obtained by extrapolation from the minimum and maximum values ​​of the moisture characteristic parameters of this type in the calibration data, respectively, and parameters b and c are determined by the lower and upper boundaries of the stable existence of the moisture characteristic parameters of this type in the calibration data, respectively.

[0020] A statistical calibration method based on the threshold interval of feature parameters is adopted, and the specific method is as follows: For each type of water (adsorbed water, crystallized water, reaction-formed water) and each type of characteristic parameter (temperature, partial pressure of water vapor, dielectric / near-infrared, weight loss rate), the numerical boundaries of the parameter in "completely not belonging to this type of water," "partially belonging to," and "completely belonging to" are determined using measured data from calibrated samples in the S2 reference library. Parameter a (left foot): When the value of this characteristic parameter is lower than this value, it definitely does not belong to this type of water, and the membership degree is 0. Basis for value: The minimum characteristic value of this type of water that did not appear in the calibration experiment (or this value minus a safety margin).

[0021] Parameter b (left shoulder): When this characteristic parameter value is reached, the water begins to fully belong to this type of water, and the membership degree rises to 1. Value basis: The lower boundary value at which the characteristic parameter of this type of water begins to stably appear in the calibration experiment.

[0022] Parameter c (right shoulder): When the value of this characteristic parameter is higher than this value, the water completely belongs to this type of water, and the membership degree decreases from 1. Value basis: The upper boundary value of the stable existence of this type of water characteristic parameter in the calibration experiment.

[0023] Parameter d (right foot): When the value of this characteristic parameter is higher than this value, it definitely does not belong to this type of water, and the membership degree is 0. The basis for the value is: the maximum characteristic value of this type of water that has not appeared again in the calibration experiment (or the value plus a safety margin).

[0024] Specific example (taking the "temperature range membership function" fT of adsorbed water as an example): Based on the calibration data of pure adsorbed water reference samples (NaCl) in the S2 reference library, the characteristic removal temperature window for adsorbed water is room temperature to 150℃, therefore: a=20 (below room temperature, definitely not within the temperature range for adsorbed water removal); b=40 (above 40℃, it enters the "perfect match" range for adsorbed water removal); c=130 (below 130℃, it is still within the "perfect match" range for adsorbed water removal); d=180 (above 160℃, definitely not within the temperature range for adsorbed water removal).

[0025] The values ​​above are for illustrative purposes only; the actual parameters will be determined based on specific calibration experimental data.

[0026] Similarly, the parameters of the membership function of the temperature range of water of crystallization are determined based on the calibration data of sodium sulfide nonahydrate (e.g., characteristic temperature range 210–240℃); the parameters of the membership function of the temperature range of water produced by the reaction are determined based on the calibration data of the calibrated reaction system (e.g., reaction temperature 100–400℃).

[0027] In some implementations, the mapping equations are expressed as: W total (t)=W ad (t)+W cr (t)+W rxn (t) M ad (t)=Φ ad [W ad (t),T(t),P(t),ε(t)] M cr (t)=Φ cr [W cr (t),T(t),P(t),ε(t)] M rxn (t)=Φ rxn [W rxn (t),T(t),P(t),ε(t)] Among them, W total W represents total weightlessness, which is equal to the integral of the real-time weightlessness rate at time 0-t. ad W represents the weight loss of adsorbed water. cr W represents the weight loss of bound water. rxn Φ indicates the weight loss of water produced in the reaction. ad Φ cr Φ rxn These represent the inverse function mapping relationships of adsorbed water, bound water, and reaction-generated water based on the aforementioned construction.

[0028] In some implementations, the α i ,β i ,γ i The calibration process specifically includes: Water classification was performed separately using multiple data dimensions to obtain the classification accuracy R. T R P R ε R W , where R · Indicates accuracy rate; Calculate the initial weighting coefficients:

[0029] Where R ad,i / j This represents the single-parameter correct discrimination rate of the i-th or j-th feature parameter for adsorbed water, where j = 1 / 2 / 3 / 4; The initial weighting coefficients were verified and fine-tuned using standard samples.

[0030] As a typical example, the weighting coefficient α i ,β i ,γ i (i = 1, 2, 3, 4) reflects the contribution of four types of characteristic parameters (temperature, water vapor partial pressure, dielectric / near-infrared, and weight loss rate) to the identification of various moisture types. The calibration process is as follows: Step 1: Single-parameter sensitivity test For each type of reference sample in the reference library, given its moisture composition, the moisture type was determined individually using four types of characteristic parameters, and the correct discrimination rate R for each parameter was recorded. T ,R P ,R ,R W .

[0031] Step 2: Normalized weight calculation Calculate the weighting coefficients for adsorbed water (ad), crystallization water (cr), and reaction-generated water (rxn) respectively:

[0032] Where Rad,i represents the single-parameter correct discrimination rate of the i-th feature parameter for adsorbed water, and Rad,j represents the single-parameter correct discrimination rate of the j-th feature parameter for adsorbed water.

[0033] Step 3: Weight Verification and Fine-tuning The calibrated weighting coefficients are substituted into the matching degree calculation formula, and all calibration data of the three types of standard samples in the benchmark library are back-substituted for verification. If the average deviation between the matching degree calculation result of a certain type of moisture and its known moisture content exceeds a preset threshold (e.g., 5%), the weighting coefficient of that type of moisture is fine-tuned (using the gradient descent method or manual experience fine-tuning) until the accuracy requirements are met.

[0034] In some implementations, the Φ ad Φ cr Φ rxn The construction process specifically includes: During the drying process of the reference sample, obtain the true value label of the moisture content; Plot the correlation curve between the truth labels and the matching degree, and fit the inverse function mapping relationship. The inverse function mapping relationship is cross-validated and corrected.

[0035] As a typical example, the inverse function mapping relation Φ ad Φ cr Φ rxn The construction process is as follows: Step 1: Baseline Data Acquisition For each type of standard sample in the reference library, the following data were recorded simultaneously at different times during the calibration process: Known content W of various types of moisture ad true (t),W ad true (t),W ad true (t) (which can be obtained through offline thermogravimetric analysis or chemometric calculations); And the corresponding matching degree calculation value M ad (t),M cr (t),M rxn (t).

[0036] Step 2: Scatter plotting and curve fitting Scatter plots were drawn for the three types of moisture: the horizontal axis represents the matching degree M(t) (0-1), and the vertical axis represents the corresponding moisture content W(t) (wt%). Nonlinear curve fitting was performed on the scatter plots to obtain Φ. ad Φ cr Φ rxn The function expression.

[0037] Taking adsorbed water as an example, the fitted mapping relationship is as follows: W ad =Φ ad (M ad )=k1 M ad k2 +k3 Where k1, k2, and k3 are fitting parameters, obtained by fitting the calibration data of the benchmark library using the least squares method.

[0038] Step 3: Cross-validation Substitute the constructed Φ function into the validation dataset in the benchmark library that was not used for fitting, and check the back-calculation accuracy. If the accuracy does not meet the requirements (e.g., the root mean square error with the true value > 0.2wt%), adjust the form of the fitting model (e.g., use polynomial fitting, piecewise fitting, etc.), and repeat step 2 until the accuracy requirements are met.

[0039] In some implementations, the method further includes the following steps: Based on the adsorbed water content, crystallization water content, and reaction product water content, curves showing the change of remaining adsorbed water over time, the change of crystallization water removal progress, and the accumulation-removal equilibrium curve of reaction product water were plotted.

[0040] In some implementations, the method further includes the following steps: Identify and output based on preset process thresholds: The endpoint of adsorbed water drying: the adsorbed water content is lower than the process-defined threshold. Endpoint for water of crystallization removal: water of crystallization removal ratio ≥ set requirement; Chemical reaction endpoint: The rate of water production in the reaction approaches zero and the system reaches equilibrium after water removal.

[0041] For a typical implementation example of the above method, see [link to relevant documentation]. Figure 1 As shown, the complete operation flow of the method provided by this invention may include: S1. Real-time in-situ multi-parameter data acquisition under closed working conditions In the closed environment of the drying chamber / reactor, four types of real-time time-series data are collected continuously and synchronously throughout the entire process: ① internal temperature of the system; ② partial pressure of water vapor in the exhaust gas of the chamber; ③ dielectric parameters / near-infrared characteristic spectral parameters of the powder material; ④ real-time weight loss rate of the system; the sampling period is adjustable from 1 to 30 seconds, and offline sampling is not required throughout the process.

[0042] S2. Construct a benchmark database of three types of water-specific characteristics. Three types of standard samples were selected for calibration, and threshold ranges for multidimensional characteristic parameters were established. Specific data types included:

[0043] 1) Pure adsorbed water reference sample: Industrial sodium chloride powder can be used, containing only particulate adsorbed water. The characteristics of adsorbed water are calibrated as follows: rapid volatilization in the low temperature range, short-term surge in water vapor partial pressure, weight loss concentrated in the low temperature range, and dielectric parameters rapidly decreasing with surface water. 2) Crystallization water reference sample: Sodium sulfide nonahydrate can be used to determine the characteristics of crystallization water: lattice breakage and water release in a specific critical temperature range, stepwise abrupt weight loss, and concentrated release of water vapor at characteristic temperature levels; 3) Reaction water production reference sample: solid reaction material (with a small amount of adsorbed water + trace amount of crystal water + chemically combined water continuously generated by the reaction) was selected to calibrate the characteristics of the reaction water: the water production rate is linearly related to the material conversion rate, the water vapor increment increases slowly and continuously with the reaction progress, and there is no concentrated burst release characteristic.

[0044] The database covers a temperature range of ambient temperature to 500℃, commonly used in industrial drying processes.

[0045] S3, Multi-parameter coupling algorithm for decoupling and separating three types of moisture content The real-time data collected by S1 is imported into the preset fitting model. Based on the feature boundary of the S2 benchmark library, mathematical decoupling and separation are performed, and three data points are output separately in real time: the instantaneous remaining adsorbed water mass fraction, the proportion of removed crystal water to the original crystal water, and the total amount of water generated by the cumulative chemical reaction. The three components are quantitatively separated from the total amount of water.

[0046] The 'preset fitting model' described in this invention is a rule-based discrimination model based on the threshold range of three types of moisture characteristic parameters, and is not an AI model involving machine learning or artificial intelligence algorithms.

[0047] Its specific operational logic is as follows: Step 1: Feature Window Matching. For each sampling time t, the four types of parameter values ​​are compared with the characteristic parameter threshold ranges of the three types of water in the S2 benchmark library to determine which type of water the system is in at the current time (multiple types may overlap).

[0048] Step 2: Contribution Allocation. When the three types of moisture characteristic intervals overlap (such as the three-moisture coexistence scenario in Example 3), a weighted allocation method based on feature differences is adopted: The main contributions of adsorbed water are: peak weight loss rate in the low temperature region and rapid decrease in dielectric parameters; The contribution of water of crystallization is mainly reflected in: stepwise weight loss and concentrated release of water vapor partial pressure within a specific temperature window; The contribution of water generated by the reaction is mainly reflected in: a continuous and gradual increase in water vapor, and a linear correlation with the reaction progress.

[0049] Step 3: Solve the simultaneous equations. Using the above formulas and real-time total weight loss data, calculate the real-time content of each of the three types of water.

[0050] Its core algorithm is a multi-parameter weighted feature matching algorithm, as follows: Let the four parameters collected in real time be: temperature T(t), water vapor partial pressure P(t), dielectric parameter ε(t), and weight loss rate W(t). For each time t, calculate the matching degree between each parameter and the threshold range of the three types of moisture characteristics in the S2 benchmark library: Adsorbed water matching degree: M ad (t)=α1*f T (T)+α2*f P (P)+α3*f ε (ε)+α4*f W (W) Water of crystallization matching degree: M cr (t)=β1*g T (T)+β2*g P (P)+β3*g ε (ε)+β4*gW (W) Reaction water compatibility: M rxn (t)=γ1*h T (T)+γ2*h P (P)+γ3*h ε (ε)+γ4*h W (W) Where f T ,g T ,h T The membership function for the temperature range (using a trapezoidal membership function), f P ,g P ,h P f is the membership function of the rate of change of water vapor partial pressure. ε ,g ε ,h ε f is the membership function of the rate of change of dielectric parameter. W ,g W ,h W α is the membership function for the rate of weightlessness. i ,β i ,γ i The weighting coefficients for each parameter (obtained based on the S2 benchmark library calibration, satisfying ∑α) i =∑β i =∑γ i =1).

[0051] The real-time concentrations of the three types of water at each time point were solved using the following simultaneous equations: W total (t)=W ad (t)+W cr (t)+W rxn (t) M ad (t)=Φ ad [W ad (t),T(t),P(t),ε(t)] M cr (t)=Φ cr [W cr (t),T(t),P(t),ε(t)] M rxn (t)=Φ rxn [W rxn (t),T(t),P(t),ε(t)] Among them W total (t) is obtained by integrating the real-time rate of weightlessness, Φ ad ,Φ cr ,Φ rxn This is an inverse function mapping relationship built based on the S2 benchmark library.

[0052] S4, Grouping and generating independent dehydration process curves Plot the following curves: ① the remaining amount of adsorbed water versus time; ② the rate of crystallization water removal; ③ the equilibrium curve of water accumulation and removal from the reaction. Visualize the conversion progress of the three types of materials.

[0053] S5, Automatic Threshold Determination of Multiple Process Endpoints Based on the preset process threshold, the system automatically identifies and outputs: 1) End point of adsorbed water drying: the adsorbed water content is lower than the process limit threshold; 2) End point for water of crystallization removal: The proportion of water of crystallization removed is greater than or equal to the set requirement; 3) Chemical reaction endpoint: The rate of water production in the reaction approaches zero, and the system reaches equilibrium after water removal.

[0054] The material characteristic parameters can be selected from two options: dielectric constant (low-cost industrial application) or near-infrared spectroscopy (high-precision laboratory scenarios); compatible equipment includes, but is not limited to, three mainstream chemical drying equipment: intermittent stirred reactor, rotary drum dryer, and fluidized bed calciner; the detected data can be synchronously connected to the equipment's built-in DCS control system to achieve automated linkage control of drying temperature and air intake.

[0055] This invention also provides a universal material dehydration process real-time online detection device that can distinguish between adsorbed water, crystallization water, and reaction-generated water, used to implement the universal material dehydration process real-time online detection method provided in any of the above embodiments, comprising: The in-situ multi-parameter real-time acquisition system acquires multi-dimensional data including internal temperature, exhaust gas water vapor partial pressure, material dielectric parameters or near-infrared characteristic spectral parameters, and real-time weight loss rate. And, a data processing system, used for: Perform multi-parameter acquisition and calibration to obtain the threshold range of multi-dimensional feature parameters; The test material is subjected to data acquisition to obtain target multidimensional data. Based on the threshold range of the multidimensional feature parameters, the target multidimensional data is decoupled and separated to obtain the adsorbed water content, crystallization water content and reaction product water content. The target multidimensional feature obtained by weighted fitting of the adsorbed water content, crystallization water content and reaction product water content matches the threshold range of the multidimensional feature parameters.

[0056] This invention also provides the application of the above-mentioned general material dehydration process real-time online detection method in industrial inorganic raw material drying processes.

[0057] The technical solution of the present invention will be further described in detail below through several embodiments and in conjunction with the accompanying drawings. However, the selected embodiments are only for illustrating the present invention and do not limit the scope of the present invention.

[0058] Example 1: Hot air drying test containing only adsorbed water and sodium chloride Experimental conditions: Industrial NaCl powder, initial adsorbed water 5.18wt%, drying temperature 60-130℃, fluidized bed operation, sampling cycle 30s.

[0059] According to the method provided by this invention, temperature, exhaust gas water vapor, dielectric and weight loss data are collected in real time. After model decoupling, only the adsorbed water component is detected. The real-time curve shows that the adsorbed water continues to decrease with drying. When the adsorbed water drops to 0.15wt%, the system automatically determines the end point of adsorbed water drying. The relative error with offline thermogravimetric detection data is <0.2%.

[0060] Table 1-1: Changes in adsorbed water content over time during the NaCl drying process

[0061] Table 1-2: Comparison and verification of the method of this invention with offline thermogravimetric analysis

[0062] Example 2: Staged Dehydration Test of Sodium Sulfide Nonahydrate Containing Adsorbed Water + Crystallization Water Experimental conditions: Na2S·9H2O, initial adsorbed water 3.2wt%, theoretical crystal water content = 9×18.015 / 240.18 × 100% =67.5wt% (molecular weight 240.18), drying process: low temperature desorption of adsorbed water at 80℃ → high temperature desorption of crystal water at 210~240℃.

[0063] The decoupling and separation of the two components: the rapid downward curve in the 80℃ range corresponds to the removal of adsorbed water, and the step-like weight loss curve after reaching 210℃ corresponds to the breakage and release of crystal lattice; the first endpoint is output when the adsorbed water meets the standard, and the second endpoint is output when the crystal water removal rate is 99.5%; the two dehydration stages can be clearly distinguished, which cannot be distinguished by conventional total water detection.

[0064] Table 2-1: Total weight loss and water separation decoupling data during the drying process of sodium sulfide nonahydrate

[0065] Table 2-2: Weight loss rate and corresponding moisture type in each temperature range

[0066] Example 3: Solid-phase synthesis reaction system with three waters coexisting The raw material powder contains adsorbed water and trace amounts of crystal water. After the materials are mixed, a solid-phase metathesis reaction occurs to continuously generate chemically combined water, with the three types of water existing simultaneously. Experimental conditions: Battery-grade LiOH·H2O, adsorbed water 1.2wt%, theoretical content of water of crystallization = 1×18.015 / 41.96 × 100% =42.9wt% (LiOH·H2O molecular weight 41.96); H2S gas is introduced, reaction temperature 373~673K (100~400℃).

[0067] This method calculates three independent curves: the original adsorbed water gradually decreases, the crystal water is released stepwise at a specific temperature, and the reaction-generated water accumulates slowly with the reaction. When the rate of increase in reaction-generated water reaches zero, the system simultaneously determines the endpoint of the chemical reaction and simultaneously checks the remaining crystal water and adsorbed water indicators, thus achieving dual endpoint control of reaction and drying.

[0068] Table 3-1: Simultaneous monitoring data of three types of moisture during the reaction process of LiOH·H2O and H2S

[0069] Key data points analysis:

[0070] Table 3-2: Mass balance verification of each component during the reaction process

[0071] Based on the above embodiments, it is clear that the technical solution provided by the present invention has at least the following advantages: 1. High versatility: A single calibration database and detection logic can simultaneously adapt to three types of industrial scenarios: drying with only adsorbed water, drying with decrystallized water, and coexistence of reaction product water and multiple forms of water. No need to remodel and calibrate when changing materials, and a unified industry detection and judgment standard can be formed. It should be noted that the characteristic temperature windows and absolute values ​​of weight loss rates for the three types of water may differ in different materials (e.g., different water binding energies). However, the characteristic curve morphology of the three types of water in the three-dimensional space of temperature-weight loss rate-water vapor partial pressure exhibits a universal pattern: adsorbed water always shows rapid volatilization in the low-temperature region (sharp peak), crystal water always shows a step-like burst release at a specific temperature level (step-like), and reaction-generated water always shows a gradual and continuous release related to the reaction progress (gentle slope-like). This invention utilizes the curve morphology characteristics rather than absolute temperature values ​​for pattern recognition, thus possessing cross-material applicability. 2. Real-time online non-destructive testing: The entire process is sealed and in-situ, without sampling or disrupting the production process, with no detection delay, making it suitable for continuous industrial production. 3. Refined water separation: Precisely separate three types of water from total water, accurately distinguishing whether the current process is in the stage of removing surface water, removing lattice water, or producing water from the reaction, thus preventing over-drying or under-drying. 4. Precise endpoint determination in stages: The endpoints of the three processes are determined independently, optimizing production energy consumption and product qualification rate, and facilitating standardized management and control of lithium battery raw materials and inorganic salt drying workshops.

[0072] It should be understood that the above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A general-purpose online real-time detection method for the dehydration process of materials, distinguishing between adsorbed water, crystallization water, and reaction-generated water, characterized in that, include: An in-situ multi-parameter real-time acquisition system is constructed. The multi-dimensional data acquired by the in-situ multi-parameter real-time acquisition system includes internal temperature, exhaust gas water vapor partial pressure, material dielectric parameters or near-infrared characteristic spectral parameters, and real-time weight loss rate. Multi-parameter acquisition and calibration were performed using pure adsorbed water reference samples, crystallization water reference samples, and reaction product water reference samples to obtain the threshold range of multi-dimensional characteristic parameters. Data is collected from the material to be tested to obtain target multidimensional data. Based on the threshold range of the multidimensional feature parameters, the target multidimensional data is decoupled and separated to obtain the adsorbed water content, crystallization water content, and reaction product water content. The target multidimensional feature obtained by weighted fitting of the adsorbed water content, crystallization water content, and reaction product water content matches the threshold range of the multidimensional feature parameters.

2. The general-purpose online detection method for material dehydration process according to claim 1, characterized in that, The threshold ranges for multidimensional characteristic parameters include the weightlessness temperature window, weightlessness rate, weightlessness curve shape, water vapor partial pressure change rate, dielectric parameter change rate, or near-infrared characteristic change rate.

3. The general-purpose online detection method for material dehydration process according to claim 1, characterized in that, Specifically, it includes: Calculate the degree of multidimensional matching between the target multidimensional data and the adsorbed water content, crystallization water content, and reaction product water content, respectively. Based on the multidimensional matching degree, establish a set of mapping equations between the adsorbed water content, crystallization water content, and reaction product water content and the internal temperature, tail gas water vapor partial pressure, material dielectric parameters, and / or near-infrared characteristic spectral parameters, respectively. Solve the system of mapping equations to determine the adsorbed water content, crystallization water content, and reaction product water content.

4. The general-purpose online detection method for material dehydration process according to claim 3, characterized in that, The calculation process for the multidimensional matching degree is expressed as follows: M ad (t)=α1*f T (T)+α2*f P (P)+α3*f ε (ε)+α4*f W (W) M cr (t)=β1*g T (T)+β2*g P (P)+β3*g ε (ε)+β4*g W (W) M rxn (t)=γ1*h T (T)+γ2*h P (P)+γ3*h ε (ε)+γ4*h W (W) Among them, M ad M represents the degree of adsorption water matching. cr M represents the degree of water binding compatibility. rxn The reaction produces water with a matching degree of t, where t represents time, T represents the rate of temperature change, and f represents the water production matching degree. T ,g T ,h T The membership function represents the temperature range, P represents the rate of change of water vapor partial pressure, and f represents the membership function. P ,g P ,h P The membership function represents the rate of change of water vapor partial pressure, ε represents the rate of change of dielectric parameters or near-infrared characteristic rate, and f ε ,g ε ,h ε The membership function represents the rate of change of dielectric parameters or the rate of change of near-infrared characteristic; W represents the rate of weight loss; f W ,g W ,h W Membership function of weightlessness rate, α i ,β i ,γ i This represents the weight coefficients calibrated based on the threshold interval of the multidimensional feature parameters, satisfying ∑α i =∑β i =∑γ i =1, where i=1 / 2 / 3 / 4.

5. The general-purpose online detection method for material dehydration process according to claim 4, characterized in that, The system of mapping equations is expressed as follows: W total (t)=W ad (t)+W cr (t)+W rxn (t) M ad (t)=Φ ad [W ad (t),T(t),P(t),ε(t)] M cr (t)=Φ cr [W cr (t),T(t),P(t),ε(t)] M rxn (t)=Φ rxn [W rxn (t),T(t),P(t),ε(t)] Among them, W total W represents total weightlessness, which is equal to the integral of the real-time weightlessness rate at time 0-t. ad W represents the weight loss of adsorbed water. cr W represents the weight loss of bound water. rxn Φ indicates the weight loss of water produced in the reaction. ad Φ cr Φ rxn These represent the inverse function mapping relationships of adsorbed water, bound water, and reaction-generated water based on the aforementioned construction.

6. The general-purpose online detection method for material dewatering process according to claim 5, characterized in that, α i ,β i ,γ i The calibration process specifically includes: Water classification was performed separately using multiple data dimensions to obtain the classification accuracy R. T R P R ε R W , where R · Indicates accuracy rate; Calculate the initial weighting coefficients: ; Where R ad,i / j This represents the single-parameter correct discrimination rate of the i-th or j-th feature parameter for adsorbed water, where j = 1 / 2 / 3 / 4; The initial weighting coefficients were verified and fine-tuned using standard samples; And / or, Φ ad Φ cr Φ rxn The construction process specifically includes: During the drying process of the reference sample, obtain the true value label of the moisture content; Plot the correlation curve between the truth labels and the matching degree, and fit the inverse function mapping relationship. The inverse function mapping relationship is cross-validated and corrected.

7. The general-purpose online detection method for material dehydration process according to claim 1, characterized in that, Also includes: Based on the adsorbed water content, crystallization water content, and reaction product water content, curves showing the change of remaining adsorbed water over time, the change of crystallization water removal progress, and the accumulation-removal equilibrium curve of reaction product water were plotted.

8. The general-purpose online detection method for material dehydration process according to claim 1, characterized in that, Also includes: Identify and output based on preset process thresholds: The endpoint of adsorbed water drying: the adsorbed water content is lower than the process-defined threshold. Endpoint for water of crystallization removal: water of crystallization removal ratio ≥ set requirement; Chemical reaction endpoint: The rate of water production in the reaction approaches zero and the system reaches equilibrium after water removal.

9. A universal material dehydration process real-time online detection device for distinguishing adsorbed water, crystallization water, and reaction-generated water, used to implement the universal material dehydration process real-time online detection method according to any one of claims 1-8, characterized in that, include: The in-situ multi-parameter real-time acquisition system acquires multi-dimensional data including internal temperature, exhaust gas water vapor partial pressure, material dielectric parameters or near-infrared characteristic spectral parameters, and real-time weight loss rate. And, a data processing system, used for: Perform multi-parameter acquisition and calibration to obtain the threshold range of multi-dimensional feature parameters; The test material is subjected to data acquisition to obtain target multidimensional data. Based on the threshold range of the multidimensional feature parameters, the target multidimensional data is decoupled and separated to obtain the adsorbed water content, crystallization water content and reaction product water content. The target multidimensional feature obtained by weighted fitting of the adsorbed water content, crystallization water content and reaction product water content matches the threshold range of the multidimensional feature parameters.

10. The application of the general-purpose material dehydration process real-time online detection method according to any one of claims 1-8 in the industrial inorganic raw material drying process.