Method for quantitative calcium sulfate phase analysis and its use

Abstract

The present disclosure as a whole relates to methods for phase analysis of calcium sulfate materials that can be used to prepare gypsum products. In various aspects, the present disclosure provides methods for providing a computer-implemented system for determining the phase content of a test calcium sulfate material; methods for determining the phase content of a test calcium sulfate material; methods for providing a plaster raw material; and production methods for gypsum products.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of priority to U.S. Provisional Patent Application No. 63 / 369,168, filed on July 22, 2022, which is hereby incorporated by reference in its entirety. Background Art 1. Technical Field

[0003] The present disclosure generally relates to methods for phase analysis of calcium sulfate materials that can be used to prepare gypsum products. 2. Technical Background

[0005] Gypsum building products (e.g., wallboard, ceiling, drywall, or gypsum board) are boards made of a gypsum core placed between two layers of paper (commonly referred to as facing paper) on the outer surface of the gypsum core. They are widely used as building materials due to their ease of manufacture, strong mechanical strength, low thermal conductivity, and sound insulation properties. The quality of gypsum board depends largely on its gypsum core, which is made by mixing a gypsum core containing mainly calcium sulfate hemihydrate (CaSO 4 0.5H 2 The plaster material O (which itself is colloquially referred to as "plaster") is hydrated to form calcium sulfate dihydrate CaSO 4 ·2H 2 In order to make better gypsum board, higher calcium sulfate hemihydrate content in the stucco is needed to form a rigid interlocking gypsum crystal network and to facilitate the control of the setting of the gypsum product.

[0006] However, real-world industrial plaster materials are often complex mixtures of calcium sulfate with different crystalline phases, usually including calcium sulfate dihydrate (CaSO 4 ·2H 2 O), calcium sulfate hemihydrate (CaSO 4 0.5H 2 O), soluble anhydrous calcium sulfate (CaSO 4 ), inert calcium sulfate (CaSO 4 ) and free water (H 2O). Variations in stucco composition can occur due to the use of raw gypsum materials from a variety of resources (e.g., natural gypsum, flue gas desulfurization gypsum, and waste gypsum). In addition, when stucco is produced from raw materials by calcination, the operating conditions also affect the final composition. When the calcination temperature is carefully controlled between 110°C and 120°C, a hemihydrated calcium sulfate content is formed. When the gypsum raw material is not sufficiently dehydrated, a large amount of dihydrated calcium sulfate remains. In contrast, when the gypsum raw material is dried too much, a completely dehydrated calcium sulfate phase (soluble anhydrous calcium sulfate, at T = 120°C to 300°C) and inert calcium sulfate (at T>300°C) are observed. In order to control the quality of gypsum products, it is desirable to be able to quantify each of the calcium sulfate crystalline phases in industrial stucco and products prepared therefrom in a quick, easy and accurate manner.

[0007] Previously reported methodological procedures for analyzing the calcium sulfate phase content in stucco include Ramen spectroscopy, X-ray diffraction (XRD), calorimetry, and thermogravimetric analysis (TGA). However, none of these studies can achieve a complete analysis of all calcium sulfate phases. These methods assume the absence of some crystalline compositions (e.g., assuming that there is no soluble anhydrous calcium sulfate in the stucco), or only characterize certain phases in the stucco sample (e.g., measuring only soluble anhydrous calcium sulfate). Unfortunately, these assumptions are unrealistic for industrial stucco containing a complex mixture of calcium sulfate phases.

[0008] There remains a need in the art for a rapid method for the rapid and substantially complete determination of calcium sulfate phase in stucco raw materials, for example, for use in gypsum board manufacturing processes. Summary of the invention

[0009] In one aspect, the present disclosure provides a method for providing a computer-implemented system for determining phase content of a test calcium sulfate material, the method comprising:

[0010] Providing a plurality of first reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), and soluble calcium sulfate anhydrite (AIII);

[0011] Inert calcium sulfate (IN) and free moisture (FM);

[0012] For each of the plurality of first reference calcium sulfate samples, providing:

[0013] a first plurality of parameters associated with the dehydration weight loss curve of the first reference calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the first reference calcium sulfate sample, and

[0014] a second plurality of parameters associated with the hydration temperature rise curve of the first reference calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of the hydration curve of the first reference calcium sulfate sample,

[0015] for each of the first reference calcium sulfate samples, providing a first plurality of parameters and a second plurality of parameters to a computing device;

[0016] The first plurality of parameters and the second plurality of parameters are associated with the known contents of DH, HH, AIII, IN and FM of the first reference calcium sulfate sample via the computing device, thereby providing a test model that associates the first plurality of parameters and the second plurality of parameters of the test calcium sulfate sample with the contents of one or more of DH, HH, AIII, IN and FM of the test sample.

[0017] In another aspect, the present disclosure provides the method as described above, further comprising:

[0018] Providing a plurality of second reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), and soluble calcium sulfate anhydrite (AIII);

[0019] Inert calcium sulfate (IN) and free moisture (FM);

[0020] For each of the plurality of second reference calcium sulfate samples, providing:

[0021] a third plurality of parameters associated with the dehydration weight loss curve of the second reference calcium sulfate sample, the third plurality of parameters comprising a third plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the second reference calcium sulfate sample, and

[0022] a fourth plurality of parameters associated with the hydration temperature rise curve of the second reference calcium sulfate sample, the fourth plurality of parameters comprising a fourth plurality of fitting parameters associated with a fit of the hydration curve of the second reference calcium sulfate sample;

[0023] For each of the plurality of second reference calcium sulfate samples, providing the third plurality of parameters and the fourth plurality of parameters to the test model; and

[0024] Using this test model, DH, HH,

[0025] The content of one or more of AIII, IN and FM.

[0026] In another aspect, the present disclosure provides a method for determining the phase content of a test calcium sulfate material, the method comprising:

[0027] Providing a dehydration weight loss curve for the test calcium sulfate material;

[0028] Providing a hydration temperature rise curve of the tested calcium sulfate material;

[0029] supply:

[0030] a first plurality of parameters associated with the dehydration weight loss curve of the test calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the test calcium sulfate sample, and

[0031] a second plurality of parameters associated with the hydration temperature rise curve of the test calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of the hydration curve of the test calcium sulfate sample;

[0032] Providing a test model (e.g., developed using the above method) that relates the first plurality of parameters and the second plurality of parameters of a test calcium sulfate sample to the content of one or more of DH, HH, AIII, IN, and FM of the test sample;

[0033] providing the first plurality of parameters and the second plurality of parameters to the test model; and

[0034] Using the test model, the content of one or more of DH, HH, AIII, IN and FM of the test calcium sulfate sample is determined.

[0035] In another aspect, the present disclosure provides a method for providing a stucco raw material, the method comprising:

[0036] providing a first stucco material;

[0037] Determining the phase content of the first stucco material according to the method described above; and

[0038] Based on the determined phase contents, the first stucco material is modified to provide a stucco feedstock having a desired phase content.

[0039] In another aspect, the present disclosure provides a method for providing a stucco raw material, the method comprising:

[0040] calcining a first sample of calcium sulfate material using a first set of calcination parameters to provide a first calcined stucco;

[0041] Determining the phase content of the first calcined sludge according to the above method;

[0042] Based on the determined phase contents, a second sample of the calcium sulfate material is calcined using a second set of calcination parameters that are different from the first set of calcination parameters to provide the stucco feedstock.

[0043] In another aspect, the present disclosure provides a method for providing a stucco raw material, the method comprising:

[0044] calcining a first calcium sulfate material using a first set of calcining parameters to provide a first calcined stucco;

[0045] Determining the phase content of the first calcined sludge according to the above method;

[0046] Based on the determined phase contents, a second calcium sulfate material different from the first calcium sulfate material is selected and calcined using a second set of calcination parameters to provide the stucco feedstock.

[0047] In another aspect, the present disclosure provides a method for producing a gypsum product, the method comprising:

[0048] Providing plaster raw materials by any of the above methods;

[0049] hydrating the first stucco raw material in an aqueous slurry;

[0050] allowing the slurry to solidify; and

[0051] The set slurry is dried to form a gypsum product.

[0052] In another aspect, the present disclosure provides a production method for providing a gypsum product, the production method comprising:

[0053] providing a first plaster raw material;

[0054] Determining the phase content of the first stucco raw material by any of the methods described above;

[0055] hydrating the first stucco raw material in an aqueous slurry;

[0056] allowing the slurry to solidify; and

[0057] drying the set slurry to form the gypsum product,

[0058] The method also includes adjusting one or more parameters of the production process based on the determined phase content of the first stucco raw material.

[0059] Additional aspects will be apparent to persons of ordinary skill in the art from the description herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a schematic diagram illustrating an example of a method for developing a test model as described herein.

[0061] Figure 2 Schematic diagrams illustrating sample preparation and data acquisition operations for various methods described herein.

[0062] Figure 3 Schematic diagram illustrating the structure of an example artificial neural network for fast phase content analysis of calcium sulfate materials.

[0063] Figure 4 Graph showing actual phase contents of synthetic stucco samples prepared in the examples.

[0064] Figure 5 Graph showing the dehydration curve of a representative synthetic stucco sample, represented by the evolution of the moisture change (%) over time.

[0065] Fig. 6A Graph of the dehydration rate of a representative synthetic stucco sample as a function of time and its corresponding fitted curve using Equation 1. Figure 6B R of all fitting curves of the dehydration rate of the synthetic plaster samples prepared in the examples 2 Histogram of the distribution.

[0066] Fig. 7A Graph of the time-varying hydration curve of a representative synthetic stucco sample and its corresponding fitted curve using Eq. 2. Figure 7B The R values ​​of all fitted curves for the hydration curves of the synthetic plaster samples prepared in Example 2 Histogram of the distribution.

[0067] Figure 8 is a representative table of the final characterization parameter matrix as described herein.

[0068] Fig. 9 Figure 2 is a graph of the loss values ​​for the training and test datasets during 500 epochs of model training.

[0069] Fig.10 Graph of expected values ​​of phase contents of synthetic stucco samples versus calculated phase contents based on calculations of a machine learning model as described herein.

[0070] Fig.11 A flow chart of the CPA analysis method for phase analysis of stucco samples is provided.

[0071] Fig.12 A phase analysis diagram of a single-phase sample of a specific CPA analysis example described in Example 2 is provided.

[0072] Fig.13 A phase analysis diagram of the industrial stucco described in Example 2 is provided.

[0073] Fig.14 Composition information for the stucco mix prepared in Example 2 is provided.

[0074] Fig.15 Provides comparison Fig.14 Graph of expected values ​​of DH, HH, III and IN in a stucco mixture of Example 2 (as calculated from the material ratios) and observed values ​​of the corresponding phases based on the phase content analysis procedure in the CPA method of Example 2. DETAILED DESCRIPTION

[0075] As mentioned above, the inventors have noticed that quality control of plaster is important in the production of gypsum building products. Based on the calculation of the calcination weight loss and hydration weight gain of three pre-treated plaster samples, many of the inventors of the present invention have developed a so-called "complete phase analysis" (CPA) method to quantify each calcium sulfate phase essentially present in industrial plaster. The CPA method is characterized by a simple experimental protocol with high accuracy, making it potentially useful for industrial quality control. However, the CPA method requires a sample preparation phase of more than 12 hours for each sample measured. Therefore, the CPA method is very time consuming, thereby delaying data feedback to the production line. Therefore, it is necessary to provide a faster quantitative test.

[0076] To address this need in the art, the present inventors have developed a method for rapid and quantitative analysis of calcium sulfate materials by using a test model that correlates data from two rapid and simple physical analyses (weight loss upon dehydration and temperature rise upon hydration) to the content of the calcium sulfate phases essentially present in industrial stucco: calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM).

[0077] Machine learning techniques can be used to develop test models from a test database of known reference calcium sulfate samples. Over the past decade, machine learning has been used in many fields to study large amounts of existing data in many fields, including materials science. Machine learning is a powerful tool for discovering mathematical relationships between input and output data sets, which is particularly useful when the data size is too large or the structure is too complex for manual analysis, and when the conceptualization of mathematical relationships is not necessary. In addition, no physical or chemical rules need to be assumed in machine learning algorithms. Instead, these relationships are derived from reference data.

[0078] In particular, artificial neural network (ANN) based algorithms provide one of the most comprehensive and general machine learning algorithms that can be often used to solve practical problems.

[0079] As will be known to those of ordinary skill in the art, ANN-based techniques can be prone to overfitting, which can make the algorithm less accurate when analyzing unseen data sets. In order to avoid overfitting and improve the general applicability of ANN techniques, common regularization protocols such as dropout techniques and early stopping techniques can be applied. In the dropout technique, randomly selected neurons are ignored during training to prevent co-dependent neurons. In the early stopping technique, model training is terminated when the algorithm begins to perform poorly on unseen data sets. These two techniques are effective in preventing overfitting in ANNs, and both techniques are applied in this example.

[0080] Here, the measurement performed on the test calcium sulfate sample is performed quickly, generally no more than one hour, and conventional equipment for sample processing is required. The obtained data is provided to a pre-trained machine learning model to determine the calcium sulfate phase present in the test calcium sulfate sample with high accuracy. The inventors have found that by providing calcium sulfate phase analysis to a gypsum product production line, the method can be particularly useful in a production facility. In addition, the inventors have found that because the method provides a rapid quantitative phase analysis of gypsum products on the production line, the analysis results can be used in near real time to affect changes in production, for example, in the process of providing plaster raw materials to a gypsum product production line, or in the process of treating the plaster raw materials to provide a gypsum product. For example, with reference to the specific method described in the following examples, based on hundreds of reference calcium sulfate samples, the calcium sulfate phase content is determined with a root mean square error (RMSE) of 2.2% and a free moisture calculation accuracy of 87.7%, and the accuracy can be increased with more training data over time.

[0081] Therefore, the method of the present disclosure can provide a fast response time for the production line to ensure the quality of the gypsum product, save time and cost, and reduce production waste.

[0082] As described above, one aspect of the present disclosure provides a method for providing a computer-implemented system for determining the phase content of a test calcium sulfate material. The method comprises:

[0083] Providing a plurality of first reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), and soluble calcium sulfate anhydrite (AIII);

[0084] Inert calcium sulfate (IN) and free moisture (FM);

[0085] For each of the plurality of first reference calcium sulfate samples, providing:

[0086] a first plurality of parameters associated with the dehydration weight loss curve of the first reference calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the first reference calcium sulfate sample, and

[0087] a second plurality of parameters associated with the hydration temperature rise curve of the first reference calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of the hydration curve of the first reference calcium sulfate sample,

[0088] providing a computing device with a first plurality of parameters and a second plurality of parameters for each of the first reference calcium sulfate samples;

[0089] The first plurality of parameters and the second plurality of parameters are associated with the known contents of DH, HH, AIII, IN and FM of the first reference calcium sulfate sample via the computing device, thereby providing a test model that associates the first plurality of parameters and the second plurality of parameters of the test calcium sulfate sample with the contents of one or more of DH, HH, AIII, IN and FM of the test sample.

[0090] The method is based on correlation of hydration data and dehydration data with phase contents of a plurality of first reference calcium sulfate samples to provide a test model. Figure 1 An example of a process flow for providing a test model is shown in FIG. Generally speaking, the process flow includes three main main steps: (1) data collection relative to a first reference calcium sulfate sample; (2) so-called feature engineering, that is, converting the data into a set of parameters for being associated with the phase content of the first reference calcium sulfate sample; and (3) associating the parameters with the known phase content of the first reference calcium sulfate sample, for example, using machine learning techniques. In the data collection step, laboratory-scale or industrial-scale experiments are performed to collect a large amount of data for model training. In the feature engineering step, the raw experimental data of each first reference calcium sulfate sample collected in the first step is converted into well-designed characterization parameters that are easy to read in a programming language. For example, the characterization parameters may include parameters developed by fitting a dehydration weight loss curve and a fitting hydration temperature rise curve, as described below. This reduces the dimension of the input data matrix of the correlation, thereby improving the computational efficiency of the correlation. In the third step, the parameters are associated with the known phase content of the first reference calcium sulfate sample. For example, machine learning can be used, such as training with experimental data having certain learning rules, wherein an artificial neural network (ANN) algorithm is used to determine the relationship between the input value and the calcium sulfate phase content. Finally, (4) validation can be used to determine the accuracy of the machine learning model and to adjust the training parameters in the machine learning model based on its evaluation over multiple iterations (or “epochs”).

[0091] Good correlation can be provided by using a relatively large amount of the first calcium reference samples. For example, if a larger database can be used for training purposes, the machine learning algorithm analysis of the large data set and the machine learning model can be more accurate. For example, in various embodiments, a plurality of the first calcium samples include at least 100 first reference calcium sulfate samples, for example at least 200 first reference calcium sulfate samples. In various embodiments, a plurality of the first reference calcium sulfate samples include at least 300 first reference calcium sulfate samples, for example at least 400 first reference calcium sulfate samples.

[0092] A first reference calcium sulfate sample may be provided, having known contents of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM), provided, for example, by combining one or more of an industrial stucco product, phase-pure DH, phase-pure HH, phase-pure AIII and phase-pure IN. "Industrial stucco product" means a stucco product that is primarily stucco but contains a large amount of one or more of DH, HH, AIII and IN. "Phase-pure" refers to a sample of at least 95% by weight of the calcium sulfate phase. The reference calcium sulfate sample advantageously consists of at least 95% by weight (e.g., at least 98% by weight or at least 98% by weight) of one or more of DH, HH, AIII, IN and FM. A person of ordinary skill in the art may use conventional methods to determine the phase content of the first reference calcium sulfate sample. Alternatively, the provided CPA method may be used. Samples prepared by adding pure phase materials may be simply calculated by the weight of the material combination.

[0093] The first reference calcium sulfate sample advantageously substantially spans the possible phase content space of test sample to be tested.For example, in various embodiments, for at least 80% (for example, at least 90% or at least 95%) of a plurality of first reference calcium sulfate samples, the known amount of DH is in the scope of 3 % by weight to 20 % by weight.In various embodiments, for at least 80% (for example, at least 90% or at least 95%) of a plurality of first reference calcium sulfate samples, the known amount of HH is in the scope of 50 % by weight to 90 % by weight.In various embodiments, for at least 80% (for example, at least 90% or at least 95%) of a plurality of first reference calcium sulfate samples, the known amount of AIII is up to 30 % by weight.In various embodiments, for at least 80% (for example, at least 90% or at least 95%) of a plurality of first reference calcium sulfate samples, the known amount of IN is in the scope of 5 % by weight to 10 % by weight.

[0094] As described above, characterization data for each of the plurality of first calcium reference samples is collected. Figure 2An example of the data collection process is illustrated in . A first reference calcium sulfate sample (e.g., in the form of a synthetic stucco mixture) is analyzed for weight loss during calcination and temperature rise during hydration to provide a curve for each. Substantially the same experimental method should be used for all first reference calcium sulfate samples to ensure that changes in the experimental results are directly related to changes in the phase content in the corresponding synthetic stucco samples.

[0095] For each of the plurality of first reference calcium samples, a dehydration test is performed to provide a dehydration weight loss curve. During dehydration, the calcium sulfate dihydrate and calcium sulfate hemihydrate contents lose their hydration water and become soluble anhydrous calcium sulfate, as shown in the following reaction:

[0096]

[0097] Therefore, the weight loss during the dehydration test is directly related to the composition of the calcium sulfate dihydrate and calcium sulfate hemihydrate content in the first reference calcium sulfate sample.

[0098] Dehydration test can be carried out in a variety of ways, for example, by using thermogravimetric analysis (TGA) to monitor the weight loss of the first reference calcium sulfate sample during high temperature dehydration. For example, in various embodiments, by at least 100 ° C, for example, at least 200 ° C or at least 250 ° C, for example, at a temperature within the range of 100 ° C to 500 ° C, or 200 ° C to 500 ° C, or 250 ° C to 500 ° C, or 100 ° C to 400 ° C, or 200 ° C to 400 ° C, or 250 ° C to 400 ° C, the weight loss of the first reference calcium sulfate sample is measured to provide a dehydration weight loss curve for each first reference calcium sulfate sample. The dehydration weight loss can be measured, for example, within the range of 5 minutes to 30 minutes. Advantageously, the dehydration weight loss is measured until the moisture change reaches a stable state. As used herein, a "dehydration weight loss curve" is a data set providing a weight loss that varies over time; it does not need to be graphically plotted. The data points may be provided at an average interval of up to 15 seconds (eg, up to 10 seconds or up to 5 seconds). One of ordinary skill in the art will select a data density that provides the desired correlation results but maintains an appropriate data set size.

[0099] The hydration test can be performed in a variety of ways, generally measuring the temperature rise when the first reference calcium sulfate sample is hydrated. The temperature rise is caused by the reaction of the hemihydrated calcium sulfate and the soluble anhydrous calcium sulfate with water and the release of heat, as shown in the following reaction:

[0100] CaSO 4 +0.5H 2 O→CaSO 4 0.5H 2 O+Q reaction 3

[0101] CaSO 4 0.5H 2 O+1.5H 2 O→CaSO 4 ·2H 2 O+Q reaction 4 Where Q is the heat released during hydration. Since both reactions 3 and 4 are highly exothermic reactions, their occurrence heats the surrounding water. As with the dehydration test, the different phase contents of the first reference calcium sulfate sample result in different hydration curves, i.e., different functions of mixture temperature and time.

[0102] For example, the hydration temperature rise curve of each first reference calcium sulfate sample is provided by measuring the time-dependent temperature rise of the mixture of the first reference calcium sulfate sample and water. In various embodiments, the temperature rise is measured in the time period within the range of 15 minutes to 60 minutes (for example, 15 minutes to 45 minutes). The weight ratio of the first reference calcium sulfate sample and water can be, for example, in the range of 1:0.5 to 1:5, for example, in the range of 1:0.5 to 1:2. Advantageously, the hydration temperature rise is measured until the temperature change reaches a stable state. As used herein, "hydration temperature rise curve" is a data set providing a time-dependent weight loss; it does not need to be graphically mapped. Data points can be provided at an average interval of up to 15 seconds (for example, up to 10 seconds or up to 5 seconds). Those of ordinary skill in the art will select to provide the desired correlation results but maintain the data density of the appropriate data set size.

[0103] Dehydrated and hydrated data will typically contain many individual data points. As will be appreciated by those of ordinary skill in the art, machine learning training using such large data sets will be resource intensive and slow. Therefore, it is desirable that the feature engineering step "converts" the raw experimental data into digital parameters that concisely contain the most important characteristics of the raw data, but are also easy to process in a programming language. In other words, the feature engineering step can be specifically designed to reduce computing power and improve computing efficiency.

[0104] Therefore, a first plurality of parameters associated with the dehydration weight loss curve of each first reference calcium sulfate sample is provided. The first plurality of parameters may include, for example, at least four parameters (e.g., 4 parameters, or 5 parameters, or 6 parameters).

[0105] The inventors have determined that useful parameters can be provided by fitting the dehydration weight loss curve to a mathematical function. Thus, the first plurality of parameters includes a first plurality of fitting parameters associated with the fit of the dehydration weight loss curve of the first reference calcium sulfate sample. For example, this can be a fit of the dehydration weight loss curve itself. As an alternative, the fit of the dehydration weight loss curve can be provided as a fit of the rate of change of moisture over time, the rate of change of moisture being the first derivative of the dehydration weight loss curve. In various embodiments, this can be fitted to a Gaussian function, for example, the equation: Dehydration rate In various embodiments, the first plurality of fit parameters includes at least three fit parameters, such as three fit parameters, or four fit parameters, or five fit parameters.

[0106] The first plurality of parameters may also include other parameters. For example, in various embodiments, the first plurality of parameters include a parameter associated with the maximum weight change of the dehydration weight loss curve. This may simply be the value of the maximum weight change of the dehydration weight loss curve.

[0107] Similarly, provide the second multiple parameters that are associated with the hydration temperature rise curve of each first reference calcium sulfate sample.The second multiple parameters can comprise for example at least four parameters, for example at least five parameters, or at least six parameters, for example 4 parameters, 5 parameters, 6 parameters, 7 parameters or 8 parameters.

[0108] The inventors have determined that useful parameters can be provided by fitting the hydration temperature rise curve to a mathematical function. Therefore, the second plurality of parameters include the second plurality of fitting parameters associated with the fitting of the hydration temperature rise curve of the first reference calcium sulfate sample. For example, this can be the fitting of the hydration temperature rise curve itself. In various embodiments, this can be fitted to a combination of Gaussian functions, for example, a combination of a positively skewed Gaussian function and a non-skewed Gaussian function. For example, the fitting can be an equation: For example, wherein the temperature change is provided as a normalized temperature change obtained by dividing the temperature change by the initial temperature. In other embodiments, the normalized temperature change can be calculated by subtracting all temperature values ​​from the temperature at time 0. In various embodiments, the second plurality of fitting parameters includes at least four fitting parameters, such as at least five fitting parameters or at least six fitting parameters, such as four fitting parameters, or five fitting parameters, or six fitting parameters, or seven fitting parameters, or eight fitting parameters.

[0109] The first plurality of parameters and the second plurality of parameters are provided to a computing device, for example, for use in developing a test model using machine learning. A person of ordinary skill will appreciate that the computing device may include a processor programmed to execute a machine learning algorithm. This may take the form of a suitably programmed general-purpose computer. As a person of ordinary skill in the art will appreciate, the computing device includes any known device suitable for computing. For example, the computing device may be a user device (e.g., a device actively operated by a user), such as a mobile device (e.g., a tablet computer or a smart phone) or a fixed device (e.g., a desktop computer or a laptop computer). Such a device may include one or more processors (e.g., a central processing unit (CPU), a graphics processing unit (GPU), a network processor, and / or an integrated circuit or controller that performs processor operations), one or more sensors (e.g., a gyroscope, an accelerometer, a camera, a touch screen, a tactile button, a keyboard, etc.), one or more network communication models (e.g., IEEE 802.11 (Wi-Fi), The computing device may include a computer system (e.g., a global positioning system (GPS), a wide area wireless interface, Ethernet, a synchronous optical network, a digital subscriber line, etc.), and one or more digital storage systems (e.g., random access memory (RAM), read-only memory (ROM), flash memory, a hard drive, a solid-state drive, a compact disk (CD), a digital video disk (DVD), and / or tape storage), all of which may be connected via a system bus or similar mechanism. In some examples, the computing device may include other components and / or peripherals (e.g., a keyboard, a mouse, a sensor, a removable memory, a printer, etc.).

[0110] The machine learning algorithm can be, for example, an artificial neural network algorithm, as described herein. The inventor has determined that the artificial neural network algorithm is particularly useful for mapping the relationship between the characterization parameter and the phase content of the first reference calcium sulfate sample. In machine learning applications, the artificial neural network decomposes complex nonlinear functions into a combination of linear transformations with learnable parameters connected by a nonlinear activation function. Along with the increase in the number of such nonlinear transformations organized in a "layer", highly complex feature spaces can be effectively learned.

[0111] In various embodiments, the artificial neural network has an input layer including at least one node for each parameter in the first plurality of parameters and the second plurality of parameters; at least two hidden layers having nodes associated with the input layer; and an output layer having at least five output nodes associated with the hidden layers. Figure 3 One such embodiment is shown in , where a large number of neurons are arranged in multiple layers in the following order:

[0112] (1) Input layer, where the characterization parameters are fed into the artificial neural network.

[0113] (2) One or more hidden layers in which neurons are interconnected.

[0114] (3) Output layer, which provides calculated values.

[0115] The neurons at the input layer first receive the values ​​of the characterization parameters and then process the signals to the next connected neurons, where weights are applied to adjust each learning process. The artificial neural network used here contains two hidden layers, each with 64 neurons. Ultimately, the characterization parameters travel from the input layer, advance through the hidden layers, and finally reach the output layer with the calculated values.

[0116] Artificial neural network algorithms are powerful machine learning models that can be used for both regression and classification problems. Regression is a category of machine learning that evaluates the relationship between some input variables and one or more numerical outputs. Classification measures the relationship between input variables and output non-numeric labels. In various embodiments, both functions can be implemented. For example, in various embodiments, the output node includes a regression node for each of DH, HH, and IN. In various embodiments, the output node includes two nodes related to AIII and FM together, for example, a regression node for the sum of AIII and FM and a classification node for classifying whether AIII or FM exists. Examples are shown in Figure 3 Among them, the first four nodes are designed to provide calculated values ​​of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), inert calcium sulfate (IN), and the total amount of soluble anhydrous calcium sulfate / free moisture content in the calcium sulfate sample (AIII / FM), all provided by regression analysis. A classification function is introduced in the fifth node of the output layer, which is called "free moisture marker". It is known that soluble anhydrous calcium sulfate and free moisture cannot coexist in real-world samples, so for quantitative purposes, they share the same node. The free moisture marker is used to indicate whether the value shown in the soluble anhydrous calcium sulfate / free moisture node represents the amount of soluble anhydrous calcium sulfate (when the free moisture marker = 0) or the amount of free moisture content (when the free moisture marker = 1).

[0117] Therefore, using artificial neural network or other means, the first plurality of parameters and the second plurality of parameters are associated with the known contents of DH, HH, AIII, IN and FM of the first reference calcium sulfate sample via a computing device. This provides a test model that associates the first plurality of parameters and the second plurality of parameters of the test calcium sulfate sample with the contents of one or more of DH, HH, AIII, IN and FM of the test sample.

[0118] Artificial neural networks can be trained by comparing the difference between the calculated values ​​from the output layer and the actual values ​​from sample measurements. The difference between the calculated value and the actual value is described by the loss value (L), as defined by the following equation: where y i is the calculated value from the output layer in the i-th group, is the true value of the i-th group, and m is the sample group fed into it. Based on this definition, an artificial neural network model with a smaller L value means a more accurate calculation relative to the actual value. Therefore, the goal of artificial neural network training is to reduce the L value over multiple iterations (called epochs), in which the weight associations associated with the parameters are adjusted according to the designed learning rules. Ideally, the L value will become smaller during the iteration process, which means that the model calculation becomes closer to the actual value.

[0119] For example, in various embodiments, providing a test model includes training an artificial neural network algorithm through multiple epochs (e.g., at least 100 epochs, or at least 200 epochs, or at least 250 epochs). In each epoch, the training artificial neural network algorithm may include comparing the calculated value of the output layer with the actual known value of one or more of DH, HH, IN, AIII and FM of the first reference calcium sulfate sample. For example, in response to the calculated value of the output layer and the actual known value of one or more of DH, HH, IN, AIII and FM of the first reference calcium sulfate sample, the weight association can be adjusted relative to one or more parameters in the first plurality of parameters and the second plurality of parameters.

[0120] The method may also include providing a plurality of second reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM). There may be, for example, at least 50, at least 100 or at least 200 second reference calcium sulfate samples; with reference to the description herein, a person of ordinary skill in the art will select a desired number that provides a meaningful validation of the test model. These second reference calcium sulfate samples can be processed in the same manner as the plurality of first reference calcium sulfate samples described above to provide a third plurality of parameters associated with the dehydration weight loss curve of the second reference calcium sulfate sample, the third plurality of parameters including a third plurality of fitting parameters associated with the fitting of the dehydration weight loss curve of the second reference calcium sulfate sample, and a fourth plurality of parameters associated with the hydration temperature rise curve of the second reference calcium sulfate sample, the fourth plurality of parameters including a fourth plurality of fitting parameters associated with the fitting of the hydration curve of the second reference calcium sulfate sample. Data associated with the second plurality of reference samples can be used in one or more of each epoch to test the validity of the model when present in the epoch relative to data that is not part of the training data set. This can prevent the complication of "overfitting", in which the artificial neural network tends to fit the training data set closely, but cannot be reliably fitted with additional data. Therefore, for each of the plurality of second reference calcium sulfate samples, the third plurality of parameters and the fourth plurality of parameters can be provided to the test model; and using the test model, the content of one or more of DH, HH, AIII, IN and FM of each second reference calcium sulfate sample can be determined. This can be, for example, relative to the content of two or more, three or more, four or more, or all five of DH, HH, AIII, IN and FM. This can be performed in multiple epochs, and can also include providing a loss value (L) for each epoch, and selecting the test model of the epoch where the loss value is significantly minimized (e.g., within 10% of the minimum value, or within 5% of the minimum value) as the final test model. Therefore, a person skilled in the art can select a final training model that is appropriately trained but not over-trained by examining the performance relative to multiple second reference calcium sulfate samples by having low loss values ​​not only for the multiple first reference calcium sulfate samples used to train the model but also for different multiple second reference calcium sulfate samples.

[0121] With a training test model such as that prepared as described above, the phase content of an unknown calcium sulfate sample can be determined quickly and easily. All that needs to be done is to provide a dehydration weight loss curve and a hydration temperature rise curve for a test calcium sulfate material, parameterize the data, and apply the data to the training model. Therefore, another aspect of the present disclosure provides a method for determining the phase content of a test calcium sulfate material, the method comprising:

[0122] Providing a dehydration weight loss curve for the test calcium sulfate material;

[0123] Providing a hydration temperature rise curve of the tested calcium sulfate material;

[0124] supply:

[0125] a first plurality of parameters associated with the dehydration weight loss curve of the test calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the test calcium sulfate sample, and

[0126] a second plurality of parameters associated with the hydration temperature rise curve of the test calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of the hydration curve of the test calcium sulfate sample;

[0127] providing a test model that relates the first plurality of parameters and the second plurality of parameters of a test calcium sulfate sample to the content of one or more of DH, HH, AIII, IN, and FM of the test sample;

[0128] providing the first plurality of parameters and the second plurality of parameters to the test model; and

[0129] Using the test model, the content of one or more of DH, HH, AIII, IN and FM of the test calcium sulfate sample is determined.

[0130] For example, the test model may be provided in any of the ways described above.

[0131] The test calcium sulfate sample advantageously has the phase content in the phase content space of the first reference calcium sulfate sample used as the training basis of the test model.For example, in various embodiments, the amount of DH in the test calcium sulfate sample is in the scope of 3 % by weight to 20 % by weight.In various embodiments, the amount of HH in the test calcium sulfate sample is in the scope of 50 % by weight to 90 % by weight.In various embodiments, the amount of AIII in the test calcium sulfate sample is up to 30 % by weight.In various embodiments, the amount of IN in the test calcium sulfate sample is in the scope of 5 % by weight to 10 % by weight.

[0132] The provision of dehydration test data and hydration test data can be carried out as described above.It will be appreciated by those of ordinary skill in the art that dehydration test data and hydration test data should be collected in a manner similar to the data used as the test model basis.Although some experimental details may be different, the method for collecting the data of the test calcium sulfate sample should produce substantially the same result.Certainly, those of ordinary skill in the art can collect data in significantly different modes, as long as data or resulting parameters can be reliably transformed into a value, and this value is substantially similar to the value provided by the program that will be used to develop the data as the test model basis.

[0133] Similarly, the first multiple parameters and the second multiple parameters of the test calcium sulfate sample can be provided in a substantially identical manner with the first multiple parameters and the second multiple parameters of the first reference calcium sulfate sample for training the test model. It will be appreciated by those of ordinary skill in the art that the calculation details may be different, but the basic fitting equation is advantageously the same, or at least the resulting parameters should be related to the data in substantially the same manner.

[0134] Once determined, the measured content of one or more of DH, HH, AIII, IN and FM of the test calcium sulfate sample can be provided to the user in any manner. They can be displayed on a screen, sent via an email or other messaging surface, and / or stored in a database.

[0135] Furthermore, and most importantly, based on the present disclosure, one of ordinary skill in the art can use the determined contents of one or more of DH, HH, AIII, IN, and FM of a test calcium sulfate sample in a variety of ways to determine and adjust various process parameters in the production of stucco raw materials and gypsum products.

[0136] For example, based on the present disclosure, one of ordinary skill in the art can use the measurements described herein when providing, for example, a stucco raw material suitable for producing a gypsum product such as gypsum building board.

[0137] For example, this information can be used as a basis for modifying a plaster material to provide a desired phase content. Therefore, another aspect of the present disclosure is a method for providing a plaster raw material. The method includes providing a first plaster material; determining the phase content of the first plaster material according to the method described herein; and, based on the determined phase content, modifying the first plaster material to provide a plaster raw material with a desired phase content. In some embodiments, the modification can be a combination of a first plaster material and one or more second plaster materials, each of which has a different phase content from the first plaster material, to provide a plaster raw material with a desired phase content. A plaster production facility may have different plaster materials with different phase contents, and may be mixed based on measurements made as described herein to provide a plurality of plaster raw materials with different desired phase contents. The various second plaster materials may, for example, be derived from materials recycled from a previous process, and / or from different mineral sources. In various embodiments, the modification of the first plaster material includes calcining the first plaster material; in the case where the phase content measurement results indicate incomplete calcination, the first plaster material may be further calcined to provide a desired phase content. In various embodiments, the modification of the first stucco material comprises adding moisture to the first stucco material, for example, where the first stucco material is over-calcined to provide an undesirably high level of dehydration. In various embodiments, the modification of the first stucco material comprises adjusting the particle size distribution of the first stucco material (e.g., by adjusting grinding conditions).

[0138] As another example, the information provided by the measurement method described herein can be used as a basis for adjusting the calcination parameters of a given calcium sulfate material. For example, another aspect of the present disclosure provides a method for providing a plaster raw material, the method comprising calcining a first sample of a calcium sulfate material using a first set of calcination parameters to provide a first calcined plaster; determining the phase content of the first calcined plaster according to the method described herein; and calcining a second sample of the calcium sulfate material using a second set of calcination parameters different from the first set of calcination parameters based on the determined phase content to provide a plaster raw material. Here, the phase content measured after calcination under the first set of calcination parameters can be used by a person of ordinary skill in the art to determine how to adjust the calcination parameters for calcining a second sample of the same material. In various embodiments, the second set of calcination parameters has a calcination temperature program different from the first set of calcination parameters. In various embodiments, the second set of calcination parameters has a calcination time different from the first set of calcination parameters.

[0139] Based on the measured phase content, various calcination parameters can be adjusted. For example, the calcination temperature can be increased, wherein a higher temperature provides a higher calcination rate, or reduced, wherein a lower temperature provides a lower calcination rate. The humidity in the calciner can be increased, wherein a higher humidity provides a lower calcination rate, or can be reduced, wherein a lower humidity provides a lower calcination rate. The throughput of the calciner can be increased, wherein a higher throughput (i.e., less time in the calciner) provides a lower degree of calcination, or can be reduced, wherein a lower throughput (i.e., more time in the calciner) provides a higher degree of calcination. Similarly, the calcination time can be extended, wherein a longer time provides a higher degree of calcination, or can be shortened, wherein a shorter time provides a lower degree of calcination. Compared with the calcination of the first sample, the calcination of the second sample providing a higher degree of calcination can advantageously provide, for example, in the stucco raw material, relatively less calcium sulfate dihydrate and / or less free moisture than in the first calcined stucco. Calcination of the first sample to provide a lower degree of calcination than when the first sample was calcined can advantageously provide, for example, relatively less anhydrous calcium sulfate and / or relatively less inert material in the stucco feedstock than in the first calcined stucco.

[0140] For example, the information provided by the measurement method described herein can be used as a basis for determining a suitable calcium sulfate raw material for a calcination method to provide a desired plaster raw material. For example, another aspect of the present disclosure provides a method for providing a plaster raw material, the method comprising: calcining a first calcium sulfate material using a first set of calcination parameters to provide a first calcined plaster; determining the phase content of the first calcined plaster according to the method as described herein; and based on the determined phase content, selecting a second calcium sulfate material different from the first calcium sulfate material and calcining it using a second set of calcination parameters to provide a plaster raw material. In some embodiments, the second set of calcination parameters may be the same as the first set of calcination parameters. In other embodiments, the second set of calcination parameters may be different from the first set of calcination parameters. Any calcination parameters may be different, including those described above. For example, in various embodiments, the second group has a calcination temperature program different from the first group. In various embodiments, the second group has a calcining furnace throughput or calcination time different from the first group.

[0141] The second calcium sulfate material can be different from the first calcium sulfate material in various ways. For example, in various embodiments, the second calcium sulfate material may have a composition different from the first calcium sulfate material, such as a different distribution of various ingredients (e.g., one or more of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); the relative amount of inert calcium sulfate (IN) and free moisture (FM) is significantly different, such as one or more of these differing by more than 2% by weight or more than 5% by weight). In various embodiments, the second calcium sulfate material may be processed differently from the first calcium sulfate material. In various embodiments, the particle size distribution of the second calcium sulfate material may be different from the particle size distribution of the first calcium sulfate material. For example, in some embodiments, compared with the first material, the second calcium sulfate material has a significantly different particle size distribution (e.g., a significantly different d50 value, e.g., differing by at least 10% from the d50 of the first calcium sulfate material). In some embodiments, the second calcium sulfate material has a significantly larger d50 value than the d50 value of the first calcium sulfate material (e.g., at least 10% greater than the d50 of the first calcium sulfate material). In other embodiments, the second calcium sulfate material has a d50 value significantly smaller than the d50 value of the first calcium sulfate material (e.g., at least 10% less than the d50 of the first calcium sulfate material). Smaller overall particle size (e.g., smaller d50) can result in higher water demand and more energy consumption when drying; larger overall particle size (e.g., larger d50) will have the opposite effect. In some such embodiments, the second calcium sulfate material may have a composition substantially similar to the first calcium sulfate material (e.g., one or more of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); the relative amounts of inert calcium sulfate (IN) and free moisture (FM) are not significantly different, e.g., each of these differs by no more than 2 wt % or no more than 1 wt %).

[0142] One of ordinary skill in the art will use material selection, material processing and calcination parameters to provide the desired stucco raw materials for the production process of gypsum products. Generally, a large amount of hemihydrate and a small amount of dihydrate, anhydrite and inert materials are desired, but one of ordinary skill in the art will understand that these processes can tolerate dihydrate, anhydrite and inert materials to a certain extent.

[0143] The stucco raw materials provided by these methods can be used to provide gypsum products using other methods familiar to those of ordinary skill in the art. For example, another aspect of the present disclosure is a method for producing a gypsum product, the method comprising: providing stucco raw materials by the above method; hydrating the first stucco raw materials in an aqueous slurry; and setting the slurry to form a gypsum product.

[0144] Knowledge of the phase content of a plaster raw material used to provide a gypsum product can also be used to determine parameters of a production process. For example, another aspect of the present disclosure provides a production method for providing a gypsum product. The production method includes providing a plaster raw material; determining the phase content of the plaster raw material according to the method described herein; hydrating the first plaster raw material in an aqueous slurry; and setting the slurry to form a gypsum product. Notably, the method also includes adjusting one or more parameters of the production method based on the determined phase content of the first plaster raw material.

[0145] The production process generally includes forming an aqueous slurry from a stucco raw material, then setting the slurry to form a wet gypsum product, and then drying the wet gypsum product to provide a dry gypsum product. Knowledge of the phase content of the stucco raw material can provide one of ordinary skill in the art with information that helps adjust various parameters associated with such a process.

[0146] For example, one of ordinary skill in the art can use knowledge of the phase content to adjust the content of the slurry. For example, in various embodiments, adjusting the content of the slurry includes selecting the amount of an accelerator (e.g., powdered calcium sulfate dihydrate or potassium sulfate) based on the determined phase content. More accelerators may be included in the slurry to cause faster solidification of the slurry (e.g., when the amount of hemihydrate is relatively low), or less accelerators may be included in the slurry to cause slower solidification of the slurry (e.g., when the amount of hemihydrate is relatively high). Similarly, in various embodiments, adjusting the content of the slurry includes selecting the amount of a retarder (e.g., diethylenetriaminepentaacetic acid and its salts) based on the determined phase content. More retarder may be included in the slurry to cause slower solidification of the slurry (e.g., when there is relatively more hemihydrate), or less retarder may be included in the slurry to cause faster solidification of the slurry (e.g., when there is relatively less hemihydrate). In various embodiments, adjusting the content of the slurry includes selecting the water demand of the slurry (i.e., the relative amount of water compared to the plaster raw material) based on the determined phase content. The water demand can be, for example, increased, such as when relatively more hemihydrate is present, but more energy is required when drying, or decreased, which requires less energy when drying. The overall density (e.g., measured by the board weight of the gypsum board) can be adjusted, for example, by adjusting the amount of the foaming agent (e.g., lauryl ether sulfate), or the foaming conditions (e.g., the time / intensity of mixing air into the slurry) can be selected based on the determined phase content. Foaming can be increased to provide a higher density / board weight, which can increase strength but consume more energy when drying per unit volume of product, or to provide a lower density / board weight, which can reduce strength but consume less energy when drying. Based on the disclosure herein, one of ordinary skill in the art can provide a desired balance of energy consumption (and therefore cost consumption when drying), board strength, and other properties by adjusting one or more parameters associated with the slurry content.

[0147] In addition, parameters associated with drying can be adjusted based on knowledge of the phase content of the stucco raw material. For example, parameters can be adjusted to provide relatively more drying or relatively less drying, such as dryer temperature (higher temperature to provide more drying; lower temperature to provide less drying); dryer humidity (lower humidity to provide more drying; higher humidity to provide less drying); and dryer throughput (less throughput / longer residence time to provide more drying); (more throughput / less residence time to provide less drying). For example, when relatively more hemihydrate is present, less drying can be provided because relatively more of the water initially present will crystallize into dihydrate. Based on the disclosure herein, one of ordinary skill in the art can provide a desired degree of drying to provide a product that is sufficiently dried to provide a desired mechanical strength, but not overdried to provide a process that operates at a desired degree of energy efficiency.

[0148] Example 1

[0149] The following examples illustrate specific embodiments of the methods of the present disclosure and various uses thereof. They are set forth for illustrative purposes only and should not be construed as limiting the scope of the present disclosure.

[0150] Data collection

[0151] First, a large number of synthetic stucco samples were prepared in order to build a customized calcium sulfate database. The samples were prepared by mixing real industrial stucco with pure calcium sulfate dihydrate, calcium sulfate hemihydrate, soluble anhydrous calcium sulfate and inert calcium sulfate samples, and were immediately evaluated by two standardized measurement procedures: dehydration test and hydration test. These can be described in Example 2 below. The raw experimental data were then labeled and saved to a local database. Since all experimental data in this database were collected under the same protocol, changes in the experimental results are directly related to changes in the phase content in the corresponding synthetic stucco samples.

[0152] Preparation of synthetic plaster samples

[0153] 490 synthetic plaster samples were prepared by mixing industrial plaster samples with various amounts of calcium sulfate dihydrate, calcium sulfate hemihydrate, soluble anhydrous calcium sulfate and / or inert calcium sulfate. Since the composition of the industrial plaster samples is easily known, the phase content of the resulting synthetic plaster samples can be easily calculated. In addition, the conventional complete phase analysis (CPA) method (see Example 2 below) was used to confirm the accurate phase content of the first 183 synthetic plaster samples. This quantitative analysis method has been shown to have high measurement accuracy, which provides more reliable and detailed information about the phase content of the synthetic plaster samples. Finally, the calculated phase content is marked as the "actual value" of each synthetic plaster sample and then stored in a local database. In the following section, these "actual values" will be used as a benchmark for training machine learning models.

[0154] The synthetic stucco samples were designed to represent actual industrial stuccos. Therefore, they typically contained 3.0% to 20.0% calcium sulfate dihydrate, 50.0% to 90.0% calcium sulfate hemihydrate, 0.0% to 30.0% soluble anhydrous calcium sulfate, and 5.0% to 10.0% inert calcium sulfate.

[0155] The actual phase contents of all 490 composite mortar samples were Figure 4 As shown in the figure, all phase contents are distributed within the above-mentioned expected range. This means that the synthetic plaster samples prepared in this way are representative of industrial plaster products. It is also worth mentioning that the free moisture content is present in some samples, but only at very low levels. It is known that free moisture cannot coexist with soluble anhydrous calcium sulfate because they react quickly with each other and form hemihydrated calcium sulfate. However, trace amounts of free moisture may be maintained in calcium sulfate samples with soluble anhydrous calcium sulfate content close to zero. Although the presence of free moisture is not common in industrial calcium sulfate, this situation is also marked and counted in the following machine learning model in an effort to improve the feasibility of the machine learning model in all possible cases.

[0156] Dehydration test

[0157] Dehydration tests were performed by monitoring the weight loss of synthetic stucco samples during dehydration using the thermogravimetric analysis (TGA) method. The 5000XL moisture analyzer performs dehydration tests and sets the dehydration temperature to 300°C. The weight loss (marked as moisture change) is recorded every eight seconds until a steady value is reached. This steady value is marked as the maximum moisture change value, which will be used in the following feature engineering steps. Typically, under this experimental setting, the dehydration process takes about 10 minutes to reach its steady state.

[0158] Hydration test

[0159] The hydration test is performed by measuring the temperature change after mixing a synthetic plaster sample with a specific amount of water. In this example, the hydration test is performed by adding 5g of the synthetic plaster sample to 5mL of distilled water in a plastic tube. The temperature change is monitored by a thermocouple while the entire system is kept in an isothermal box to minimize environmental damage. The entire hydration process is recorded every second and usually takes about 30 minutes to complete.

[0160] Feature Engineering

[0161] The dehydration and hydration curves collected in this example contain a large number of data points (e.g., about 125 data points per dehydration curve and about 2000 data points per hydration curve). In this project, each dehydration curve was parameterized using the value of the maximum moisture change value and by fitting the dehydration rate (i.e., the rate of change of moisture evolution as the first derivative of the original dehydration curve) to a Gaussian function. Each hydration curve was fit to a modified Gaussian function.

[0162] Maximum moisture change

[0163] An example of a dehydration curve is Figure 5 A very important feature for the dehydration test is the maximum moisture change value, which corresponds to the sum of the calcium sulfate dihydrate and calcium sulfate hemihydrate contents in the synthetic stucco sample, as shown in reactions 1 and 2. Therefore, this value is determined as the first characterization parameter. The features can be extracted by finding the maximum value (labeled "Cal_max") in each dehydration curve. All Cal_max values ​​are then stored in a design matrix (labeled "Characterization_Parameter_1") of dimension m×1, where m is the number of experimental groups.

[0164] Fitting dehydration curve

[0165] In order to represent a complete data set of dehydration curves but also maintain a minimum number of characterization parameters, it was found that the dehydration rate (% / s) during the dehydration test (i.e., the first derivative of the dehydration weight loss curve) can be related to time t 1 The Gaussian function of the function is a good fit:

[0166]

[0167] where t 1 is the elapsed time of the dehydration test, and a, b, c are three non-zero independent moisture change rate parameters of the Gaussian function. Fig. 6A Examples of experimental data (scatter points) and their corresponding fitted curves (lines) are given. 2 The Gaussian function with a value of 0.9971 fits accurately. Figure 6BAs shown, the R of each fitting curve 2 The value is greater than 0.990, which means that all experimental dehydration data can be well represented. The three parameters obtained (cal_rate_a, cal_rate_b and cal_rate_c) are also stored into a matrix named "characterization_parameters_2" with dimensions of m x 3.

[0168] Fitting hydration curve

[0169] Similar to the case of the dehydration curve, using as time t 2 The experimental data of the hydration curve were engineered into six parameters using a modified Gaussian function of the function. In order to eliminate the influence of ambient temperature differences, the temperature was first normalized in each experimental group. Generally speaking, a typical hydration curve contains two main peaks, such as Fig. 7A Therefore, a combination of positively skewed and non-skewed Gaussian functions was found to be the best fit to the hydration test data via the following equation:

[0170]

[0171] where t 2 is the elapsed time of the calcination test, and a, b, c, d, e, and f are six independent temperature variation parameters. The first term in Equation 2 is a positively skewed Gaussian function, which is used to describe the first peak (at about 100 seconds); on the other hand, the second term in Equation 2 is designed to describe the second peak in the hydration curve (at about 1500 seconds). Fig. 7A Examples of experimental data during hydration and their corresponding fitted curves are given. 2 The values ​​are as high as 0.9994, which means that the fitted curves convincingly represent the experimental data. Figure 7B The R of each fitted curve is shown. 2 The value is greater than 0.995. The parameters (hyd_a, hyd_b, hyd_c, hyd_d, hyd_e, and hyd_f) are then stored into a matrix named “featured_parameters_3” of dimension m×6 for the machine learning model.

[0172] Machine learning models using artificial neural networks (ANN)

[0173] In the previous section, the experimental data has been converted into ten characterization parameters (i.e., maximum humidity change, three independent humidity change parameters, and six independent temperature change parameters), and three matrices have been generated: characterization_parameter_1 (m×1), characterization_parameter_2 (m×3), and characterization_parameter_3 (m×6). The final input matrix (m×10) can be established by combining these three matrices. Figure 8An overview of the final characterization parameter matrix is ​​provided in . In this final input matrix, the parameters in each column are normalized in the range [0, 1] in order to reduce computational cost and minimize bias error.

[0174] The ANN algorithm was applied to the machine learning model in order to map the relationship between the characterization parameters and the phase content of the synthetic plaster samples. Figure 3 Show and describe.

[0175] By comparing the calculated values ​​from the output layer with the actual values ​​from sample preparation (i.e., Figure 4 The ANN is trained by measuring the difference between the calculated and actual values. The difference between the calculated and actual values ​​is described by the loss value (L) as described above. The machine learning model is trained for 500 epochs and the change in the loss value is Fig. 9 As shown in the figure, the experimental data is divided into a training data set and a validation data set at a ratio of 80:20. The training data set is first fed into the machine learning model to train the weights in the ANN. Fig. 9 As shown, the loss value of the machine learning model continues to decrease within the epoch of the training data set. This means that when the model iterates enough epochs, the model calculation becomes more accurate. The validation data set is also fed into the training machine learning model of each epoch, and its corresponding loss value is calculated. The reason for using the validation data set is to check whether the model is effective for unseen data sets. It is found that the loss value of the validation data set is also reduced in the first 100 epochs, which means that the model becomes more accurate in this time period of calculating the phase content of the validation data set. However, even when the loss value of the training data set continues to decrease, the loss value of the validation data set gradually increases after about 300 epochs, thereby showing the beginning of overfitting. Therefore, early stopping techniques are applied as regularization to control the overfitting of ANN. In this example, it is most appropriate to use the test model of the epoch (e.g., 300 epochs as in this example) where the loss from the validation set data is significantly minimized as the final test model. Of course, other data sets will show different overfitting behaviors, and will show different expected stopping points to control overfitting.

[0176] Performance Overview

[0177] After completing the training of the machine learning model, its performance was first visually evaluated by comparing the calculated values ​​with the actual phase content values. Fig.10 A comparison between the expected and calculated phase content values ​​is shown. The expected values ​​for each phase content in the synthetic stucco samples (x-axis) were calculated from the sample preparation stage described above, whereas the calculated values ​​(y-axis) were obtained from the machine learning model. If the resulting data point lies on the diagonal line in each figure, it means that the calculated phase content is exactly the same as the expected value. Fig.10The results in Figure 2 show that, except for very few outliers, most of the data points are very close to the diagonal of each phase. These results provide an intuitive explanation of the excellent accuracy of the current method of machine learning model for phase content analysis.

[0178] Finally, the performance of the machine learning model was quantitatively evaluated by two aspects: the root mean square error (RMSE) of the calculated phase content values ​​and the accuracy of the calculated free moisture content. First, the RMSE analysis indicates the difference between the calculated values ​​from the machine learning model and the actual phase content values ​​in the synthetic stucco samples, as calculated by the following equation:

[0179]

[0180] Based on this definition, a lower RMSE value would indicate a more accurate calculation by the machine learning model. In this study, the RMSE value of the calculated stucco phase content in the test dataset was calculated to be 2.2% when the model was trained for an iteration number of 300 epochs. Meanwhile, the accuracy of the free moisture content calculation in this method was determined to be 87.7%. Therefore, the low RMSE value and high type accuracy prove the concept that this machine learning model is an accurate and effective method to calculate the phase content in calcium sulfate materials.

[0181] Therefore, the inventor has developed a fast and quantitative phase analysis method to measure the calcium sulfate phase content in calcium sulfate materials and products using a machine learning model in conjunction with an artificial neural network (ANN) algorithm. The method has a simple and fast measurement procedure and only requires conventional equipment for sample handling. The calcium sulfate material is dehydrated and its weight / water content is recorded during the dehydration test. The same calcium sulfate material is also hydrated by mixing the material with water to be fully hydrated. Its temperature change is also recorded. After data acquisition and feature engineering, the characterized experimental data is fed into the trained machine learning model, wherein the calculation of the corresponding phase content of the calcium sulfate material is easily calculated automatically. The accuracy of the analytical method is demonstrated by its low root mean square error (RMSE) value calculated about phase content and the high type accuracy of measuring about free moisture. Therefore, the method is considered to be applicable to the quality control of industrial calcium sulfate materials and products in large-scale operations.

[0182] Example 2

[0183] Example 2 provides the so-called "CPA method" for determining the phase content of a calcium sulfate sample. This example describes a quantitative phase analysis that measures the amount of all calcium sulfate phases in a calcium sulfate sample. Broadly, the phase content is calculated using the calcination weight loss and hydration weight gain of the initial sample, the semi-hydrated sample, and the fully hydrated sample. Uncomplicated sample preparation and routine measurement procedures make it suitable for on-site quality control measures in industrial settings.

[0184] In gypsum wallboard manufacturing, calcium sulfate materials are typically provided in calcined form for use as raw materials. These materials are typically fine white powders and are often referred to as "stucco" after the hemihydrate phase that is typically desired. Depending on the characteristics of the gypsum raw material and the calcination conditions, stucco may contain CaSO 4 ·xH 2 Different crystalline phases of the O system, including the main content of hemihydrate:

[0185] Calcium sulfate dihydrate (DH): It is possible to find some traces of the dihydrate ("gypsum") in stucco due to incomplete calcination. Insufficient time or temperature in the calcination kiln or large particle size of the gypsum raw material are possible causes of incomplete calcination. In the latter, calcination to the hemihydrate occurs on the surface of the large particles, while their cores may provide a significant dihydrate content.

[0186] Calcium sulfate hemihydrate (HH): HH is the main component of the stucco mixture and is formed from DH by an endothermic dehydration reaction typically between 110°C and 120°C as shown in reaction (R-1):

[0187]

[0188] Generally speaking, HH can occur in two forms: in wet processes (eg, autoclaving) the α-form predominates, and in dry processes (eg, calcination) the β-form predominates. The β-form of HH is more relevant to building board production.

[0189] Soluble calcium sulfate anhydrite (AIII): Soluble anhydrite also has an alpha form and a beta form, which are prepared by further dehydration (usually 120°C to 300°C) of the alpha hemihydrate and the beta hemihydrate, respectively.

[0190]

[0191] Depending on the calcination conditions, AIII can be formed directly from DH without forming an intermediate HH phase.

[0192]

[0193] AIII, also known as γ-CaSO 4, is very unstable and spontaneously rehydrates to HH if water vapor is present. Generally, the presence of AIII in stucco is undesirable because it increases the water demand and makes it difficult to control the hydration process.

[0194] Insoluble calcium sulfate anhydrite (AII and AI): Further heating of AIII to temperatures above 300 °C results in the formation of a calcite called β-CaSO 4 or insoluble anhydrite of AII. There is also a so-called α-CaSO 4 A high temperature phase of anhydrite, or AI, which is stable only above 1180°C. AII has very limited hydration, and AI does not show solubility in water. AII is relevant to commercial applications of gypsum, and in this context is primarily classified as an inert phase (IN).

[0195] In a typical building board preparation, stucco is mixed with water and other solid or liquid additives to prepare a slurry that is allowed to set (by hydration of the hemihydrate to form interlocking dihydrate crystals) to form a "gypsum" board material. The phase content of the stucco greatly affects the properties of the slurry, such as water demand and setting time, which ultimately control the properties of the final drywall product. More specifically, controlling the setting time is important in modern drywall production, which operates at high speeds.

[0196] This example provides a method for quantifying CaSO based on gravimetric and thermogravimetric analysis (GA and TGA). 4 ·xH 2 Phase analysis method for all phases in the O system. This method is a significant improvement of the Dweck and Lasota method, which is described in J. Dweck, EIP Lasota, Quality control of commercial plasters by thermogravimetry, Thermochim. Acta. 318 (1998) 137–142. doi: 10.1016 / S0040-6031 (98) 00338-4, which is hereby incorporated by reference in its entirety. Here, new samples, measurement results and equations including the AIII phase are included in the analysis. The sample preparation and measurements described in the examples are simple enough to be performed in the factory, i.e., a fully equipped laboratory is not required.

[0197] Theory / Calculation

[0198] Samples and analysis

[0199] The quantitative phase analysis of a given stucco material was based on gravimetric and thermogravimetric phase analysis of three samples made from the stucco material, as follows:

[0200] ORG (original): This is the sample obtained from the stucco mix without any modification.

[0201] HUM (humidified and dried): This sample was prepared by conditioning the stucco in a humidity chamber. The humidity level and the duration of conditioning were adjusted to ensure that all AIII in the stucco was converted to HH after conditioning. After conditioning, the sample was dried at a temperature high enough to remove any free moisture (FM) but low enough to retain the water of crystallization.

[0202] HYD (Hydrated and Dry): This sample was prepared by complete hydration of the stucco. Complete hydration means that there is enough water and time for all hydratable phases (HH and AIII) to be converted to DH.

[0203] Two types of measurements are performed on the samples:

[0204] Gravimetric Analysis (GA): This is basically a measure of the weight change of the HUM and HYD samples after conditioning. The output of this analysis is the weight gain of the HUM and HYD samples, recorded as a percentage

[0205]

[0206] Where %ΔW g is the percentage of weight gain during the GA analysis, W f and W 0 are the weights of samples with and without modification, respectively.

[0207] Thermogravimetric analysis (TGA): This analysis measures the isothermal weight loss at a temperature high enough to remove all crystal water and free water. The absolute value of the weight loss is % ΔW l It is reported as a percentage (similar to equation (T-1)) and measured for all three samples.

[0208] Model Assumptions

[0209] This analysis method is based on two assumptions:

[0210] 1. FM and AIII do not coexist in stucco. Since AIII can be rapidly converted to HH in the presence of moisture, it is assumed that FM is stable only when AIII is absent in stucco.

[0211] 2. The only hydratable phases in the stucco are HH and AIII. In other words, only the partially hydrated calcium sulfate phase CaSO 4 ·xH 2 O is HH (x = 0.5). Stucco mixes usually contain some paper dust which comes from recycled drywall and which can absorb water. The amount of these impurities is very small and their effect on weight gain and loss is negligible.

[0212] A consequence of the first hypothesis is the need to separate the formulations of stucco containing AIII and FM. The presence of AIII or FM can be easily distinguished from the weight gain of the HUM samples. If AIII is present in the plaster, the weight of the sample should increase after conditioning in the humidity chamber. The water of crystallization will not be removed in the subsequent drying step, however, FM will be easily removed during the drying stage. Therefore, it can be written

[0213]

[0214] where D, H, A, F and I are the weight percentages of DH, HH, AIII, FM and IN phases in the stucco mixture, respectively. The model has four unknown quantities ( The situation is very rare and can be combined with the other two cases), so four linear equations are needed to solve for all the unknowns. One equation is naturally given by equation (T-2), while the other three equations can be provided by gravimetric and thermogravimetric analysis.

[0215] Model constants

[0216] The model uses the absolute values ​​of the theoretical weight gains and losses of pure DH, HH, and AIII phases as constants. The symbols used to represent these constants are It is the absolute value of the relative weight change when phase 1 transforms into phase 2. The values ​​of all constants used in this model can be easily calculated from the molecular weight of each phase of calcium sulfate. These values ​​are also given in the supplementary information below.

[0217] Thermogravimetric phase analysis (TGPA)

[0218] The three weight losses of ORG, HUM and HYD samples measured by thermogravimetric analysis provide the remaining three equations required to calculate the phase contents. According to the model assumptions, AIII and FM cannot coexist and each case should be treated separately.

[0219]

[0220] The initial weight of all three samples can be calculated as:

[0221] W 0 =W DH +W HH +W AIII +W IN , (T-3)

[0222] where the terms on the right side of equation (T-3) are the weights of each phase in the stucco. The weight loss of the ORG sample measured by TGA is due to the dehydration of the DH and HH phases (the reverse reaction of (R-2) and (R-3)).

[0223] Therefore, one can simply write

[0224]

[0225] It should be noted that the phase percentages are calculated as

[0226]

[0227] In equation (T-5), the numerator is the weight of each phase in the stucco and the denominator is the initial weight of the sample (before conditioning and calcination). When the stucco is conditioned in a humidifier to prepare the HUM sample, all AIII contents are converted to HH, and the new weight of HH, W′ HH , which can be written as

[0228]

[0229] Therefore, the weight of the conditioned HUM sample was read as

[0230]

[0231] For the measured weight loss of HUM after calcination, it can be written

[0232]

[0233] Divide the numerator and denominator of the fraction on the right by W 0 And rearranging equation (T-8) gives

[0234]

[0235] A similar calculation can be applied to the HYD sample. In the HYD sample, all HH and AIII are converted to DH, and the weight of the conditioned sample can be written as

[0236]

[0237] For the measured weight loss of the HYD sample, it can be written

[0238]

[0239] For the HUM sample, equation (T-11) can be rewritten as the following linear relationship.

[0240]

[0241] Equations (T-4), (T-8) and (T-12) establish a system of three linear equations that can be solved for D, H and A. Knowing the three phase contents, I (inert phase content) can be easily calculated from equation (T-2).

[0242]

[0243] When FM is present in the stucco, the weight loss measured by TGA comes from the dehydration of DH, HH and the removal of FM. Therefore, it can be written

[0244]

[0245] As described above, conditioning the stucco to prepare the HUM samples removed free moisture and the weight of the samples decreased. In the HYD samples, all HH was converted to DH and FM was also removed during the drying step.

[0246]

[0247] For the stucco samples with AIII, the following equation for the weight loss measured for the HUM and HYD samples when the stucco contained FM can be written.

[0248]

[0249] Similarly, equations (T-13), (T-15) and (T-16) can be used to establish a linear system of equations to solve for D, H and F. Both systems of linear equations developed for thermogravimetric phase analysis can be solved analytically. However, it is preferred to solve them as constrained optimization problems, where the phase content is constrained between 0 and 100 to avoid non-physical negative values ​​of the phase content. Matrix representations of these linear systems of equations are shown in the Supplementary Information.

[0250] Gravimetric Phase Analysis (GPA)

[0251] Since only the weight gain of the HUM and HYD samples could be measured, the gravimetric phase analysis lacked an equation to solve for all four phase contents. The final equation can be derived from any of the three weight losses measured by TGA.

[0252]

[0253] When the plaster contains AIII, the weight of the sample increases after conditioning in the humidifier, and the weight gain of the measured HUM sample is due to the conversion of AIII to HH. Therefore, it can be written

[0254]

[0255] This equation gives the weight percentage of phase AIII only. For the weight gain of the HYD sample, we can write

[0256]

[0257] Knowing A from equation (T-17), equation (T-18) readily gives H. A and H can then be substituted into any of equations (T-4), (T-9), and (T-12) to calculate D. The weight percent of the inert phase is then calculated from equation (T-2) as described above.

[0258]

[0259] If the stucco contains free moisture, the free moisture is removed and the weight decreases after the HUM sample is prepared. Therefore, the weight gain measured for the HUM sample is equal to the negative value of the free moisture weight percentage.

[0260]

[0261] Free moisture is also removed in the HYD sample, and the weight gain can be written as

[0262]

[0263] F and H can again be substituted into any of equations (T-13), (T-15) and (T-16) to calculate D, and then I can be calculated from equation (T-2) (T-2).

[0264] Once ORG, HUM and HYD samples were prepared for TGA analysis, and can be easily measured by simple weight measurements. This means that gravimetric analysis can be easily used to cross-check the accuracy of thermogravimetric phase analysis and improve the reliability of the calculations. Finally, the phase analysis can be summarized by classifying all equations into two groups:

[0265] Phase analysis of stucco containing AIII using equations (T-2), (T-4), (T-9), (T-12), (T-17) and (T-18) to calculate DH, HH, AIII in stucco

[0266] and weight percentage of inert phase.

[0267] Phase analysis of stucco containing free moisture using equations (T-2), (T-13),

[0268] (T-15), (T-16), (T-19) and (T-20) to calculate the weight percentage of DH, HH, free moisture and inert phase in the stucco.

[0269] Fig.11 The flow chart shown summarizes the entire process used for phase analysis.

[0270] An example of a characterization method is provided below. One of ordinary skill in the art can adapt the method as desired.

[0271] Materials and methods

[0272] Material

[0273] Calcium sulfate dihydrate (ACS reagent 98%) and calcium sulfate hemihydrate (purum ≥ 97%) were purchased from Sigma-Aldrich. The most reliable source of laboratory-grade AIII is Drierite, an industrial desiccant made from calcined gypsum. Drierite (no indicator, 8 mesh) was also purchased from Sigma-Aldrich. High-purity AIII and HH samples were prepared from DH reagent purchased from Sigma-Aldrich. To this end, DH was first calcined at 200°C to obtain a high-purity AIII sample, which was conditioned in a humidifier (75% humidity, 45°C) to prepare high-purity HH. Two types of reagents (purchased and synthesized) were used to prepare synthetic "plaster" mixtures with various compositions. If necessary, insoluble calcium sulfate was used as an inert phase. Industrial-grade plaster samples provided by CertainTeedGypsum, Inc. were also analyzed. Industrial plaster samples are mixtures of different types of synthetic gypsum (FGDG) and recycled gypsum (WG) and are used as raw materials in building board production plants.

[0274] Sample preparation

[0275] To prepare the HUM samples, the plaster was first conditioned overnight (at least 18 hours) in a humidity chamber (LH-1.5, Associated Environmental Systems, MA, USA) with 75% humidity and 45°C temperature. The HUM samples thus prepared were then moved to an oven and dried at 45°C for two hours to remove any possible free moisture. Finally, the samples were cooled to room temperature (about 23°C) in a vacuum oven. These HUM samples were sealed and stored in a desiccator for further use.

[0276] It is well known that the AIII phase can be readily rehydrated to HH even at very low humidity. Although industrial methods have been established to condition stucco mixtures at high water pressures and temperatures in order to reduce the AIII content, conditions for complete conversion of AIII to HH at low pressures and temperatures have not been fully investigated. In this study, the optimal humidity and temperature combination for complete conversion of AIII to HH was achieved by trial and error - but one of ordinary skill in the art can adopt the methods described herein in their own analysis.

[0277] HYD samples were prepared by complete hydration of the stucco mixture. To achieve complete hydration, the stucco was mixed with enough water to completely submerge the powder and form a slurry. The samples were covered and allowed to stand at room temperature for 2 hours. The samples were then placed in a vacuum oven where they were dried at 45°C overnight (at least 18 hours).

[0278] Test and Measurement

[0279] HYD samples were prepared by complete hydration of the stucco mixture. To achieve complete hydration, the stucco was mixed with enough water to completely submerge the powder and form a slurry. The samples were covered and allowed to stand at room temperature for 2 hours. The samples were then placed in a vacuum oven where they were dried at 45°C overnight (at least 18 hours).

[0280] The weight gain of HUM and HYD samples after hydration was measured by a balance with a reliability of 0.001 g (Denver Instrument PI-225 D.3 Pinnacle Analytical Balance). The weight loss of ORG, HUM and HYD samples due to complete calcination (complete dehydration) was measured using a TGA instrument and / or a moisture analyzer. TA Instruments Q500 TGA was used for isothermal calcination at 300° C. for 30 minutes. The weight loss of ORG, HUM and HYD samples due to complete calcination (complete dehydration) was measured using a moisture analyzer (obtained from Arizona Instrument LLC). The samples tested by the 5000XL were heated to 300°C and held at that temperature until the rate of weight change became less than 0.1 g / s. All weight losses used in this particular study were measured with a moisture analyzer. Although a TGA instrument can provide higher precision (0.1°C and 0.1 μg) compared to a moisture analyzer (1°C and 1 mg), it is believed that a moisture analyzer is the appropriate tool for stucco phase analysis for the following reasons:

[0281] 1. The average difference between all phase contents calculated by the two instruments is less than 1%.

[0282] 2. Moisture analyzers are easy to set up and operate in industrial environments.

[0283] 3. The samples analyzed by TGA are usually very small (about 50 mg) compared to the amount that can be analyzed by a moisture analyzer (about 5 g). When testing a sample of an inhomogeneous mixture such as plaster, a sufficiently large amount needs to be analyzed to provide a meaningful statistical representation of the sample. In theory, repeating the TGA measurement many times can solve this problem. In addition to the fact that the high precision provided by the TGA is not necessary for this analysis, considering the cost and time of TGA analysis, a moisture analyzer seems to be a highly suitable candidate for this application.

[0284] For each of the ORG, HUM and HYD samples, at least three 10 g samples were prepared. The values ​​reported for each phase content in this study are the average values ​​calculated from these three samples, and the error bars represent the standard deviation from the average.

[0285] Results and discussion

[0286] Single phase samples

[0287] As described in the previous section, two types of single-phase samples of HH and AIII were used. The first type was a reagent purchased directly from Sigma-Aldrich and labeled HH SA and Drierite. The second type of samples was produced in-house from purchased DH using the above procedure. These homemade HH and AIII samples were then labeled HH DH and AIII DH . Fig.12 Results of phase analysis performed on these samples are shown; detailed information is also provided in the Supplementary Information, Table S1.

[0288] Fig.12 It was confirmed that the DH reagent (DH) purchased from Sigma-Aldrich had very high purity and contained almost 100% DH. SA The reagents showed that the HH content (about 89%) was lower than the required purity of the sample (≥97%). The product of these reagents showed that both the complex EDTA and X-ray diffraction tests were used to characterize DH, but only the former was performed to characterize HH. SA Complexed EDTA is a volumetric analysis that can measure calcium content, but it cannot identify the phase from which the calcium came. SAThe calcium in drierite can come from a hydratable calcium-containing phase that is not a hemihydrate or an inert calcium-containing phase. In the case of drierite, purity or AIII content has not been reported, as anhydrous calcium sulfate is commonly used as a desiccant and chemical reagent. The phase analysis described here detected about 69% AIII and about 22% IN content in drierite. Some HH was also detected, most likely due to moisture adsorbed from the environment by some AIII. Fig.12 As shown, high-purity HH can be synthesized from DH samples. DH and AIII DH Samples. Ten synthetic stucco samples were prepared by combining high purity samples; their phase analysis is explained in the next section.

[0289] Synthetic plaster samples

[0290] The phase contents of the synthetic plaster samples were adjusted in a way to cover the range of phase contents typically observed during the different stages of industrial plaster preparation. Since single-phase samples with known phase contents were used, the phase contents of the synthetic plaster samples were also known. The expected phase contents of the synthetic plaster samples are shown in Table 1 below. In this table, the phase content column refers to the expected phase content in each synthetic plaster sample, however the corresponding formulation of each single-phase sample mixture is also listed on the right-hand side. More specifically, the values ​​of the phase contents were calculated by adding up each phase content in the single-phase sample mixture.

[0291] Table 1: Phase contents and single-phase compositions used to prepare synthetic stucco samples .

[0292]

[0293] After preparing the synthetic stucco samples, phase analysis was performed and the root mean square error (RMSE) between the expected and measured phase contents was calculated as

[0294]

[0295] The average RMSE for the TGPA and GPA analyses of the ten synthetic stucco samples were 0.45% and 1.06%, respectively. The lower accuracy of the GPA analysis may be because the analysis uses two different types of measurements, whereas all measurements used in the TGPA are of the same type and precision. Overall, the AIII phase shows the largest relative error between the measured and expected values. This is most likely due to the nature of the AIII phase, which is very unstable. During the preparation of the synthetic stucco samples, AIII can quickly absorb moisture.

[0296] Industrial plaster samples

[0297] Industrial plaster samples were collected at different stages of the plaster preparation process midstream in a drywall production plant. Two samples (Sample A and Sample B) were collected at the initial stage of plaster preparation. In both samples, calcination was performed to remove most of the DH content. Due to their specific preparation conditions, Sample A and Sample B were expected to have relatively high AIII contents. The third sample (Sample C) was collected from the output of the production plant, where the AIII content had been mostly converted to HH. The phase analysis of these samples is shown in Fig.13 middle.

[0298] Fig.13 The two calcination mills are shown to operate in a similar manner. There is very little difference in the phase content of the stucco produced by Sample A and Sample B. As expected, both Sample A and Sample B stucco contain a large amount of AIII due to over-calcination. Once these stuccos are combined and conditioned in the mixer, the AIII content in the stucco is significantly reduced. A large amount of AIII in the stucco is tolerable as long as it helps to reduce the DH content. Fig.13 As shown, the DH content does not change after the calcination stage, since high temperatures are required to remove DH.

[0299] Mixture of industrial plaster samples and single-phase samples

[0300] To further test the method for industrial stucco products, 50 sets of mixture samples were prepared by mixing industrial stucco samples with single-phase samples. The main motivation for this part of the study was to simulate the small differences in phase content in industrial stucco products by manually adding different amounts of single-phase content and then confirm the sensitivity and reliability of the phase analysis procedure. The formation of the stucco single-phase mixture used in this study is shown in Fig.14 More details of the mixtures are provided in the Supplementary Information in Tables S2 and S3. As mentioned above, the average RMSE in the TGPA analysis (0.45%) is smaller than the average RMSE in the GPA analysis (1.06%), which means that at least in the range of the stucco phase content in this study, the TGPA analysis is more accurate than the GPA analysis. Therefore, the following data were analyzed by the TGPA method.

[0301] Fig.15 A comparison between the expected values ​​for each phase content in a single-phase sample of stucco based on sample preparation (x-axis) and the observed values ​​(y-axis) obtained from the experiment and the phase content analysis procedure is shown. The diagonal line in each figure refers to the case where the observed value exactly matches the expected value. The deviation of the data point from the upper left corner means that the experimental data provides a value for the corresponding phase that is larger than the expected value. In contrast, the deviation towards the lower right direction indicates a value that is underestimated by the analysis compared to the expectation.

[0302] The results demonstrate that the phase analysis method developed in this study can provide reliable values ​​for the phase contents in stucco single-phase mixtures. The average RMSE value for these 50 sets of stucco single-phase samples is 1.14%, as calculated by equation (T-21), which is slightly higher than the average RMSE value (0.45%) obtained for the above 10 sets of synthetic stucco samples. This is most likely due to the fact that industrial stucco samples are much more complex than synthetic stucco samples and can have higher variations during sampling. However, the low value of RMSE still indicates the good quality of the phase analysis using the procedure described herein.

[0303] The calculated values ​​of RMSE for DH, HH, AIII and IN were 0.78%, 1.70%, 1.58% and 0.60%, respectively. In general, the analysis of DH and IN contents was very reliable, while the RMSE values ​​for HH and AIII contents were slightly higher, but still convincing. Fig.15 It can be appreciated that the data points for DH content are mostly located around the diagonal line, indicating an excellent prediction of the expected value. Fig.15 The experimental values ​​of IN content in the graph are slightly more scattered; however, this should be attributed to the relatively narrow scales of both the x-axis and the y-axis in the figure. It is worth noting that in areas where the expected IN content is low, the observed IN content is generally overestimated. Based on the above discussion, it is possible that the single-phase content used for the mixture (especially the HH and AIII samples) may contain some IN content, which may lead to lower than actual IN expected values. Therefore, the observed values ​​of IN content are trustworthy in these cases. The main reason why DH and IN have good RMSE in the phase analysis test is because of the fact that they are not sensitive to free moisture in the air.

[0304] On the other hand, it is not unexpected that the RMSE for HH and AIII contents is slightly higher than that for DH and IN contents. As mentioned previously, AIII content is very sensitive to free moisture and can be converted to HH during sample preparation. Therefore, some AIII may be converted to HH during the experiment, which simultaneously results in a decrease in AIII weight and an increase in HH weight. This hypothesis is also supported by the following results: when the expected value is small, the observed AIII content is usually less than the expected AIII content ( Fig.15 ). In fact, there are several groups with observed AIII content values ​​of zero.

[0305] in conclusion

[0306] An accurate phase analysis method is provided that measures the weight percentage of all calcium sulfate phases in a stucco sample. The phase contents are calculated using weight loss and weight gain values ​​measured after calcination and hydration of three samples prepared from the stucco mixture, respectively:

[0307] ORG sample: original stucco without any conditioning.

[0308] HUM samples: The stucco mix was conditioned at 45°C and 75% humidity to convert all AIII to HH.

[0309] HYD Sample: The stucco mix was fully hydrated by adding water to convert all AIII and HH to DH.

[0310] The method proposed in this study has significant advantages over previously reported methods. First, the method is able to distinguish all soluble calcium sulfate phases and quantify their contents. In addition, it requires simple sample preparation, and the measurement is ideal for on-site analysis in an industrial setting. The phase content of the stucco mixture can greatly affect the production rate, water demand, and final product properties in drywall production. The method proposed in this study can be used as a quality control measure midstream in drywall production to control the quality of the stucco mixture.

[0311] Of course, one of ordinary skill in the art may use other methods to measure the phase content of a known sample.

[0312] It will be apparent to those skilled in the art that various modifications and variations may be made to the processes and devices described herein without departing from the scope of the present disclosure. Therefore, the present disclosure is intended to cover these modifications and variations of the present invention as long as they fall within the scope of the appended claims and their equivalents.

[0313] Further aspects of the present disclosure are provided by the following enumerated embodiments, which can be combined and arranged in any number and in any combination as long as it is technically or logically consistent.

[0314] Embodiment 1. A method for providing a computer-implemented system for determining the phase content of a test calcium sulfate material, the method comprising:

[0315] Providing a plurality of first reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM);

[0316] For each of the plurality of first reference calcium sulfate samples, providing:

[0317] a first plurality of parameters associated with the dehydration weight loss curve of the first reference calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the first reference calcium sulfate sample, and

[0318] a second plurality of parameters associated with the hydration temperature rise curve of the first reference calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of the hydration curve of the first reference calcium sulfate sample,

[0319] providing a computing device with a first plurality of parameters and a

[0320] The second multiple parameters;

[0321] The first plurality of parameters and the second plurality of parameters are associated with the known contents of DH, HH, AIII, IN and FM of the first reference calcium sulfate sample via the computing device, thereby providing a test model that associates the first plurality of parameters and the second plurality of parameters of the test calcium sulfate sample with the contents of one or more of DH, HH, AIII, IN and FM of the test sample.

[0322] Embodiment 2. The method according to embodiment 1, wherein the plurality of first reference calcium sulfate samples comprises at least 100 first reference calcium sulfate samples, such as at least 200 first reference calcium sulfate samples.

[0323] Embodiment 3. The method according to embodiment 1, wherein the plurality of first reference calcium sulfate samples comprises at least 300 first reference calcium sulfate samples, such as at least 400 first reference calcium sulfate samples.

[0324] Embodiment 4. The method of any one of Embodiments 1 to 3, wherein each first reference calcium sulfate sample comprises one or more of an industrial stucco product, phase pure DH, phase pure HH, phase pure AIII, and phase pure IN.

[0325] Embodiment 5. The method of any one of Embodiments 1 to 4, wherein for at least 80% (e.g., at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of DH is in the range of 3 wt % to 20 wt %.

[0326] Embodiment 6. The method of any one of Embodiments 1 to 5, wherein for at least 80% (e.g., at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of HH is in the range of 50 wt % to 90 wt %.

[0327] Embodiment 7. The method of any one of Embodiments 1 to 6, wherein for at least 80% (e.g., at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of AIII is up to 30 wt%.

[0328] Embodiment 8. The method of any one of Embodiments 1 to 7, wherein for at least 80% (e.g., at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of IN is in the range of 5 wt % to 10 wt %.

[0329] Embodiment 9. The method according to any one of Embodiments 1 to 8, wherein the dehydration weight loss curve of each first reference calcium sulfate sample is provided by thermogravimetrically measuring the weight loss of the first reference calcium sulfate sample at a temperature in the range of at least 100°C, such as at least 200°C, or at least 250°C, for example, from 100°C to 500°C, or from 200°C to 500°C, or from 250°C to 500°C, or from 100°C to 400°C, or from 200°C to 400°C, or from 250°C to 400°C.

[0330] Embodiment 10. The method of embodiment 9, wherein the weight loss is measured over a period of time ranging from 5 minutes to 30 minutes.

[0331] Embodiment 11. A method according to embodiment 9 or embodiment 10, wherein the dehydration weight loss is advantageously measured until the moisture change reaches a steady state.

[0332] Embodiment 12. The method according to any one of embodiments 1 to 11, wherein the fitting of the dehydration weight loss curve is a fitting of the moisture change rate over time.

[0333] Embodiment 13. The method according to embodiment 12, wherein the time-varying moisture variation is fitted to a Gaussian function, for example,

[0334] Embodiment 14. A method according to any one of Embodiments 1 to 13, wherein the first plurality of fitting parameters includes at least three fitting parameters.

[0335] Embodiment 15. The method of any one of Embodiments 1 to 14, wherein the first plurality of parameters comprises parameters associated with the maximum weight change of the dehydration weight loss curve.

[0336] Embodiment 16. The method according to embodiment 15, wherein the parameter associated with the maximum weight change of the dehydration weight loss curve is the maximum weight change of the dehydration weight loss curve.

[0337] Embodiment 17. The method of any one of Embodiments 1 to 16, wherein the first plurality of parameters comprises at least four parameters.

[0338] Embodiment 18. The method according to any one of Embodiments 1 to 17, wherein the hydration temperature rise curve of each first reference calcium sulfate sample is provided by measuring the temperature rise of the mixture of the first reference calcium sulfate sample and water over time.

[0339] Embodiment 19. A method according to embodiment 18, wherein the temperature rise is measured over a time period ranging from 15 minutes to 60 minutes, such as from 15 minutes to 45 minutes.

[0340] Embodiment 20. The method of embodiment 18 or embodiment 19, wherein the weight ratio of the first reference calcium sulfate sample to water is in the range of 1:0.5-1:5, such as 1:0.5-1:2.

[0341] Embodiment 21. The method of any one of Embodiments 1 to 20, wherein the hydration temperature rise curve is fit to a combination of Gaussian functions, for example, a combination of a positively skewed Gaussian function and a non-skewed Gaussian function.

[0342] Embodiment 22. A method according to embodiment 21, wherein the hydration temperature rise curve is fitted to the following For example, where the temperature change is provided as a normalized temperature change by dividing the temperature change by the initial temperature.

[0343] Embodiment 23. A method according to any one of Embodiments 1 to 22, wherein the second plurality of fitting parameters includes at least four parameters, such as at least five parameters.

[0344] Embodiment 24. A method according to any one of Embodiments 1 to 22, wherein the second plurality of fitting parameters includes at least six parameters.

[0345] Embodiment 25. A method according to any one of Embodiments 1 to 24, wherein the second plurality of parameters includes at least four parameters, such as at least five parameters.

[0346] Embodiment 26. A method according to any one of Embodiments 1 to 24, wherein the second plurality of parameters includes at least six parameters.

[0347] Embodiment 27. A method according to any one of Embodiments 1 to 26, wherein the computing system includes a processor programmed to execute a machine learning algorithm.

[0348] Embodiment 28. A method according to Embodiment 27, wherein the machine learning algorithm is an artificial neural network algorithm.

[0349] Embodiment 29. A method according to embodiment 28, wherein the artificial neural network has an input layer, which includes at least one node for each parameter of the first plurality of parameters and the second plurality of parameters; at least two hidden layers, which have nodes associated with the input layer; and an output layer, which has at least five output nodes associated with the hidden layer.

[0350] Embodiment 30. A method according to embodiment 29, wherein the output node includes a regression node for each of DH, HH and IN.

[0351] Embodiment 31. A method according to embodiment 30, wherein the output node includes two nodes related to AIII and FM together, such as a regression node for the sum of AIII and FM, and a classification node for classifying whether AIII or FM exists.

[0352] Embodiment 32. A method according to any one of Embodiments 28 to 31, wherein providing the test model includes training the artificial neural network algorithm through multiple epochs (e.g., at least 100 epochs, or at least 200 epochs, or at least 250 epochs).

[0353] Embodiment 33. A method according to Embodiment 32, wherein in each epoch, training the artificial neural network algorithm includes comparing the difference between the calculated value of the output layer and the known value of one or more of DH, HH, IN, AIII and FM of the first reference calcium sulfate sample.

[0354] Embodiment 34. A method according to embodiment 33, wherein in each epoch, the weight association is adjusted relative to one or more of the parameters of the first plurality of parameters and the second plurality of parameters in response to the difference between the calculated value of the output layer and the actual known value of one or more of DH, HH, IN, AIII and FM of the first reference calcium sulfate sample.

[0355] Embodiment 35. The method according to any one of embodiments 1 to 34, further comprising:

[0356] providing a plurality of second reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM);

[0357] For each of the plurality of second reference calcium sulfate samples, providing:

[0358] a third plurality of parameters associated with the dehydration weight loss curve of the second reference calcium sulfate sample, the third plurality of parameters comprising a third plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the second reference calcium sulfate sample, and

[0359] a fourth plurality of parameters associated with the hydration temperature rise curve of the second reference calcium sulfate sample, the fourth plurality of parameters comprising a fourth plurality of fitting parameters associated with a fit of the hydration curve of the second reference calcium sulfate sample;

[0360] For each of the plurality of second reference calcium sulfate samples, the test model is provided with

[0361] the third plurality of parameters and the fourth plurality of parameters; and

[0362] Using this test model, DH, HH,

[0363] The content of one or more of AIII, IN and FM.

[0364] Embodiment 36. The method of embodiment 35, wherein providing the third plurality of parameters and the fourth plurality of parameters to the test model and determining the content of one or more of DH, HH, AIII, IN and FM of each second reference calcium sulfate sample is performed in multiple epochs.

[0365] Embodiment 37. A method according to embodiment 36, the method also includes providing a loss value (L) for each epoch, and selecting a test model of the epoch in which the loss value is significantly minimized as the final test model.

[0366] Embodiment 38. A method for determining the phase content of a test calcium sulfate material, the method comprising:

[0367] Providing a dehydration weight loss curve for the test calcium sulfate material;

[0368] Providing a hydration temperature rise curve of the tested calcium sulfate material;

[0369] supply

[0370] a first plurality of parameters associated with the dehydration weight loss curve of the test calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the test calcium sulfate sample, and

[0371] a second plurality of parameters associated with the hydration temperature rise curve of the test calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of the hydration curve of the test calcium sulfate sample;

[0372] A test model is provided, which tests the first plurality of parameters and the second plurality of parameters of the calcium sulfate sample with one of DH, HH, AIII, IN and FM of the test sample.

[0373] The content of one or more of them is related;

[0374] providing the first plurality of parameters and the second plurality of parameters to the test model; and

[0375] Using the test model, the content of one or more of DH, HH, AIII, IN and FM of the test calcium sulfate sample is determined.

[0376] Embodiment 39. A method according to embodiment 38, wherein the test model is created by a method according to any one of embodiments 1 to 36.

[0377] Embodiment 40. The method according to embodiment 38 or embodiment 39, wherein the amount of DH in the test calcium sulfate sample is in the range of 3 wt % to 20 wt %.

[0378] Embodiment 41. The method according to any one of Embodiments 38 to 40, wherein the amount of HH in the test calcium sulfate sample is in the range of 50 wt% to 90 wt%.

[0379] Embodiment 42. The method according to any one of embodiments 38 to 41, wherein the amount of AIII in the test calcium sulfate sample is up to 30 wt%.

[0380] Embodiment 43. The method according to any one of Embodiments 38 to 42, wherein the amount of IN in the test calcium sulfate sample is in the range of 5 wt.% to 10 wt.%.

[0381] Embodiment 44. The method of any one of Embodiments 38 to 43, wherein the dehydration weight loss curve of the test sample is provided in substantially the same manner as the dehydration weight loss curve of the first reference calcium sulfate sample used to train the test model.

[0382] Embodiment 45. The method according to any one of embodiments 38 to 44, wherein the dehydration weight loss curve of the test sample is provided by a method as described in any one of embodiments 9 to 11.

[0383] Embodiment 46. The method of any one of Embodiments 38 to 45, wherein the first plurality of parameters of the test calcium sulfate sample are provided in substantially the same manner as the first plurality of parameters of the first reference calcium sulfate sample used to train the test model.

[0384] Embodiment 47. The method of any one of Embodiments 38 to 46, wherein the first plurality of parameters is provided as described in any one of Embodiments 12 to 17.

[0385] Embodiment 48. The method of any one of Embodiments 38 to 47, wherein the hydration temperature rise curve of the test sample is provided in substantially the same manner as the hydration temperature rise curve of the first reference calcium sulfate sample used to train the test model.

[0386] Embodiment 49. The method according to any one of embodiments 38 to 48, wherein the hydration temperature rise curve of the test sample is provided by the method according to any one of embodiments 18 to 20.

[0387] Embodiment 50. The method of any one of Embodiments 38 to 49, wherein the second plurality of parameters of the test calcium sulfate sample is provided in substantially the same manner as the second plurality of parameters of the second reference calcium sulfate sample used to train the test model.

[0388] Embodiment 51. The method of any one of Embodiments 38 to 50, wherein the second plurality of parameters is provided as described in any one of Embodiments 21 to 26.

[0389] Embodiment 52. A method for providing a stucco raw material, the method comprising:

[0390] providing a first stucco material;

[0391] determining the phase content of the first stucco material according to the method of any one of embodiments 38 to 51; and

[0392] Based on the determined phase contents, the first stucco material is modified to provide a stucco feedstock having a desired phase content.

[0393] Embodiment 53. The method of embodiment 52, wherein the modification of the first stucco material comprises combining the first stucco material with one or more second stucco materials to provide a stucco feedstock having a desired phase content, the second stucco materials each having a different phase content than the first stucco material.

[0394] Embodiment 54. The method of Embodiment 52 or Embodiment 53, wherein the modification of the first stucco material comprises calcining the first stucco material.

[0395] Embodiment 55. The method of any one of Embodiments 52 to 54, wherein the modification of the first stucco material comprises adding moisture to the first stucco material.

[0396] Embodiment 56. The method of any one of Embodiments 52 to 55, wherein modification of the first stucco material comprises adjusting the particle size distribution of the first stucco material (eg, by adjusting grinding conditions).

[0397] Embodiment 57. A method for providing a stucco raw material, the method comprising:

[0398] calcining a first sample of calcium sulfate material using a first set of calcination parameters to provide a first calcined stucco;

[0399] Determining the phase content of the first calcined stucco according to the method of any one of embodiments 38 to 51;

[0400] Based on the determined phase contents, a second sample of the calcium sulfate material is calcined using a second set of calcination parameters that are different from the first set of calcination parameters to provide the stucco feedstock.

[0401] Embodiment 58. The method of Embodiment 57, wherein the second group has a higher calcination temperature program than the first group.

[0402] Embodiment 59. The method of Embodiment 57, wherein the second group has a lower calcination temperature program than the first group.

[0403] Embodiment 60. The method of any one of Embodiments 57 to 59, wherein the second group has a higher calcination moisture than the first group.

[0404] Embodiment 61. The method of any one of Embodiments 57 to 59, wherein the second group has a lower calcination moisture than the first group.

[0405] Embodiment 62. The method of any one of Embodiments 57 to 61, wherein the second group has a higher calciner throughput than the first group.

[0406] Embodiment 63. The method of any one of Embodiments 57 to 61, wherein the second group has a lower calciner throughput than the first group.

[0407] Embodiment 64. The method of any one of Embodiments 57 to 63, wherein the second group has a longer calcination time than the first group.

[0408] Embodiment 65. The method of any one of Embodiments 57 to 63, wherein the second group has a shorter calcination time than the first group.

[0409] Embodiment 66. A method for providing a stucco feedstock, the method comprising:

[0410] calcining a first calcium sulfate material using a first set of calcining parameters to provide a first calcined stucco;

[0411] Determining the phase content of the first calcined stucco according to the method of any one of embodiments 38 to 51;

[0412] Based on the determined phase contents, a second calcium sulfate material different from the first calcium sulfate material is selected and calcined using a second set of calcination parameters to provide the stucco feedstock.

[0413] Embodiment 67. The method of Embodiment 66, wherein the second calcium sulfate material has a different composition than the first calcium sulfate material.

[0414] Embodiment 68. A method according to embodiment 67, wherein the second calcium sulfate material has significantly different relative amounts of one or more of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM) compared to the first calcium sulfate material, for example, one or more of which differ by greater than 2 weight % or greater than 5 weight %.

[0415] Embodiment 69. The method of any one of Embodiments 66 to 68, wherein the second calcium sulfate material is processed differently from the first calcium sulfate material.

[0416] Embodiment 70. A method according to any one of Embodiments 66 to 69, wherein the second calcium sulfate material has a significantly different particle size distribution compared to the first material (e.g., a significantly different d50 value, e.g., differing by at least 10% from the d50 of the first calcium sulfate material).

[0417] Embodiment 71. A method according to Embodiment 70, wherein the second calcium sulfate material has a d50 value that is significantly greater than the d50 value of the first calcium sulfate material (e.g., at least 10% greater than the d50 of the first calcium sulfate material).

[0418] Embodiment 72. A method according to Embodiment 70, wherein the second calcium sulfate material has a d50 value that is significantly smaller than the d50 value of the first calcium sulfate material (e.g., at least 10% smaller than the d50 of the first calcium sulfate material).

[0419] Embodiment 73. The method of any one of Embodiments 66 to 72, wherein the second calcium sulfate material has a substantially similar composition to the first calcium sulfate material (e.g., one or more of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); the relative amounts of inert calcium sulfate (IN) and free moisture (FM) are not significantly different, e.g., each of these differs by no more than 2 wt. % or no more than 1 wt. %).

[0420] Embodiment 74. The method of any one of Embodiments 66 to 73, wherein the second group has a higher calcination temperature program than the first group.

[0421] Embodiment 75. The method of any one of Embodiments 66 to 73, wherein the second group has a lower calcination temperature schedule than the first group.

[0422] Embodiment 76. The method of any one of Embodiments 66 to 75, wherein the second group has a higher calcination moisture than the first group.

[0423] Embodiment 77. The method of any one of Embodiments 66 to 75, wherein the second group has a lower calcination moisture than the first group.

[0424] Embodiment 78. The method of any one of Embodiments 66 to 77, wherein the second group has a higher calciner throughput humidity than the first group.

[0425] Embodiment 79. The method of any one of Embodiments 66 to 77, wherein the second group has a lower calciner throughput than the first group.

[0426] Embodiment 80. The method of any one of Embodiments 66 to 79, wherein the second group has a longer calcination time than the first group.

[0427] Embodiment 81. The method of any one of Embodiments 66 to 79, wherein the second group has a shorter calcination time than the first group.

[0428] Embodiment 82. A method for producing a gypsum product, the method comprising:

[0429] Providing a stucco raw material by the method according to any one of embodiments 52 to 81;

[0430] hydrating the first stucco raw material in an aqueous slurry; and

[0431] allowing the slurry to solidify; and

[0432] The set slurry is dried to form a gypsum product.

[0433] Embodiment 83. A production method for providing a gypsum product, the method comprising:

[0434] providing a first plaster raw material;

[0435] Determining the phase content of the first stucco raw material according to the method of any one of embodiments 38 to 51;

[0436] hydrating the first stucco raw material in an aqueous slurry;

[0437] allowing the slurry to solidify; and

[0438] drying the set slurry to form the gypsum product,

[0439] The method also includes adjusting one or more parameters of the production process based on the determined phase content of the first stucco raw material.

[0440] Embodiment 84. The production method according to embodiment 83, wherein adjusting one or more parameters of the production method comprises adjusting the content of the slurry,

[0441] Embodiment 85. A production method according to Embodiment 84, wherein adjusting the content of the slurry includes selecting the amount of a promoter (e.g., powdered calcium sulfate dihydrate or potassium sulfate) based on the measured phase content.

[0442] Embodiment 86. A production method according to Embodiment 84 or Embodiment 85, wherein adjusting the content of the slurry includes selecting the amount of retarder (e.g., diethylenetriaminepentaacetic acid and its salts) based on the measured phase content.

[0443] Embodiment 87. A production method according to any one of Embodiments 83 to 86, wherein adjusting the content of the slurry includes selecting the water demand of the slurry (i.e., the relative amount of water compared to the stucco raw material) based on the determined phase content.

[0444] Embodiment 88. The production method of any one of Embodiments 83 to 87, wherein adjusting the content of the slurry comprises selecting the density of the gypsum product (e.g., by board weight) based on the determined phase content.

[0445] Embodiment 89. The production method of any one of Embodiments 82 to 88, wherein adjusting one or more parameters of the production method comprises adjusting one or more parameters associated with drying.

Claims

1. A method for providing a computer-implemented system for determining the phase content of a test calcium sulfate material, the method comprising: Providing a plurality of first reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM); For each of the plurality of first reference calcium sulfate samples, providing: The dehydration weight loss curve associated with the first reference calcium sulfate sample a first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of a dehydration weight loss curve of the first reference calcium sulfate sample, and The hydration temperature rise curve associated with the first reference calcium sulfate sample a second plurality of parameters, the second plurality of parameters comprising a second plurality of fit parameters associated with a fit of a hydration curve of the first reference calcium sulfate sample, providing the first plurality of parameters and the second plurality of parameters for each of the first reference calcium sulfate samples to a computing device; The first plurality of parameters and the second plurality of parameters are associated with the known contents of DH, HH, AIII, IN and FM of the first reference calcium sulfate sample via the computing device, thereby providing a test model that associates the first plurality of parameters and the second plurality of parameters of the test calcium sulfate sample with the contents of one or more of DH, HH, AIII, IN and FM of the test sample.

2. The method of claim 1, wherein the plurality of first reference calcium sulfate samples comprises at least 300 first reference calcium sulfate samples, such as at least 400 first reference calcium sulfate samples.

3. The method according to claim 1, wherein For at least 80% (eg, at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of DH is in the range of 3 wt % to 20 wt %. For at least 80% (eg, at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of HH is in the range of 50 wt % to 90 wt %. For at least 80% (eg, at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of AIII is up to 30 wt%. For at least 80% (eg, at least 90% or at least 95%) of the plurality of first reference calcium sulfate samples, the known amount of IN is in the range of 5 wt % to 10 wt %.

4. The method of claim 1, wherein the dehydration weight loss curve for each first reference calcium sulfate sample is provided by thermogravimetrically measuring the weight loss of the first reference calcium sulfate sample at a temperature in the range of at least 100°C, such as at least 200°C, or at least 250°C, such as in the range of 100°C to 500°C, or 200°C to 500°C, or 250°C to 500°C, or 100°C to 400°C, or 200°C to 400°C, or 250°C to 400°C.

5. The method according to claim 1, wherein the fitting of the dehydration weight loss curve is the fitting of the moisture change rate over time. The method of claim 1 , wherein the first plurality of parameters comprises at least four parameters.

7. The method of claim 1, wherein a hydration temperature rise curve for each first reference calcium sulfate sample is provided by measuring a time-varying temperature rise of a mixture of the first reference calcium sulfate sample and water.

8. The method of claim 1, wherein the hydration temperature rise curve is fit to a combination of Gaussian functions, such as a combination of a positively skewed Gaussian function and a non-skewed Gaussian function.

9. The method of claim 1, wherein the second plurality of fitting parameters comprises at least four parameters, such as at least five parameters.

10. The method of claim 1, wherein the second plurality of parameters comprises at least four parameters, such as at least five parameters.

11. The method of claim 1 , wherein the computing system comprises a processor programmed to execute a machine learning algorithm, the machine learning algorithm being an artificial neural network algorithm.

12. The method according to claim 1, further comprising: providing a plurality of second reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM); For each of the plurality of second reference calcium sulfate samples, providing: The dehydration weight loss curve associated with the second reference calcium sulfate sample a third plurality of parameters, the third plurality of parameters comprising a third plurality of fitting parameters associated with a fit of a dehydration weight loss curve of the second reference calcium sulfate sample, and The hydration temperature rise curve associated with the second reference calcium sulfate sample a fourth plurality of parameters comprising a fourth plurality of fit parameters associated with a fit of a hydration curve of the second reference calcium sulfate sample; for each of the plurality of second reference calcium sulfate samples, providing the third plurality of parameters and the fourth plurality of parameters to the test model; as well as Using the test model, the content of one or more of DH, HH, AIII, IN and FM of each second reference calcium sulfate sample is determined.

13. A method for determining the phase content of a test calcium sulfate material, the method comprising: Providing a dehydration weight loss curve for the test calcium sulfate material; Providing a hydration temperature rise curve of the tested calcium sulfate material; supply: a first plurality of parameters associated with a dehydration weight loss curve of the test calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the test calcium sulfate sample, and a second plurality of parameters associated with a hydration temperature rise curve of the test calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of a hydration curve of the test calcium sulfate sample; providing a test model that relates the first plurality of parameters and the second plurality of parameters of a test calcium sulfate sample to the content of one or more of DH, HH, AIII, IN, and FM of the test sample; providing the first plurality of parameters and the second plurality of parameters to the test model; and Using the test model, the content of one or more of DH, HH, AIII, IN and FM of the test calcium sulfate sample is determined.

14. The method of claim 13, wherein the test model is created by a method comprising the steps of: Providing a plurality of first reference calcium sulfate samples, each sample having a known content of calcium sulfate dihydrate (DH), calcium sulfate hemihydrate (HH), soluble calcium sulfate anhydrite (AIII); inert calcium sulfate (IN) and free moisture (FM); For each of the plurality of first reference calcium sulfate samples, providing: a first plurality of parameters associated with a dehydration weight loss curve of the first reference calcium sulfate sample, the first plurality of parameters comprising a first plurality of fitting parameters associated with a fit of the dehydration weight loss curve of the first reference calcium sulfate sample, and a second plurality of parameters associated with the hydration temperature rise curve of the first reference calcium sulfate sample, the second plurality of parameters comprising a second plurality of fitting parameters associated with a fit of the hydration curve of the first reference calcium sulfate sample, providing the first plurality of parameters and the second plurality of parameters for each of the first reference calcium sulfate samples to a computing device; and The first plurality of parameters and the second plurality of parameters are associated with the known contents of DH, HH, AIII, IN and FM of the first reference calcium sulfate sample via the computing device, thereby providing a test model that associates the first plurality of parameters and the second plurality of parameters of the test calcium sulfate sample with the contents of one or more of DH, HH, AIII, IN and FM of the test sample.

15. A method for providing a plaster raw material, the method comprising: Provide first plaster material ； Determining the phase content of the first stucco material according to the method of claim 13 or claim 14; and Based on the determined phase contents, the first stucco material is modified to provide a stucco feedstock having a desired phase content.

16. A method for providing a plaster raw material, the method comprising: A first sample of calcium sulfate material is calcined using a first set of calcination parameters to provide a first calcined stucco ； Determining the phase content of the first calcined mortar according to the method of claim 13 or claim 14; and Based on the determined phase contents, a second sample of the calcium sulfate material is calcined using a second set of calcination parameters different from the first set of calcination parameters to provide the stucco feedstock.

17. A method for providing a plaster raw material, the method comprising: The first calcium sulfate material is calcined using a first set of calcination parameters to provide a first calcined stucco ； Determining the phase content of the first calcined mortar according to the method of claim 13 or claim 14; Based on the determined phase contents, a second calcium sulfate material different from the first calcium sulfate material is selected and calcined using a second set of calcination parameters to provide the stucco feedstock.

18. A method for producing a gypsum product, the method comprising: Providing a stucco raw material by the method according to any one of claims 15 to 17; hydrating the first stucco raw material in an aqueous slurry; as well as allowing the slurry to solidify; as well as The set slurry is dried to form the gypsum product.

19. A method for providing a gypsum product, the method comprising: providing a first plaster raw material; Determining the phase content of the first plaster raw material according to the method of claim 13 or claim 14; hydrating the first stucco raw material in an aqueous slurry; allowing the slurry to solidify; as well as drying the set slurry to form the gypsum product, The method also includes adjusting one or more parameters of the production process based on the determined phase content of the first stucco raw material.

20. The production method according to claim 19, wherein adjusting one or more parameters of the production method comprises adjusting the content of the slurry, wherein adjusting the content of the slurry comprises one or more of the following: Selecting the amount of accelerator (e.g., powdered calcium sulfate dihydrate or potassium sulfate) based on the determined phase content; selecting the amount of retarder (e.g., diethylenetriaminepentaacetic acid and its salts) based on the determined phase content; and The water demand of the slurry (ie, the relative amount of water compared to the stucco raw material) is selected based on the determined phase contents.

21. The production method of claim 19, wherein adjusting one or more parameters of the production method comprises adjusting one or more parameters associated with drying.

Images