Online quantitative model of moisture content in hydroxychloroquine sulfate granules using near-infrared spectroscopy and its establishment method and detection method

By establishing a near-infrared spectral online quantitative model of hydroxychloroquine sulfate particles, the shortcomings of offline detection in the existing technology are solved, and the rapid and accurate detection of moisture of hydroxychloroquine sulfate particles is achieved, and the production efficiency and product quality are improved.

CN113834795BActive Publication Date: 2024-12-03SHANGHAI PHARMACEUTICALS HOLDING CO LTD +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010514508.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-06-08
Publication Date
2024-12-03
Estimated Expiration
2040-06-08

AI Technical Summary

Technical Problem

In the prior art, near-infrared spectroscopy analysis technology is used to detect the moisture content in hydroxychloroquine sulfate particles limited to offline mode, and the detection results cannot be measured quickly and in real time, which is not conducive to improving production efficiency.

Method used

A near-infrared spectral online quantitative model of moisture of hydroxychloroquine sulfate particles was established. By collecting near-infrared spectral samples during the drying process, removing abnormal samples, selecting appropriate modeling bands and main factors, and establishing them using partial least squares regression. and monitor moisture content in a fluidized bed in real time.

Benefits of technology

It realizes rapid and accurate detection of moisture of hydroxychloroquine sulfate particles, efficient and convenient monitoring of drying endpoints, the correlation coefficient R of the online quantitative model reaches more than 0.9, and has good error stability, and is suitable for industrial production.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113834795B_ABST
    Figure CN113834795B_ABST
Patent Text Reader

Abstract

The invention discloses a method for establishing a near infrared spectrum online quantitative model of the moisture content of hydroxychloroquine sulfate particles, the established model, and a method for detecting the moisture content of hydroxychloroquine sulfate particles. The steps of the establishment method are as follows: collecting near infrared spectrum samples of hydroxychloroquine sulfate particles in the drying process, sampling after each collection, and detecting the moisture content of the samples obtained by sampling; removing abnormal samples from the collected near infrared spectrum samples, and dividing the remaining near infrared spectrum samples after removing the abnormal samples into a calibration set and a verification set; selecting a modeling band of 4800-5500cm ‑1 and 6800~7400cm ‑1 , the number of principal factors is selected to be 5 to 7, and the partial least squares regression is used to establish an online quantitative model. The online quantitative model of the present invention has high accuracy and good stability, can be applied to industrial production, realizes rapid and accurate detection of the moisture content of hydroxychloroquine sulfate particles, and efficiently and conveniently monitors the drying end point.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a method for establishing a near-infrared spectrum online quantitative model of the moisture content of hydroxychloroquine sulfate particles, the established model, and a method for detecting the moisture content of hydroxychloroquine sulfate particles. Background Art

[0002] Hydroxychloroquine sulfate is a commonly used drug for the clinical treatment of rheumatoid arthritis and systemic lupus erythematosus. An important quality control point in the production process of hydroxychloroquine sulfate granules is hydroxychloroquine sulfate granulation. The process of hydroxychloroquine sulfate granulation is wet granulation. The quality control of wet granulation mainly detects a series of indicators such as granule moisture, mixing uniformity and particle size during the granulation process. Among them, the moisture content affects both the particle size distribution and the degree of mixing, so it needs to be measured first. At present, hydroxychloroquine sulfate granules use a rapid moisture meter to measure the moisture content during the drying process to monitor the drying end point. This analytical method samples and measures the moisture during the drying process, which will bring about the influence of human interference, and the measurement time of the moisture meter is long. While waiting for the measurement results, the moisture content of the material in the fluidized bed is still changing. Therefore, a more efficient and accurate method for detecting the moisture content of hydroxychloroquine sulfate granules is urgently needed.

[0003] Near infrared spectroscopy (NIRS) analysis technology is one of the core technologies of process analytical technology (PAT). It can detect key quality data such as raw material quality, intermediate material quality and process parameters to ensure the quality of finished products. It has the characteristics of convenient testing, fast analysis speed and high analysis efficiency.

[0004] Using NIR analysis technology to monitor the moisture content of hydroxychloroquine sulfate granules in real time will help improve the production efficiency of hydroxychloroquine sulfate and the quality of the final product, and contribute to the future development of full automation of production. At present, the report of using NIR analysis technology to monitor the moisture content of hydroxychloroquine sulfate granules is limited to the establishment of the offline quantitative model of hydroxychloroquine sulfate granules moisture (Li Shasha, Chen Hui, Zhao Yunli, etc. Hydroxychloroquine sulfate granules moisture content determination near infrared quantitative model establishment [J]. Journal of Shenyang Pharmaceutical University, 2019, 036 (007): 593-599.). However, offline detection is not conducive to improving the production efficiency of hydroxychloroquine sulfate granules and the quality of the final product, and is not conducive to the development of the full automation direction of hydroxychloroquine sulfate granules production. Therefore, a method for establishing a near infrared spectroscopy online quantitative model of moisture in hydroxychloroquine sulfate granules is urgently needed. Summary of the invention

[0005] The problem to be solved by the present invention is to overcome the defect that the water content in hydroxychloroquine sulfate granules detected by near infrared spectroscopy analysis technology in the prior art is limited to an offline mode, the test result cannot be measured quickly and in real time, which is not conducive to improving production efficiency, and a method for establishing a near infrared spectroscopy online quantitative model of the water content of hydroxychloroquine sulfate granules is provided, and the model built, and a detection method for the water content of hydroxychloroquine sulfate granules. The online quantitative model of the present invention has higher accuracy and better stability, can be applied to industrial production, so as to realize fast and accurate detection of the water content of hydroxychloroquine sulfate granules, and efficient and convenient monitoring drying end point.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] One of the technical solutions provided by the present invention is: a method for establishing an online quantitative model of near-infrared spectroscopy of moisture in hydroxychloroquine sulfate particles. The establishment method comprises the following steps:

[0008] (1) collecting near infrared spectrum samples of hydroxychloroquine sulfate particles in the drying process, and taking samples after each collection, and detecting the moisture content of the samples obtained by sampling; wherein, when the drying temperature is below A, the sampling time interval is 30 to 32 seconds, and when the drying temperature is greater than A, the sampling time interval is 115 to 125 seconds; A is 34 to 36°C;

[0009] (2) removing abnormal samples from the collected near-infrared spectrum samples, and dividing the remaining near-infrared spectrum samples after removing the abnormal samples into a calibration set and a validation set;

[0010] (3) Select the modeling band as 4800-5500cm -1 and 6800~7400cm -1 The number of principal factors was selected as 5 to 7, and partial least squares regression was used to establish an online quantitative model of near-infrared spectroscopy of moisture in hydroxychloroquine sulfate particles.

[0011] In step (1), the hydroxychloroquine sulfate granules are granules prepared in the hydroxychloroquine sulfate preparation production process, generally hydroxychloroquine sulfate wet granules prepared by wet granulation. The drying process is carried out in a conventional drying device in the art, generally a fluidized bed.

[0012] The specific equipment setting for collecting near-infrared spectrum samples of the hydroxychloroquine sulfate particles in the drying process and taking samples after each collection is completed can be: spectral windows (such as sapphire windows) and sampling ports are set on both sides of the lower part of the fluidized bed boiler body, which can be symmetrically arranged. Preferably, the spectral window is close to the sampling port to ensure the spatial correspondence between the spectrum and the sample.

[0013] During the drying process, the hydroxychloroquine sulfate particles are placed in the fluidized bed boiler body. After the fluidized bed is air-dried, the near-infrared spectrum samples can be collected by the near-infrared spectrometer probe fixed on the sapphire window. After each collection, samples are taken from the sampling port, preferably immediately, to ensure the time correspondence between the spectrum and the sample. In order to avoid the uneven drying process of the fluidized bed to the material as much as possible, the difference between the moisture value of the material affects the modeling, the air inlet temperature and frequency can be controlled, the material temperature rises slowly, the agglomeration phenomenon is alleviated, and the uniformity of the fluidized bed to the material drying process is improved. In order to avoid the accumulation of wet materials at the sampling port and affect the moisture value of the sample, the materials accumulated at the sampling port can be cleaned up in time.

[0014] The near infrared spectrometer is generally a near infrared spectrometer with a probe, such as a Thermo Antaris MX near infrared spectrometer. The probe of the near infrared spectrometer is preferably a diffuse reflection optical fiber probe.

[0015] The scanning range for collecting near-infrared spectrum samples can be the wavelength range of the near-infrared spectrometer used. For example, when the near-infrared spectrometer used is Thermo Antaris MX, the scanning range can be 10000-4000 cm -1 .

[0016] The number of scans, resolution and gain value for collecting near-infrared spectrum samples can be selected according to the conventional methods in the art. Preferably, the number of scans is 64 and the resolution is 8 cm -1 , the gain value is 8×.

[0017] Before collecting the near-infrared spectrum sample each time, it is preferred to collect a background spectrum.

[0018] As the drying process proceeds, the operation of collecting near-infrared spectrum samples and then sampling is performed continuously for multiple times. The A is preferably 35°C. The sampling time interval is preferably: when the drying temperature is below 35°C, the sampling time interval is 30s, and when the drying temperature is greater than 35°C, the sampling time interval is 120s. The temperature at the drying end point can be above 48°C, preferably 50°C.

[0019] The near infrared spectrum samples collected can be targeted at multiple batches of hydroxychloroquine sulfate particles as much as possible, to include more batch differences, improve the applicability of the model in actual production. The batches of preferred hydroxychloroquine sulfate particles are more than 10 batches. Usually, the number of near infrared spectrum samples collected in each batch is more than 40. In one embodiment of the present invention, 12 batches are collected, a total of 519 near infrared spectrum samples.

[0020] The instrument for detecting the moisture content of the sample obtained by sampling can be a conventional moisture meter in the art, such as HX204 halogen moisture meter. The moisture content of the sample is generally 1.00-6.50%, covering the drying process optimization range of 1.00-2.50%.

[0021] In step (2), eliminating abnormal samples from the collected near-infrared spectral samples can improve the accuracy and stability of the online quantitative model.

[0022] The method for removing abnormal samples from the collected near-infrared spectrum samples can be a conventional method in the art, such as a Chauvelet criterion test combined with a leverage value-studentized residual map. The number of abnormal samples is generally more than 10.

[0023] The classification method of the calibration set and the validation set may be a conventional classification method in the art, such as the SPXY classification method or the y-equal spacing method.

[0024] The ratio of the calibration set to the validation set may be 2:1 to 3:1, for example, 2.1:1.

[0025] The calibration set can be used to establish the online quantitative model. The number of the calibration set can be more than 300. The validation set can be used to validate the model and evaluate the accuracy of the model. The number of the validation set can be more than 100.

[0026] Step (2) preferably includes preprocessing the near-infrared spectrum samples remaining after removing the abnormal samples.

[0027] The preprocessing method may be to separate the scattering signal from the chemical information in the near infrared spectrum, or to separate the scattering signal from the chemical information in the near infrared spectrum, enhance the spectral signal and filter out noise.

[0028] Wherein, the separation of the scattered signal and chemical information in the near-infrared spectrum can eliminate the influence of the scattering caused by the uneven distribution of particles on its spectrum, and the method used can be multiplicative signal correction (MSC) or standard normal variate transformation (SNV). The method of enhancing the spectral signal can be first-order derivative or second-order derivative, preferably first-order derivative. The method of filtering out noise can be Savitzky-Golay smoothing filter or Karl Norris smoothing filter, preferably Karl Norris smoothing filter.

[0029] The preprocessing method is preferably MSC, or SNV, or MSC, first-order derivative and Karl Norris smoothing filter, or SNV, first-order derivative and Karl Norris smoothing filter; more preferably MSC, first-order derivative and Karl Norris smoothing filter, or SNV, first-order derivative and KarlNorris smoothing filter.

[0030] In step (3), the modeling band is preferably 4935-5336 cm -1 and 6911~7297cm -1 The number of principal factors is preferably 6.

[0031] The number of principal factors can be selected according to conventional methods in the art, for example, according to the PRESS diagram in TQAnlyst9.5, the number of principal factors corresponding to the minimum root mean square error of cross validation (RMSECV) and PRESS is selected. Selecting the correct number of principal factors can fully utilize the information of near-infrared spectra and avoid overfitting in the process of establishing an online quantitative model.

[0032] The software for establishing the online quantitative model may be conventional model building software in the art, such as MATLAB 2014a or TQAnlyst 9.5.

[0033] The second technical solution provided by the present invention is: an online quantitative model of the near-infrared spectrum of the moisture content of hydroxychloroquine sulfate particles established by the above-mentioned establishment method.

[0034] The online quantitative model of the present invention can be evaluated by the conventional method in this area. For example: the correlation coefficient R of the online quantitative model is greater than 0.9, indicating that the accuracy of the online quantitative model is higher. The data points of the calibration set and the validation set are evenly distributed within the model range, and the calibration set is distributed within the range of the validation set, indicating that the model established with the calibration set is effective for the validation of the validation set. The absolute value of the difference between the validation set error root mean square and the calibration set error root mean square (|RMSEP-RMSEC|) approaches 0, indicating that the online quantitative model has better stability. The ratio (RPD) of the validation set standard deviation (SDV) and the prediction set standard deviation (SEP) is greater than 5, indicating that the prediction result of the online quantitative model is acceptable. The moisture content (predicted value) of the hydroxychloroquine sulfate granules obtained by the online quantitative model and the moisture content (reference value) of the hydroxychloroquine sulfate granules measured by the moisture meter are subjected to paired sample t test to show that there is no statistical difference, indicating that the error of the online quantitative model is small and the stability is better.

[0035] The third technical solution provided by the present invention is: a method for detecting the moisture content of hydroxychloroquine sulfate particles.

[0036] The detection method comprises the following steps:

[0037] (1) collecting near infrared spectra of hydroxychloroquine sulfate particles during the drying process;

[0038] (2) importing the near infrared spectrum collected in step (1) into the online quantitative model of the near infrared spectrum of the moisture of the hydroxychloroquine sulfate particles to obtain the moisture content of the hydroxychloroquine sulfate particles.

[0039] In step (1), the scanning range for collecting the near-infrared spectrum can be the wavelength range of the near-infrared spectrometer used. For example, when the near-infrared spectrometer used is Thermo Antaris MX, the scanning range can be 10000-4000cm -1 .

[0040] The resolution and gain value of the near infrared spectrum can be selected according to the conventional method in the art. Preferably, the resolution is 8 cm -1 , the gain value is 8×.

[0041] The number of scans for collecting the near infrared spectrum may be 2 to 3 times.

[0042] In step (2), the moisture content of the hydroxychloroquine sulfate particles can be used to monitor the drying end point. For example, when the moisture content of the hydroxychloroquine sulfate particles is 1.00-2.50%, drying can be stopped.

[0043] On the basis of being in accordance with the common sense in the art, the above-mentioned preferred conditions can be arbitrarily combined to obtain the preferred embodiments of the present invention.

[0044] The reagents and raw materials used in the present invention are commercially available.

[0045] The positive and progressive effects of the present invention are: the near-infrared spectroscopy online quantitative model of the moisture content of hydroxychloroquine sulfate particles of the present invention has high accuracy and good stability, the correlation coefficient R of the online quantitative model is above 0.9, |RMSEP-RMSEC| is close to 0, and RPD is above 5. The online quantitative model of the present invention is applied to industrial production, and the t-test shows no statistical difference; the moisture content of hydroxychloroquine sulfate particles can be quickly and accurately detected, and the drying end point can be monitored efficiently and conveniently. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The near-infrared spectrogram of the hydroxychloroquine sulfate particles in the drying process collected in Example 1.

[0047] Figure 2 The near infrared spectroscopy online quantitative model of the moisture content of hydroxychloroquine sulfate particles established by the method of Example 1.

[0048] Figure 3The figure is a moisture change trend diagram of the predicted value and reference value of the moisture content of the hydroxychloroquine sulfate particles in Example 6.

[0049] Figure 4 The figure is a trend diagram of the moisture content of the hydroxychloroquine sulfate particles detected in Example 7.

[0050] Figure 5 The near-infrared spectroscopy online quantitative model of the moisture content of hydroxychloroquine sulfate particles established by the method of Comparative Example 1. DETAILED DESCRIPTION

[0051] The present invention is further described below by way of examples, but the present invention is not limited to the scope of the examples. The experimental methods in the following examples without specifying specific conditions are carried out according to conventional methods and conditions, or selected according to the product specifications.

[0052] Example 1 Establishment of near infrared spectroscopy online quantitative model for moisture content of hydroxychloroquine sulfate particles

[0053] (1) Collecting near-infrared spectra and sampling

[0054] Hydroxychloroquine sulfate granules (Shanghai Shangyao Zhongxi Pharmaceutical Co., Ltd.) prepared by wet granulation in the production process of hydroxychloroquine sulfate preparations were placed in a fluidized bed for drying. After the fluidized bed began to be air-dried, near-infrared spectral samples (such as Figure 1 ), and take samples after each collection is completed, and test the moisture content of the samples obtained.

[0055] A sapphire window is installed in the lower part of the fluidized bed pot, and the probe of the near-infrared spectrometer is fixed on the window. The sampling port is on the other side of the fluidized bed, close to the sapphire window. The near-infrared spectrometer is a ThermoAntaris MX near-infrared spectrometer, and the probe is a diffuse reflection fiber probe; the scanning range is 10000~4000cm -1 , the number of scans is 64 times, the resolution is 8cm -1 , and the gain value is 8×. A background spectrum is collected before collecting the near-infrared spectrum sample each time.

[0056] When the drying temperature is below 35°C, the sampling time interval is 30s; when the drying temperature is greater than 35°C, the sampling time interval is 120s. A batch of hydroxychloroquine sulfate granules is divided into 4 fluidized beds to complete drying. Each fluidized bed collects about 10 near-infrared spectral samples. A batch can collect about 40 near-infrared spectral samples. A total of 12 batches are collected, totaling 519 near-infrared spectral samples; sampling is taken after each collection, and a total of 519 hydroxychloroquine sulfate granule samples are taken. The moisture content of the samples obtained by sampling is detected by HX204 halogen moisture analyzer and is 1.00-6.50% (reference value).

[0057] (2) The Chauville criterion test combined with the leverage value-studentized residual map was used to eliminate 13 abnormal spectra (23, 58, 61, 70, 79, 95, 96, 97, 139, 223, 276, 350 and 473) from the near-infrared spectrum samples collected in step (1).

[0058] The near-infrared spectra of hydroxychloroquine sulfate granule samples after removing abnormal spectra were divided into a calibration set and a validation set using the SPXY classification method, of which the calibration set consisted of 344 samples and the validation set consisted of 162 samples. Principal component analysis showed that the validation set was evenly distributed in the calibration set, indicating that the calibration set covered the moisture content range of the validation set and the division was reasonable.

[0059] The remaining near-infrared spectrum samples after removing abnormal samples are preprocessed, wherein the preprocessing methods are MSC, first-order derivative and Karl Norris smoothing filter.

[0060] (3) Select the modeling band as 4935-5336cm -1 ; Select the value 6 corresponding to the minimum of the root mean square error of interactive validation (RMSECV) and PRESS in the PRESS diagram of TQAnlyst9.5 as the main factor number; Use partial least squares regression to establish the online quantitative model of near infrared spectroscopy of moisture in hydroxychloroquine sulfate particles (such as Figure 2 ).

[0061] Example 2 Establishment of near infrared spectroscopy online quantitative model for moisture content of hydroxychloroquine sulfate particles

[0062] In the method for establishing the online quantitative model in Example 2, in step (2), the near-infrared spectral samples remaining after removing the abnormal samples are preprocessed, wherein the preprocessing methods are SNV, first-order derivative and Karl Norris smoothing filtering; the rest are the same as in Example 1.

[0063] Example 3 Establishment of near infrared spectroscopy online quantitative model for moisture content of hydroxychloroquine sulfate particles

[0064] In the method for establishing the online quantitative model in Example 3, in step (2), the near-infrared spectral samples remaining after the abnormal samples are removed are preprocessed, wherein the preprocessing method is MSC; the rest are the same as in Example 1.

[0065] Example 4 Establishment of near infrared spectroscopy online quantitative model for moisture content of hydroxychloroquine sulfate particles

[0066] In the method for establishing the online quantitative model in Example 4, in step (2), the near-infrared spectral samples remaining after the abnormal samples are removed are preprocessed, wherein the preprocessing method is SNV; the rest are the same as in Example 1.

[0067] Example 5 Establishment of near infrared spectroscopy online quantitative model for moisture content of hydroxychloroquine sulfate particles

[0068] In the method for establishing the online quantitative model in Example 5, in step (2), the near-infrared spectral samples remaining after the abnormal samples are removed are not preprocessed; the rest are the same as in Example 1.

[0069] Comparative Example 1 Establishment of near infrared spectroscopy online quantitative model for moisture content of hydroxychloroquine sulfate particles

[0070] In Comparative Example 1, the sampling time interval is 120s; the rest are the same as in the embodiment.

[0071] from Figure 5 It can be seen that the online quantitative model in comparative example 1 is biased towards the side with low moisture content, and the prediction results will be biased.

[0072] Effect Example 1

[0073] (1) The online quantitative models in Examples 1 to 5 were evaluated for accuracy. The results are shown in Table 1.

[0074] Table 1

[0075] Example RMSEC / % RMSEP / % RMSECV / % R Example 1 0.408 0.435 0.446 0.9372 Example 2 0.405 0.436 0.446 0.9372 Example 3 0.467 0.409 0.481 0.9309 Example 4 0.469 0.411 0.483 0.9306 Example 5 0.457 0.499 0.489 0.9206

[0076] In Examples 1 to 5, the absolute value of the difference between the root mean square error of the validation set and the root mean square error of the calibration set (|RMSEP-RMSEC|) is close to 0, indicating that the established online quantitative model has good stability.

[0077] like Figure 2 As shown, the data points of the calibration set and the validation set in Example 1 are evenly distributed within the model range, and the calibration set is distributed within the range of the validation set, indicating that the online quantitative model established with the calibration set is effective for the validation set. And the ratio (RPD) of the validation set standard deviation (SDV) to the prediction set standard deviation (SEP) of Example 1 is 5.18, which is greater than 5, indicating that the prediction results of the online quantitative model are acceptable.

[0078] Example 6 Detection of moisture content of hydroxychloroquine sulfate granules

[0079] The online quantitative model in Example 1 was applied to detect the moisture content of hydroxychloroquine sulfate particles in the laboratory:

[0080] A near infrared spectrum was collected for a batch of hydroxychloroquine sulfate particles in the drying process; the collected near infrared spectrum was introduced into the online quantitative model in Example 1 to obtain the moisture content (predicted value) of the hydroxychloroquine sulfate particles; wherein, when collecting the near infrared spectrum, the near infrared spectrometer was a Thermo Antaris MX near infrared spectrometer, and the probe was a diffuse reflection optical fiber probe; the scanning range was 10000-4000cm-1 , the scanning times are 2 to 3 times, and the resolution is 8 cm -1 , the gain value is 8×.

[0081] After each collection, samples were taken and the moisture content (reference value) of the samples was tested with a moisture meter.

[0082] The predicted value and the reference value were subjected to a paired sample t test, with the significance level set at 0.05. The test result |t|=0.195, which is less than the t two-tail critical value of 1.990, and F=1.10<1.46, indicating that there is no statistical difference between the predicted value and the reference value. Figure 3 It can be seen that the two curves have good coincidence, which further shows that the online quantitative model of Example 1 has small error and good stability.

[0083] Example 7 Detection of moisture content of hydroxychloroquine sulfate granules

[0084] The online quantitative model in Example 1 is applied to real-time monitoring of the moisture content of hydroxychloroquine sulfate particles in the workshop production process:

[0085] The hydroxychloroquine sulfate particles were dried and fed into a fluidized bed in two batches. Near-infrared spectra were collected from the hydroxychloroquine sulfate particles during the drying process. When collecting the near-infrared spectra, the near-infrared spectrometer was a ThermoAntaris MX near-infrared spectrometer, and the probe was a diffuse reflection optical fiber probe. The scanning range was 10,000 to 4,000 cm -1 , the scanning times are 2 to 3 times, and the resolution is 8 cm -1 , the gain value is 8×.

[0086] The collected near infrared spectrum (prediction set sample) was imported into the online quantitative model in Example 1 to obtain the moisture content (prediction value) of the hydroxychloroquine sulfate particles; wherein, the RMSEP of the prediction set was 0.265.

[0087] Sampling was performed after each collection was completed. The number of samples in each batch was 10. After sampling, the moisture content (reference value) was determined using a moisture meter.

[0088] A paired sample t-test was performed on the predicted values ​​of the prediction set samples and the reference values, with the significance level set at 0.05. The test result |t|=0.265, which is less than the t two-tailed critical value of 2.110, and F=0.78<3.18, indicating that there is no statistical difference between the predicted values ​​and the true values.

[0089] Importing the model into the Result Operation software can reflect the changes in the moisture content of hydroxychloroquine sulfate granules during the production process of the above workshop in real time. Figure 4It can be seen that the moisture content steadily decreases until it stabilizes. During the drying process, 5 samples were taken online and the reference values ​​were measured with a moisture meter and marked on Figure 4 In the figure (the five values ​​marked by the triangle symbol), it can be seen that the predicted values ​​have good coincidence with the reference values, indicating that the model has a small error and good stability.

Claims

1. A method for establishing an online quantitative model of near-infrared spectroscopy of hydroxychloroquine sulfate granules water content, characterized in that: It includes the following steps: (1) collecting near infrared spectrum samples of the hydroxychloroquine sulfate particles in the drying process and taking samples after each collection, and detecting the moisture content of the samples obtained by sampling; wherein, when the drying temperature is below 35° C., the sampling time interval is 30 to 32 seconds, and when the drying temperature is greater than 35° C., the sampling time interval is 115 to 125 seconds; (2) removing abnormal samples from the collected near-infrared spectrum samples, and dividing the remaining near-infrared spectrum samples after removing the abnormal samples into a calibration set and a validation set; (3) Select the modeling band as 4800-5500cm -1 and 6800~7400cm -1 , the number of principal factors was selected as 5 to 7, and partial least squares regression was used to establish an online quantitative model of moisture near infrared spectroscopy of hydroxychloroquine sulfate granules; In step (1), the near-infrared spectrum samples are collected from the hydroxychloroquine sulfate particles in the drying process, and the sampling equipment is set to: a spectrum collection window and a sampling port are arranged on both sides of the lower part of the fluidized bed boiler body, and a probe of a near-infrared spectrometer is fixed on the spectrum collection window, and the spectrum collection window is close to the sampling port; when the fluidized bed is dried by air inlet, the near-infrared spectrum samples are collected by the near-infrared spectrometer probe fixed on the spectrum collection window, and samples are immediately taken from the sampling port after each collection is completed; In step (1), the scanning range of the near infrared spectrum sample collection is 10000~4000cm -1 , the number of scans is 64 times, the resolution is 8cm -1 , the gain value is 8×; In step (1), the terminal temperature of the drying is 50°C.

2. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 1, is characterized in that: In step (1), when the drying temperature is below 35°C, the sampling time interval is 30s, and when the drying temperature is greater than 35°C, the sampling time interval is 120s; And / or, the batches of the hydroxychloroquine sulfate granules are more than 10 batches.

3. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 1, is characterized in that: In step (1), the probe of the near-infrared spectrometer used for collecting near-infrared spectrum samples is a diffuse reflection optical fiber probe.

4. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 1, is characterized in that: In step (2), the method for eliminating abnormal samples in the collected near-infrared spectrum samples is a method of combining the Chauvinist criterion test with a leverage value-studentized residual plot; And / or, the classification method of the calibration set and the validation set is the SPXY classification method or the y equal spacing method.

5. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 1 or 4, is characterized in that: Step (2) includes preprocessing the near-infrared spectrum samples remaining after removing the abnormal samples.

6. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 5, is characterized in that: The preprocessing method is to separate the scattered signal and chemical information in the near infrared spectrum; Alternatively, the preprocessing method is to separate the scattering signal and chemical information in the near-infrared spectrum, enhance the spectral signal and filter out noise.

7. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 6, is characterized in that: The method for separating the scattering signal and chemical information in the near infrared spectrum is multivariate scattering correction or standard canonical transformation.

8. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 6, is characterized in that: The method for enhancing the spectral signal is first-order derivative or second-order derivative.

9. the method for establishing the online quantitative model of hydroxychloroquine sulfate granule moisture near infrared spectrum as claimed in claim 6, is characterized in that: The method for filtering out noise is Savitzky-Golay smoothing filtering or Karl Norris smoothing filtering.

10. The method for establishing a near-infrared spectrum online quantitative model of hydroxychloroquine sulfate granules moisture as claimed in claim 5, is characterized in that: The preprocessing method includes multivariate scatter correction, first-order derivative and Karl Norris smoothing filter; Alternatively, the preprocessing method is standard canonical transformation, first-order derivative and Karl Norris smoothing filtering.

11. The method for establishing a near-infrared spectrum online quantitative model of hydroxychloroquine sulfate granules moisture as claimed in claim 1, is characterized in that: In step (3), the modeling band is 4935-5336cm -1 and 6911~7297cm -1 ; And / or, the number of principal factors is 6.

12. The method for establishing a near-infrared spectrum online quantitative model of hydroxychloroquine sulfate granules moisture as claimed in claim 1 or 2, characterized in that: The method for eliminating abnormal samples in the collected near-infrared spectrum samples is a combination of the Chauvinist criterion test and the leverage value-studentized residual map; The classification method of the calibration set and the validation set is the SPXY classification method; Step (2) includes preprocessing the near-infrared spectrum samples remaining after the abnormal samples are removed; the preprocessing method is multivariate scattering correction, first-order derivative and Karl Norris smoothing filtering.

13. An online quantitative model of near-infrared spectroscopy of moisture in hydroxychloroquine sulfate particles established by the method as described in any one of claims 1 to 12.

14. A method for detecting moisture in hydroxychloroquine sulfate particles, characterized in that: It includes the following steps: (1) collecting near infrared spectra of hydroxychloroquine sulfate particles during the drying process; (2) importing the near infrared spectrum collected in step (1) into the online quantitative model of near infrared spectrum of moisture of hydroxychloroquine sulfate particles as claimed in claim 13 to obtain the moisture content of the hydroxychloroquine sulfate particles.