Detection system and detection method for medical intermediates

Abstract

The invention discloses a medical intermediate detection system and method, and the system comprises a production monitoring module which obtains the production process video data of a target product, and controls the operation of a product production line; the data storage module is used for storing target product production process data, production process video data and finished product video data; and the analysis processing module analyzes the video data of the production process of the target product. By acquiring the video data of the production process, the operation condition of the production line can be monitored in real time, and measures can be taken quickly when abnormity is found. When abnormity is found, automatic early warning can be carried out, a production line is promoted to stop running, and the product quality problem caused by misoperation or equipment faults is avoided; the related data of the whole life cycle from raw material input to finished product output of the target product is stored in a centralized manner, so that the production process of each batch of products is traceable, the traceability and transparency of production are enhanced, and the product quality is ensured.

Description

Technical Field

[0001] The invention belongs to the field of pharmaceutical intermediate production detection systems, and in particular relates to a pharmaceutical intermediate detection system and a detection method. Background Art

[0002] Pharmaceutical intermediates are an important part of the drug research and development process. They can be used to manufacture precursors, intermediates and end products of drugs. Their quality and effect have a crucial impact on the efficacy and safety of drugs. If the impurity content of the intermediates is too high or the reaction is incomplete, it may cause problems in the drug production process, thereby affecting the efficacy and safety of the drugs. Therefore, in the process of drug research and development, it is necessary to strictly control and manage the quality and effect of pharmaceutical intermediates to ensure the efficacy and safety of drugs. At the same time, it is also necessary to continuously optimize the research and development process and technology, improve the quality of intermediates, and provide more reliable support for drug research and development.

[0003] Currently, the production monitoring and management of pharmaceutical intermediates is still mainly done manually, and manual monitoring cannot ensure that the quality of the target product produced throughout the entire process meets the requirements, which will affect the effect of the final product, leading to waste of resources and increased production costs.

[0004] Therefore, we need to design a detection system and method for pharmaceutical intermediates to solve these problems. Summary of the invention

[0005] The problem to be solved by the present invention is to provide a pharmaceutical intermediate production detection system, which realizes comprehensive intelligent management of the production process by integrating functions such as data acquisition, real-time monitoring, intelligent early warning, quality supervision and visual management, improves production efficiency, ensures product quality, reduces resource waste, and ensures production safety.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A method for detecting a pharmaceutical intermediate comprises the following steps:

[0008] S1. Obtain near infrared spectrum data of the target product to be detected;

[0009] S2, collecting multiple samples of the target product to be tested and numbering them, and placing the samples of the target product to be tested in the photoacoustic cell respectively for testing in sequence;

[0010] S3, the processed light is projected onto a chopper, and light with adjustable intensity and period is obtained through the chopper. After the light is redirected by a reflector and focused by a lens, it is vertically projected into the acousto-optic pool; the signal received by the microphone in the acousto-optic pool is sequentially input into a computer after passing through a preamplifier and a lock-in amplifier, and the computer stores and processes the received signal to obtain spectral data of multiple target product samples;

[0011] S4. Select appropriate near-infrared spectrum modeling band and near-infrared raw spectrum preprocessing method: the near-infrared spectrum raw data is preprocessed by standard normalized transformation (SNV) and second-order derivative method (Savitzky-Golay smoothing);

[0012] S5. Establish near-infrared quantitative calibration model for key quality control indicators in target products: Use chemometric software to correlate the obtained near-infrared spectral information with the standard values ​​measured by the reference method, use the principal component-based back-propagation artificial neural network method (PCA-BPANN) to establish a quantitative calibration model between near-infrared spectroscopy and key quality control indicators, cross-validate the calibration model, and evaluate the model performance with multiple parameters including correlation coefficient (R), mean square error of calibration set (RMSEC), mean square error of prediction set (RMSEP), relative deviation (RSEP), and relative analytical error (RPD);

[0013] S6. Determine the near infrared spectrum data of the sample of the target product to be detected, import it into the quantitative calibration model in S5, and obtain the detection result.

[0014] Preferably, in S5, the following steps are performed: principal component analysis (PCA) is performed on the near-infrared spectrum of the preprocessed and selected band to extract the main characteristic variables of the target product sample spectrum, which are used as the number of input layer units of the reverse artificial neural network BP-ANN to construct a two-layer artificial neural network model; after testing, the transfer function from the input layer to the hidden layer is a tangent S-shaped transfer function (Tansig), and the transfer function from the hidden layer to the output layer is a linear transfer function (Purelin), and the Levenberg-Marquardt rule is used to train the network, the learning rate is 0.1, the number of repeated training is 1000, and the training termination target is 4×10-7; finally, a reverse artificial neural network (PCA-BPANN) quantitative correction model based on principal component analysis of each key quality control indicator in the target product sample is obtained through training with the calibration set sample data, and the model is comprehensively examined through comprehensive evaluation indicators, and the stability and predictive ability of the trained PCA-BPANN quantitative model are tested using the validation set samples.

[0015] A detection system for pharmaceutical intermediates, including a production monitoring module, a data storage module and an analysis and processing module:

[0016] The production monitoring module obtains the production process video data of the target product and controls the operation of the product production line;

[0017] A data storage module, used to store target product production process data, production process video data and finished product video data;

[0018] The analysis and processing module analyzes the target product production process video data, obtains the production process steps of the target product in the production process video data, and compares them with the target product production process data. When the production process steps of the production process video data do not conform to the production process steps in the target product production process data, a warning signal is sent to the production monitoring module, and the production equipment is controlled to stop running through the production monitoring module; when the production process steps of the production process video data completely conform to the production process steps in the target product production process data, the production process video data is bound to the corresponding target product to obtain the finished product video data;

[0019] The quality detection module is used to detect the target product quality corresponding to the finished product video data.

[0020] The quality detection module shown uses Fourier transform infrared spectroscopy to detect the quality of the target product;

[0021] The quality inspection module performs the steps of the above-mentioned method for inspecting a pharmaceutical intermediate.

[0022] The advantages and positive effects of the present invention are:

[0023] By acquiring video data of the production process, it is not only possible to monitor the operation of the production line in real time, but also to take prompt measures when abnormalities are found. When abnormalities are found, it can automatically warn and stop the production line, avoiding product quality problems caused by operational errors or equipment failures; the relevant data of the entire life cycle of the target product from raw material input to finished product output is centrally stored, so that the production process of each batch of products can be traced, enhancing the traceability and transparency of production, ensuring product quality, and providing data support for further optimization of production; through intelligent analysis of the video data of the production process, it can automatically identify the operation links that do not meet the process requirements, issue warnings in time, reduce the omission rate of manual inspection, improve the accuracy of detection, and greatly reduce labor costs; Fourier transform infrared spectroscopy technology is used to detect the quality of finished products, combined with near-infrared spectroscopy modeling and principal component analysis (PCA) + reverse artificial neural network (BP-ANN) methods, to achieve rapid and accurate measurement of key quality control indicators of products. Compared with traditional chemical analysis methods, it has the advantages of non-destructive, high efficiency, and low cost. It can reduce the production cost of enterprises, improve the profitability of enterprises, and also improve the qualified rate and consistency of products, enhancing the competitive advantage of enterprises in the market. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0025] Figure 1 It is a system block diagram of a pharmaceutical intermediate production and detection system of the present invention;

[0026] Figure 2 The present invention discloses a quality inspection flow chart of a pharmaceutical intermediate production inspection system. DETAILED DESCRIPTION

[0027] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and the like are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0028] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0029] The present invention will be further described below in conjunction with the accompanying drawings:

[0030] like Figure 1 and Figure 2 As shown, an intelligent detection system and detection method for pharmaceutical intermediates include a detection system including a production monitoring module, a data storage module and an analysis and processing module:

[0031] The production monitoring module obtains the production process video data of the target product and controls the operation of the product production line;

[0032] A data storage module, used to store target product production process data, production process video data and finished product video data;

[0033] The analysis and processing module analyzes the target product production process video data, obtains the production process steps of the target product in the production process video data, and compares them with the target product production process data. When the production process steps of the production process video data do not conform to the production process steps in the target product production process data, a warning signal is sent to the production monitoring module, and the production equipment is controlled to stop running through the production monitoring module; when the production process steps of the production process video data completely conform to the production process steps in the target product production process data, the production process video data is bound to the corresponding target product to obtain the finished product video data;

[0034] The quality detection module is used to detect the target product quality corresponding to the finished product video data.

[0035] The quality detection module shown uses Fourier transform infrared spectroscopy to detect the quality of the target product;

[0036] The detection method includes the following steps:

[0037] S1. Obtain near infrared spectrum data of the target product to be detected;

[0038] S2, collecting multiple samples of the target product to be tested and numbering them, and placing the samples of the target product to be tested in the photoacoustic cell respectively for testing in sequence;

[0039] S3, the processed light is projected onto a chopper, and light with adjustable intensity and period is obtained through the chopper. After the light is redirected by a reflector and focused by a lens, it is vertically projected into the acousto-optic pool; the signal received by the microphone in the acousto-optic pool is sequentially input into a computer after passing through a preamplifier and a lock-in amplifier, and the computer stores and processes the received signal to obtain spectral data of multiple target product samples;

[0040] S4. Select appropriate near-infrared spectrum modeling band and near-infrared raw spectrum preprocessing method: The near-infrared spectrum raw data is preprocessed using standard normalized transformation (SNV) and second-order derivative method (Savitzky-Golay smoothing); they are used to eliminate the influence of baseline drift, noise and solid particles on the spectrum.

[0041] S5. Establish near-infrared quantitative calibration model for key quality control indicators in target products: Use chemometric software to correlate the obtained near-infrared spectral information with the standard values ​​measured by the reference method, use the principal component-based back-propagation artificial neural network method (PCA-BPANN) to establish a quantitative calibration model between near-infrared spectroscopy and key quality control indicators, cross-validate the calibration model, and evaluate the model performance with multiple parameters including correlation coefficient (R), mean square error of calibration set (RMSEC), mean square error of prediction set (RMSEP), relative deviation (RSEP), and relative analytical error (RPD);

[0042] About two-thirds of the samples were randomly selected as the calibration set, and the remaining one-third of the samples were used as the validation set for prediction. A BP-ANN network model with 2 hidden layers was established. According to the precision of the model and the accuracy of the prediction, the optimal number of hidden layer units was found. The transfer function from the input layer to the hidden layer was the tangent S-shaped transfer function (Tansig), and the transfer function from the hidden layer to the output layer was the linear transfer function (Purelin). At the same time, the Levenberg-Marquardt rule was used to train the network, with a learning rate of 0.1, 1000 repeated training times, and a training termination target of 4×10-7. Finally, the PCA-BP-ANN quantitative model of the target sample and quality control indicators was obtained after training with the calibration set sample data;

[0043] The established quantitative calibration model uses parameters such as correlation coefficient (R), mean square error of calibration set (RMSEC), and mean square error of prediction set (RMSEP) to examine the model performance. When the model correlation coefficient R is closer to 1, the analysis accuracy is higher. When the RMSEC and RMSEP values ​​are smaller and closer to each other, the model has a higher prediction ability. At the same time, the prediction relative error (RSEP) is used to evaluate the model's prediction ability for unknown samples. When the RSEP value is less than 10% and the smaller it is, the evaluation model has a good prediction ability and can meet the requirements of rapid online detection of the active ingredient index components of the target product.

[0044] S6. Determine the near infrared spectrum data of the sample of the target product to be tested, import it into the quantitative calibration model in S5, and obtain the test result.

[0045] In S5, the following steps are performed: the near-infrared spectra of the preprocessed and selected bands are subjected to principal component analysis (PCA) to extract the main characteristic variables of the target product sample spectrum, which are used as the input layer unit number of the reverse artificial neural network BP-ANN to construct a two-layer artificial neural network model; the transfer function from the input layer to the hidden layer is tested to be a tangent S-shaped transfer function (Tansig), and the transfer function from the hidden layer to the output layer is a linear transfer function (Purelin). At the same time, the Levenberg-Marquardt rule is used to train the network, with a learning rate of 0.1, 1000 repeated training times, and a training termination target of 4×10-7; finally, the principal component analysis-based reverse artificial neural network (PCA-BPANN) quantitative calibration model of each key quality control indicator in the target product sample is obtained through training with the calibration set sample data, and the model is comprehensively examined through comprehensive evaluation indicators, and the stability and predictive ability of the trained PCA-BPANN quantitative model are tested using the validation set samples.

[0046] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for detecting a pharmaceutical intermediate, characterized in that: The following steps are involved: S1. Obtain near infrared spectrum data of the target product to be detected; S2, collecting multiple samples of the target product to be tested and numbering them, and placing the samples of the target product to be tested in the photoacoustic cell respectively for testing in sequence; S3, the processed light is projected onto a chopper, and light with adjustable intensity and period is obtained through the chopper. After the light is redirected by a reflector and focused by a lens, it is vertically projected into the acousto-optic pool; the signal received by the microphone in the acousto-optic pool is sequentially input into a computer after passing through a preamplifier and a lock-in amplifier, and the computer stores and processes the received signal to obtain spectral data of multiple target product samples; S4. Select appropriate near-infrared spectrum modeling band and near-infrared raw spectrum preprocessing method: the near-infrared spectrum raw data is preprocessed using standard regularization transformation and second-order derivative method; S5. Establish near-infrared quantitative calibration model for each key quality control index in the target product: use chemometrics software to correlate the obtained near-infrared spectral information with the standard value measured by the reference method, use the principal component-based reverse artificial neural network method to establish a quantitative calibration model between the near-infrared spectrum and key quality control indicators, cross-validate the calibration model, and evaluate the model performance with multiple parameters including correlation coefficient, calibration set mean square error, prediction set mean square error, relative deviation, and relative analytical error; S6. Determine the near infrared spectrum data of the sample of the target product to be detected, import it into the quantitative calibration model in S5, and obtain the detection result.

2. A method for detecting a pharmaceutical intermediate according to claim 1, characterized in that: In S5, the following steps are performed: principal component analysis is performed on the near-infrared spectra of the preprocessed and selected bands to extract the main characteristic variables of the target product sample spectrum, which are used as the number of input layer units of the reverse artificial neural network BP-ANN to construct a two-layer artificial neural network model; the transfer function from the input layer to the hidden layer is tested to be a tangent S-shaped transfer function, and the transfer function from the hidden layer to the output layer is a linear transfer function. At the same time, the Levenberg-Marquardt rule is used to train the network, with a learning rate of 0.1, 1000 repeated training times, and a training termination target of 4×10-7; finally, the reverse artificial neural network quantitative correction model based on principal component analysis of each key quality control indicator in the target product sample is obtained through training with the calibration set sample data, and the model is comprehensively examined through comprehensive evaluation indicators, and the stability and predictive ability of the trained PCA-BPANN quantitative model are tested using the validation set samples.

3. A detection system for pharmaceutical intermediates, characterized in that: Including production monitoring module, data storage module and analysis and processing module: The production monitoring module obtains the production process video data of the target product and controls the operation of the product production line; A data storage module, used to store target product production process data, production process video data and finished product video data; The analysis and processing module analyzes the target product production process video data, obtains the production process steps of the target product in the production process video data, and compares them with the target product production process data. When the production process steps of the production process video data do not conform to the production process steps in the target product production process data, a warning signal is sent to the production monitoring module, and the production equipment is controlled to stop running through the production monitoring module; when the production process steps of the production process video data completely conform to the production process steps in the target product production process data, the production process video data is bound to the corresponding target product to obtain the finished product video data; A quality detection module is used to detect the quality of the target product corresponding to the finished product video data, and the quality detection module executes the steps of a method for detecting a pharmaceutical intermediate as described in claim 1 or 2.

Images