A medicine material on-line detection device

Abstract

The application discloses a kind of drug material online detection equipment, it is related to drug material detection technical field, solve the detection distortion problem caused by drug sample slight tilt.The equipment includes equipment shell, control console and detection bin, and detection bin is built-in installation state detection component and infrared detection component.Control console integrated circuit board integrates sensing acquisition, data processing and control module, and data processing module contains slight tilt identification, spectrum offset correction, component data correction unit, and the real component content is calculated after correcting spectral distortion by identifying inclination angle using pressure data.Control module determines sample qualified state in combination with preset standard, controls equipment operation and sends control instruction to external production system.The equipment improves detection accuracy and operation convenience, realizes detection and production linkage, adapts to online quality control demand, and provides reliable support for drug production quality control.

Description

Technical Field

[0001] This invention relates to the field of pharmaceutical material testing technology, and in particular to an online testing device for pharmaceutical materials. Background Technology

[0002] In online component analysis of pharmaceutical materials, infrared spectroscopy is widely used due to its speed and non-destructive nature. However, this technology has extremely high requirements for the stability of the sample mounting posture: when the pharmaceutical sample (such as tablets or films) is slightly tilted (not a serious tilt, which cannot be identified by conventional mounting and testing), the infrared light path will shift, causing the characteristic peak positions in the spectral data to shift and the intensity to be distorted, thus resulting in incorrect component analysis results.

[0003] While existing online infrared detection equipment can perform basic component detection, it suffers from key technical defects: pharmaceutical samples (such as tablets and films) are prone to slight tilts of 0.5° to 3°. This angle cannot be detected by conventional installation methods but causes infrared light path shift, resulting in spectral peak position shift and intensity distortion, which in turn leads to errors in component content calculation. At the same time, the tactile sensors in existing equipment are only used to determine whether the sample is installed correctly and are not correlated with infrared spectral data. They lack a targeted error correction mechanism and cannot solve the detection distortion problem caused by tilt, which seriously affects the accuracy of the detection results. Therefore, it is necessary to provide an online detection device for pharmaceutical materials to solve the above technical problems. Summary of the Invention

[0004] This invention provides an online detection device for pharmaceutical materials, which solves the technical problem of infrared spectrum shift and component detection data distortion caused by slight tilting of pharmaceutical samples.

[0005] To solve the above-mentioned technical problems, the present invention provides an online detection device for pharmaceutical materials, including a device housing, a control console on one side of the top surface of the device housing, and a detection chamber on the other side of the top surface of the device housing. The detection chamber contains an installation status detection component and an infrared detection component. The installation status detection component includes at least one set of tactile sensors for collecting contact pressure distribution data between the pharmaceutical sample and the sample support structure. The infrared detection component consists of an infrared light source and an infrared detector for collecting infrared spectral data of the pharmaceutical sample.

[0006] The control console is equipped with an integrated circuit board, which includes a sensor acquisition module, a data processing module, and a control module.

[0007] The sensing and acquisition module is used to collect contact pressure distribution data and infrared spectral data;

[0008] The data processing module includes a micro-tilt recognition unit, a spectral shift correction unit, and a component data correction unit;

[0009] The micro-tilt recognition unit is used to identify the micro-tilt angle of the drug sample;

[0010] The spectral shift correction unit is used to correct infrared spectral shifts based on a small tilt angle.

[0011] The component data correction unit is used to calculate the true component content of the drug sample based on the corrected spectral data.

[0012] The control module is used to control the operating status of the equipment and send control commands to the external production system based on the corrected component content and slight tilt angle information output by the data processing module, combined with the preset component qualification standards and tilt threshold.

[0013] Preferably, the infrared light source in the infrared detection assembly is disposed on one inner wall of the detection chamber, and the infrared detector is disposed on the other inner wall of the detection chamber; a base is provided on the bottom surface of the detection chamber between the infrared light source and the infrared detector, and an I-beam is fixedly connected to the top surface of the base. A first through hole coaxial with the infrared detector is opened on the I-beam, and two opposing limiting plates are fixedly connected to the four sides of the I-beam.

[0014] Preferably, the sample carrying structure includes a plate inserted into the I-beam, the plate having a second through hole coaxial with the first through hole and having the same diameter, and an annular groove for placing the drug sample is formed on one end face of the plate at the second through hole, and a limiting ring for pressing the drug sample is magnetically attached to the annular groove.

[0015] Preferably, a rectangular block is fixed to the bottom surface of the I-beam, and multiple arc-shaped grooves are provided at equal intervals on the bottom surface of the rectangular block. Two second rectangular grooves are symmetrically opened on the top surface of the base, located on both sides of the I-beam and adapted to the rectangular block. The tactile sensor is disposed on the bottom surface of the second rectangular groove and adapted to the arc-shaped groove.

[0016] Preferably, the micro-tilt recognition unit includes a pressure feature extraction subunit and an angle calculation subunit;

[0017] The pressure feature extraction subunit is used to extract the pressure values ​​and pressure differences in different regions of the sample bearing structure after preprocessing and feature quantization of the contact pressure distribution data.

[0018] The angle calculation subunit converts the extracted pressure difference value into a tiny tilt angle of the sample based on a preset pressure difference and tilt angle mapping model.

[0019] Preferably, the spectral shift correction unit includes an offset determination subunit and a spectral correction subunit;

[0020] The offset determination subunit determines the positional offset and intensity distortion of the characteristic peaks in the infrared spectrum based on the tilt angle output by the micro tilt recognition unit and by calling the preset tilt angle-spectral offset correspondence.

[0021] The spectral correction subunit adjusts the position of the characteristic peaks according to the positional offset and compensates for the spectral intensity according to the intensity distortion to generate corrected spectral data.

[0022] Preferably, the tilt angle-spectral shift correspondence is established in the following manner:

[0023] Apply different minute tilt angles to the standard sample, collect infrared spectral data at the corresponding angles, compare the spectral data at the standard angle, record the characteristic peak position offset and intensity distortion, and form a mapping relationship database.

[0024] Preferably, the component data correction unit includes a model calling subunit and a content calculation subunit;

[0025] The model calls a subunit to load a preset corrected spectral-component content correlation model;

[0026] The content calculation subunit inputs the corrected spectral data output by the spectral shift correction unit into the correlation model to calculate and output the true component content of the drug sample.

[0027] Preferably, the corrected spectrum-component content correlation model is a mathematical model obtained by training corrected spectral data of multiple samples with known component contents at different tilt angles through partial least squares regression, support vector machine or neural network algorithm.

[0028] Compared with related technologies, the online detection device for pharmaceutical materials provided by this invention has the following beneficial effects:

[0029] 1. This solution collects contact pressure distribution data through a tactile sensor, accurately identifies the three-dimensional tilt angle through a micro-tilt recognition unit, corrects spectral distortion through a spectral offset correction unit, and finally calculates the true component content through an association model. This constructs a complete error correction closed loop consisting of tilt recognition, spectral correction, and component calculation, completely solving the detection distortion problem caused by tilt and significantly improving the accuracy of drug component detection.

[0030] 2. This solution designs a sample-bearing structure consisting of a single plate, a magnetic limiting ring, and a spring clip. Mechanical limiting and magnetic fixation ensure precise sample alignment, while a tactile sensor detects the installation position, automatically verifying and providing feedback on the installation status to avoid human error. Simultaneously, the pull-out design and hook-shaped locking function simplify the sample replacement process, eliminating the need for continuous force from staff, thus ensuring sample installation stability while improving operational efficiency.

[0031] 3. The control module of this solution can automatically determine the sample's qualification status based on the corrected component content and tilt angle, control the equipment's start / stop, alarm, and other operating statuses, and simultaneously send instructions to the external production system to continue production, pause material feeding, or adjust the process. This linkage mechanism allows the test results to directly guide production control, avoid the generation of batches of unqualified products, solve the problem of disconnect between testing and production, and help achieve efficient and accurate quality control of the production process.

[0032] The solution addresses the core shortcomings of existing equipment—detection accuracy, ease of operation, and production integration—by focusing on three key dimensions: accuracy, ease of operation, and production integration. It not only resolves the core defect of detection distortion caused by slight tilting in existing equipment but also improves installation stability and operational efficiency through structural optimization. Furthermore, it enables coordinated control between the detection and production systems, ultimately achieving precise, efficient, and intelligent online detection of pharmaceutical materials. This provides comprehensive and reliable technical support for quality control in the pharmaceutical production process and effectively compensates for several technical shortcomings of existing equipment. Attached Figure Description

[0033] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0034] Figure 1 This is a first-view schematic diagram of the overall structure proposed in this invention;

[0035] Figure 2 This is a schematic diagram of the overall structure for removing the cover plate proposed in this invention;

[0036] Figure 3 This is a schematic diagram of the overall structure of the detection chamber proposed in this invention;

[0037] Figure 4 This is a schematic cross-sectional view of the spring box structure proposed in this invention;

[0038] Figure 5 This is a block diagram showing the overall module connection relationship proposed in this invention;

[0039] Figure 6 This is a block diagram illustrating the entire process of detection, calibration, and control proposed in this invention.

[0040] The numbers in the diagram are: 1. Equipment casing; 2. Control console; 3. Limiting ring; 4. Cover plate; 5. Infrared light source; 6. Infrared detector; 7. Base; 8. I-beam plate; 9. Straight plate; 10. Spring box; 11. Tactile sensor; 12. Hook-shaped locking block. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “group,” “class,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0043] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0044] Please refer to the following: Figures 1-6 An online testing device for pharmaceutical materials includes a housing 1, a control console 2 on one side of the top surface of the housing 1, and a testing chamber on the other side of the top surface of the housing 1. The testing chamber contains an installation status detection component and an infrared detection component. The installation status detection component includes at least one set of tactile sensors 11 for collecting contact pressure distribution data between the pharmaceutical sample and the sample-supporting structure. The infrared detection component consists of an infrared light source 5 and an infrared detector 6 for collecting infrared spectral data of the pharmaceutical sample.

[0045] The control console 2 is equipped with an integrated circuit board, which includes a sensor acquisition module, a data processing module, and a control module.

[0046] The sensor acquisition module is used to collect contact pressure distribution data and infrared spectral data;

[0047] The data processing module includes a micro-tilt recognition unit, a spectral shift correction unit, and a component data correction unit;

[0048] The micro-tilt recognition unit is used to identify the micro-tilt angle of the drug sample;

[0049] The spectral shift correction unit is used to correct infrared spectral shifts based on minute tilt angles;

[0050] The component data correction unit is used to calculate the true component content of a drug sample based on the corrected spectral data.

[0051] The control module is used to determine the corrected component content C and the three-dimensional micro-tilt angle based on the output of the data processing module. Combined with the preset ingredient qualification standards With tilt threshold It controls the operating status of equipment and sends control commands to external production systems.

[0052] When the component content is corrected And a small tilt angle If the sample is found to be in good condition and the equipment is operating normally, a normal maintenance signal is generated. The normal maintenance signal is used to trigger the control cover 4 drive assembly to remain closed, the infrared light source 5 switch assembly to remain open for continuous detection, and the sample carrying structure moving assembly to maintain its current position to ensure the continuity of subsequent sample detection. At the same time, a normal status signal is sent to the control console 2, and the component is displayed as qualified and the tilt is normal on the display screen. The indicator light configured on the control console 2 is green.

[0053] If a small tilt angle An emergency adjustment signal is generated. The emergency adjustment signal is used to trigger the control of the infrared light source 5 switch component to shut down, so as to avoid continuous acquisition of distorted spectrum; control the cover plate 4 drive component to open, so as to facilitate the staff to adjust the sample installation posture; control the sample support structure moving component to reset to the initial position, so as to provide a reference for reinstalling the sample; at the same time, a tilt warning signal is sent to the control console 2, the display screen pops up a sample tilt exceeding the standard prompt, the indicator light flashes yellow, and the tilt angle value is recorded simultaneously for the staff to refer to.

[0054] If the component content or Then a shutdown detection signal is generated; the shutdown detection signal is used to trigger the control of the infrared light source 5 to turn off and the cover plate 4 to open, and the sample carrying structure moving component to move the unqualified sample out of the detection station;

[0055] If a small tilt angle and The system generates two abnormal signals. The two abnormal signals take priority in executing the emergency control action for tilt exceeding the threshold, and then the shutdown action for component abnormality is superimposed. The console 2 displays the two types of abnormal information simultaneously, and prompts the processing priority in the order of tilt exceeding the standard → component unqualified.

[0056] Coordination and control with external production systems, including:

[0057] When a normal maintenance signal is generated, a test pass → continuous production instruction is sent to an external production system (such as a manufacturing execution system, MES). This instruction includes the average composition content and tilt angle statistics of the current batch of samples, which are recorded and traced by the production system to maintain uninterrupted production.

[0058] When an emergency adjustment signal is generated, a pause feeding → pending adjustment instruction is sent to the external production system, prompting the production system to temporarily stop feeding new samples to the testing equipment to avoid invalid testing of batch samples due to installation problems.

[0059] When a shutdown detection signal is generated, a "process adjustment → troubleshooting" instruction is sent to the external production system, along with the component data and detection timestamp of the non-conforming sample. This triggers the production system to start a self-check of process parameters (such as raw material ratio, tableting pressure, etc.). Production is resumed only after the parameters are adjusted to be qualified.

[0060] When a dual abnormal signal is generated, an emergency shutdown → comprehensive investigation instruction is sent to the external production system, requiring the production system to stop the operation of the relevant production line and simultaneously push the abnormal data to the quality control module to assist staff in locating installation errors or process defects.

[0061] The control module receives real-time feedback signals from the execution components (such as whether the cover 4 is fully open or whether the infrared light source 5 is properly closed). If the action is not executed properly, a fault prompt will be issued through the control console 2. At the same time, all control commands, equipment status, and external system linkage results are stored in the storage unit configured in the data processing module, supporting historical data query and production traceability to ensure the traceability of the detection and production process.

[0062] Specifically, the infrared light source 5 in the infrared detection assembly is set on one inner wall of the detection chamber, and the infrared detector 6 is set on the other inner wall of the detection chamber; a base 7 is provided on the bottom surface of the detection chamber between the infrared light source 5 and the infrared detector 6, and an I-beam plate 8 is fixedly connected to the top surface of the base 7. The I-beam plate 8 has a first through hole coaxial with the infrared detector 6, and two opposing limiting plates are fixedly connected to all four sides of the I-beam plate 8.

[0063] It should be noted that the outer shell 1 is made of cold-rolled steel plate with an epoxy resin coating, which is rust-proof, corrosion-resistant, and has high structural strength, protecting its internal components. The control console 2 is made of ABS engineering plastic and is equipped with an LCD screen. The infrared light source 5 is a silicon carbide rod, which has high heating efficiency and high infrared radiation intensity, suitable for the deep component detection needs of pharmaceutical materials. The infrared detector 6 is a photon detector, which has fast response speed and high detection sensitivity, and can accurately capture the light signal after the infrared light source 5 penetrates the sample. The base 7 is made of aluminum alloy, which is lightweight and has good heat dissipation. The I-beam plate 8 is made of acrylic plate, which has high transparency and is easy to observe the internal installation status. The cover plate 4 is made of tempered glass, which is pressure-resistant, wear-resistant, and does not affect the light path. This invention integrates the control console 2, detection chamber, and internal components through the outer shell 1. The infrared light source 5 and the infrared detector 6 form a detection light path, and the I-beam plate 8 and the limiting plate provide a sample installation benchmark, realizing the basic integration of detection function and installation positioning, and ensuring the orderly cooperation of all components of the device. The above-mentioned components together constitute the basic framework of the detection equipment, providing structural support for subsequent accurate sample installation and stable detection.

[0064] Specifically, the sample carrying structure includes a straight plate 9 inserted on the I-beam plate 8. The straight plate 9 has a second through hole that is coaxial with the first through hole and has the same diameter. The outward end face of the straight plate 9 has an annular groove for placing the drug sample at the second through hole. A limiting ring 3 for pressing the drug sample is magnetically attached at the annular groove.

[0065] It should be noted that the outer wall of the I-plate 9 has a first rectangular groove at the upper end of the limiting ring 3; the upper ends of both sides of the I-plate 9 have symmetrical slots; the inner end face of the two side wings of the I-plate 8 is provided with a corresponding slot; the inner wall of the I-plate 8 has a limiting groove for the sliding of the slot; and one end of the slot is fixedly connected to a limiting block that is slidably set in the limiting groove.

[0066] It should be further explained that the I-plate 9 is made of low-carbon steel with a galvanized surface, which not only has magnetic properties but also prevents rust. It is suitable for the magnetic attraction of the limiting ring 3 and has high structural strength and is not easily deformed. The main body of the limiting ring 3 is made of ABS engineering plastic, and a permanent magnetic material block is fixed only at one end facing the I-plate 9. It achieves adsorption and fixation with the I-plate 9 through local magnetic attraction, ensuring magnetic stability. The first rectangular groove facilitates finger insertion and pulling. This structure, through the limiting groove of the I-plate 8 and the I-plate 9, achieves precise guidance for sample installation. The limiting ring 3 is tightly adsorbed with the magnetically attracted I-plate 9 through local magnetic attraction, which not only fixes the reagent tablets but also makes it easy to observe the installation status with the naked eye. The first rectangular groove facilitates the subsequent pulling of the I-plate 9, improving the stability, visibility and ease of operation of sample installation.

[0067] The inner wall of the slot is frosted to increase friction with the card block and improve engagement stability. The card block is made of 45# steel and is heat-treated for high hardness and wear resistance. The limiting block is made of nylon, which has low sliding resistance and is not easy to wear on the inner wall of the I-beam plate 8. This structure provides a detachable fixing structure for the I-beam plate 9 through the cooperation of the card block and the slot. The limiting block and the limiting slot restrict the sliding direction of the card block, ensuring that the card block is accurately engaged in the slot, preventing the I-beam plate 9 from shifting during the testing process, and improving the stability of sample installation.

[0068] The two sides of the I-beam plate 8 have outward-facing end faces connected to the limiting groove and fixedly connected to the spring box 10. The spring box 10 contains a spring, one end of which is fixedly connected to the limiting block, and the other end of which is fixedly connected to the outward-facing end of the inner wall of the spring box 10. A pull rod is sleeved inside the spring, and one end of the pull rod extends out of the outward-facing outer wall of the spring box 10 and is fixedly connected to a pull block.

[0069] A U-shaped limiting block in the same direction as the spring is fixed to the bottom surface of the spring box 10, and a hook-shaped locking block 12 adapted to the U-shaped limiting block is fixed to the bottom surface of the pull block. The hook head of the hook-shaped locking block 12 abuts against the inward end of the U-shaped limiting block.

[0070] It should be noted that the spring box 10 is made of aluminum alloy, which has high structural strength and is easy to process; the spring is a cylindrical helical compression spring made of 60Si2Mn spring steel, which has a high elastic limit and long fatigue life, and can provide continuous clamping force for the locking block; the pull rod is made of 304 stainless steel, which is rust-proof, corrosion-resistant and not easily deformed; the pull block is made of PVC material with anti-slip texture on the surface, making it easy for operators to pull; this structure uses the elastic force of the spring to push the limit block and the locking block, so that the locking block automatically locks into the slot without manual operation. The pull rod and pull block facilitate the subsequent unlocking of the locking block, which takes into account both the automation of sample fixing and the convenience of unlocking, and improves the operating efficiency of the equipment;

[0071] The U-shaped limiting block is made of stainless steel and is fixed to the spring box 10 by welding, resulting in a stable structure. The hook-shaped locking block 12 is made of elastic plastic and has a certain deformation capacity, making it easy to hook and release the U-shaped limiting block. Through the cooperation of the hook-shaped locking block 12 and the U-shaped limiting block, the structure can achieve state locking after pulling the pull block to unlock the locking block, without requiring continuous force from the operator, freeing up their hands for subsequent pulling of the straight plate 9, further improving the convenience and user-friendliness of the equipment operation.

[0072] Specifically, a rectangular block is fixed to the bottom surface of the I-beam plate 9, and multiple arc-shaped grooves are provided at equal intervals on the bottom surface of the rectangular block. The top surface of the base 7 has two second rectangular grooves located on both sides of the I-beam plate 8 and adapted to the rectangular block. The tactile sensor 11 is set in the bottom surface of the second rectangular groove and adapted to the arc-shaped groove.

[0073] It should be further noted that the rectangular block is made of ABS plastic; the inner wall of the arc-shaped groove is smooth to avoid scratching the tactile sensor 11; the tactile sensor 11 is a capacitive touch sensor with high detection sensitivity and small size, suitable for arc-shaped groove contact scenarios; the inner wall of the second rectangular groove is provided with a positioning boss to ensure that the rectangular block is accurately embedded; this structure provides a secondary positioning reference for the straight plate 9 through the cooperation of the rectangular block and the second rectangular groove. When the arc-shaped groove contacts the tactile sensor 11, it triggers an installation signal, realizing automatic verification of the sample installation status, avoiding detection errors caused by human judgment errors, and improving the effectiveness of detection data and the safety of equipment operation.

[0074] Specifically, the micro-tilt recognition unit includes a pressure feature extraction subunit and an angle calculation subunit;

[0075] The pressure feature extraction subunit is used to preprocess and quantize the contact pressure distribution data collected by the tactile sensor 11, and then extract the pressure values ​​and pressure differences in different regions of the sample's bearing structure. Specifically, the process includes the following steps:

[0076] The contact pressure distribution data between the bottom surface of the plate 9 and the top surface of the base 7 is collected in real time by the tactile sensor 11 and recorded as follows: Where i and j represent the two-dimensional coordinates of the tactile sensor 11 in the second rectangular slot, determined based on the installation position of the tactile sensor, and t represents the data sampling time. Specifically, it represents the contact pressure value detected by the tactile sensor 11 at the coordinate (i,j) at sampling time t;

[0077] Because the testing environment contains noise such as equipment vibration and electromagnetic interference, it is necessary to obtain contact pressure distribution data. Noise reduction was performed, specifically using a moving average filtering algorithm, to obtain the noise-reduced contact pressure distribution data. The formula is: Where N represents the length of the filtering window and is a positive integer, and n represents the index of the historical sampling time used to calculate the moving average, with a value ranging from 0 to N-1; the filtering process removes stress data fluctuations caused by random noise, ensuring the accuracy of subsequent feature extraction;

[0078] Using the axis of the I-beam (8) as the boundary, the acquisition area of ​​the tactile sensor 11 is divided into two regions (left region L and right region R); using the center line of the I-beam 8 perpendicular to the axis as the boundary, it is divided into two regions (front region F and rear region B). The average pressure value of each region is calculated, including the average pressure of the left region. Average pressure in the right region Average pressure in the front region and the average pressure in the rear region :

[0079] The formula for the average pressure in the left region is: The formula for the average pressure in the right region is: The formula for the average pressure in the front region is: The formula for the average pressure in the rear region is: ;

[0080] in , , , These represent the number of tactile sensors 11 in the left, right, front, and rear areas, respectively, reflecting the overall contact pressure level in the left, right, front, and rear areas.

[0081] Based on the average pressure of each region mentioned above, the pressure difference is calculated to quantify directional imbalance, including:

[0082] left and right pressure difference The formula is The difference in pressure between the left and right sides is represented by absolute values; the greater the difference, the higher the probability of tilting in the left or right direction.

[0083] Pressure difference in the front and rear directions The formula is By demonstrating the pressure difference in the front and rear directions, a basis is provided for judging the tilt in the front and rear directions;

[0084] The angle calculation subunit, based on a preset pressure difference and tilt angle mapping model, converts the extracted pressure difference value into a small tilt angle of the sample. The specific process is as follows:

[0085] The pressure difference-tilt angle mapping model was established through calibration experiments to ensure a quantitative correspondence between the pressure difference and the tilt angle: a series of known minute tilt angles were applied to a standard sample with the same dosage form and known thickness as the drug sample to be tested. Each sensor 11 collects data. Corresponding pressure difference (Including the pressure difference in the left and right directions) Pressure difference in the front and rear directions ); use linear regression algorithm to ( , By fitting the data pairs, a mapping function is obtained: ;in The angle of the sample tilt. The pressure difference is represented by k and b, which are the preset proportional coefficient and correction coefficient, respectively. Both are determined by the calibration data of the standard sample (calculated by the least squares method; different dosage forms need to be recalibrated to update k and b). This model establishes the conversion relationship between the pressure difference and the tilt angle, solving the problem that pressure data cannot directly reflect the degree of tilt.

[0086] The pressure difference in the left and right directions output by the pressure feature extraction subunit. Pressure difference in the front and rear directions Substituting these values ​​into the mapping function, we obtain the tilt angles in the corresponding directions, including:

[0087] Slight tilt angle in the left and right directions The formula is ;like The larger the value, the more significant the optical path offset in the left and right directions;

[0088] Slight tilt angle in the forward and backward direction The formula is ;like The larger the value, the more significant the optical path offset in the forward and backward directions;

[0089] By combining the tilt angles in the left-right and front-back directions, a spatial angle synthesis algorithm is used to calculate the three-dimensional micro-tilt angle of the sample. The formula is as follows: This step fully reflects the spatial tilt state of the sample's installation posture, avoiding the defect that a single-direction angle cannot reflect the overall tilt.

[0090] The original pressure data is denoised and characterized by denoising and region quantization through the pressure feature extraction subunit, providing reliable input for angle calculation;

[0091] The angle calculation subunit achieves quantitative conversion and overall characterization of pressure difference to three-dimensional tilt angle through preset model and spatial synthesis. The two work together to form a complete link consisting of data acquisition, preprocessing, feature extraction and angle calculation, ensuring accurate identification of small tilt of drug sample from 0.5° to 3°, laying the data foundation for subsequent spectral shift correction.

[0092] Specifically, the spectral shift correction unit includes an offset determination subunit and a spectral correction subunit;

[0093] The offset determination subunit, based on the tilt angle output by the micro-tilt recognition unit, calls a preset tilt angle-spectral offset correspondence model to determine the positional offset and intensity distortion of characteristic peaks in the infrared spectrum. The specific process is as follows:

[0094] A pre-defined tilt angle-spectral shift correspondence model was constructed, and the model was calibrated using standard samples to ensure the accuracy of the shift calculation.

[0095] A standard sample with the same dosage form and thickness as the drug sample to be tested was selected, and a series of known three-dimensional micro-tilt angles were applied to the standard sample. After each angle is applied, infrared spectral data of the standard sample are acquired using the infrared detection component. Simultaneously, standard spectral data of standard samples were acquired under 0° tilt (no offset reference state). ( (The wavelength is the infrared light wavelength, determined by the wavelength range of infrared light source 5).

[0096] Feature extraction was performed on each set of data to determine the reference peak positions of key component characteristic peaks in the standard spectrum. (from standard spectrum) (Determining the wavelength corresponding to the maximum absorbance) and reference intensity (Depend on (After determining the absorbance value at the specified location), the characteristic peak positions of the spectrum under tilted conditions are then extracted. and characteristic peak intensity Calculate each group Corresponding peak position shift and intensity distortion The influence of reference strength differences is eliminated through normalization.

[0097] Using linear regression algorithm to and The data pairs are fitted separately to establish corresponding relationship models, including:

[0098] Peak position offset model, the model expression is as follows ;

[0099] The intensity distortion model is expressed as follows: ;

[0100] in, , These represent the scaling factor and correction factor corresponding to the peak position shift model, respectively. , The proportional coefficient and correction coefficient, respectively, represent the intensity distortion model. They are calculated from the calibration data of standard samples using the least squares method. Different drug dosage forms require recalibration to update the coefficients and ensure model adaptability. This model is used to transform the abstract tilt angle into quantifiable spectral distortion parameters, solving the problem of the correlation between tilt state and spectral error.

[0101] The offset determination subunit receives the real-time three-dimensional micro-tilt angle output by the micro-tilt recognition unit. Substituting these values ​​into the two preset models (peak position shift model and intensity distortion model), the real-time peak position shift of the current sample is calculated. and real-time intensity distortion ;

[0102] The spectral correction subunit provides targeted distortion quantification data, ensuring that the correction process is targeted and avoiding secondary errors caused by indiscriminate correction.

[0103] The spectral correction subunit adjusts the position of characteristic peaks based on the positional offset, compensates for spectral intensity based on intensity distortion, and generates corrected spectral data. The specific process is as follows:

[0104] Infrared detector 6 receives the infrared light signal transmitted through the drug sample and acquires the raw infrared spectral data. The raw infrared spectral data includes wavelengths With corresponding absorbance intensity The mapping relationship;

[0105] The key characteristic peaks in the original spectrum are calibrated using a peak-shifting algorithm, as shown in the following formula: ;in The corrected characteristic peak position. The characteristic peak positions in the original spectrum (by (Determine the wavelength corresponding to the local maximum absorbance in the middle); parameters This is a sign function used to determine the original peak position relative to the reference peak position. The direction of the offset, if If the peak position is redshifted, the output will be +1; if the peak position is redshifted, the output will be +1 The output is -1, indicating a blue shift in the peak position, ensuring that the correction direction is opposite to the shift direction. This step corrects the characteristic peak position deviation caused by the optical path shift, so that the peak position returns to the standard position, ensuring the accuracy of component identification.

[0106] For the spectrum after peak position adjustment, an intensity normalization compensation algorithm is used to correct intensity distortion, as shown in the following formula. ;in The intensity of the characteristic peak after compensation, To correct the peak position The corresponding original strength (by) (Obtained through interpolation); By compensating for the light intensity loss or gain caused by the tilt of the optical path, the true intensity of the characteristic peaks is restored, providing accurate intensity data support for subsequent component content calculations;

[0107] Integrate all wavelength-intensity data after characteristic peak position adjustment and intensity compensation to form complete calibrated spectral data. And transmit it to the component data correction unit;

[0108] The offset determination subunit converts the tilt angle into spectral distortion parameters through calibration modeling and real-time calculation, providing a target for correction. The spectral correction subunit accurately restores the original spectrum through peak position adjustment and intensity compensation. Together, they form a complete link of distortion quantification and accurate correction, ensuring that the infrared spectral data can truly reflect the component characteristics of the drug sample, laying a core foundation for the accuracy of subsequent component content calculation.

[0109] Specifically, the component data correction unit includes a model call subunit and a content calculation subunit;

[0110] The model calls a sub-unit to load a pre-defined corrected spectral-component content correlation model. The specific process is as follows:

[0111] Select standard samples with the same dosage form and composition as the drug sample to be tested, and determine their true component content through manual precise value assignment or standard analytical methods (such as high performance liquid chromatography). A series of known minute tilt angles are applied to each standard sample. Corrected spectral data at each tilt angle were collected and generated sequentially. Including corrected peak positions With strength ;

[0112] Partial Least Squares Regression (PLSR) algorithm is used to train the correlation model between the corrected spectral data and the true component content data. The core form of the formula is as follows: Where C represents the predicted content of the drug sample's components. For the intercept term, (k=1,2,...,m) are regression coefficients, all obtained by fitting the training data of the standard samples using the PLSR algorithm, and m is the number of principal components of the model, determined based on the cross-validation error; To correct the spectral data Feature extraction functions (such as band absorbance integral, characteristic peak intensity ratio, etc.) aim to transform high-dimensional spectral data into low-dimensional feature vectors, thereby improving the computational efficiency and stability of the model. By constructing this model, a quantitative mapping relationship between the corrected spectrum and the true component content can be effectively established, solving the error problem caused by directly calculating the component content using the original spectrum or the uncorrected spectrum.

[0113] The model calling subunit stores the trained association model in the storage unit of the data processing module. During the detection process, it receives the corrected spectral data of the sample to be detected from the spectral shift correction unit in real time. It also invokes the association model to provide a calculation interface for the content calculation subunit;

[0114] The content calculation subunit inputs the corrected spectral data output by the spectral shift correction unit into the correlation model, calculates and outputs the true component content of the drug sample, and the specific process is as follows:

[0115] Corrected spectral data output by the spectral shift correction unit Extract spectral features consistent with those used during the training of the association model. ;

[0116] Extracted spectral features Substituting into the spectrum-component content correlation model, the component content C of the drug sample is calculated using the following formula. ; Through quantitative calculation using the spectrum-component content correlation model, the output can accurately reflect the content data of drug components, thus solving the problem of misjudgment of component content caused by spectral distortion;

[0117] The content calculation subunit transmits the calculated component content C to the console 2 for display, and also transmits it to the control module;

[0118] The model calling subunit provides a mathematical model for calculating component content through offline training and online loading. The content calculation subunit realizes the quantitative conversion from corrected spectrum to component content through feature extraction and model substitution. The two work together to form a complete link of model support and data calculation, ensuring that the final output component content data can truly reflect the quality status of drug samples and provide accurate basis for drug production quality control and process regulation.

[0119] Specifically, the corrected spectrum-component content correlation model is a mathematical model obtained by training corrected spectral data of multiple samples with known component contents at different tilt angles through partial least squares regression, support vector machine or neural network algorithm.

[0120] This invention provides an online detection device for pharmaceutical materials, the core of which is to solve the detection distortion problem caused by slight tilting of pharmaceutical samples, and to achieve accurate online detection of pharmaceutical components. The working principle of this invention is as follows:

[0121] Sample installation and positioning: Open the test chamber cover 4, place the drug sample into the annular groove on the outer end face of the I-plate 9, and fix it magnetically by the limiting ring 3; after visually confirming that the sample is flush with the annular groove and accurately aligned with the second through hole on the I-plate 9, insert the I-plate 9 into the limiting groove formed by the limiting plates on the four sides of the I-plate 8 until the rectangular block on the bottom surface of the I-plate 9 is completely embedded in the second rectangular groove on the top surface of the base 7, completing the initial positioning; the spring in the spring box 10 pushes the locking block through the limiting block, causing the locking block to slide into the locking grooves on both sides of the I-plate 9 for secondary fixation. At the same time, the arc groove on the bottom surface of the rectangular block makes precise contact with the tactile sensor 11 in the second rectangular groove. The tactile sensor 11 transmits the installation signal to the control console 2, which displays the installation status. If there is an abnormal contact, an alarm is issued to ensure that the sample does not shift significantly.

[0122] Data Acquisition: After the cover plate 4 is closed, the control panel 2 controls the infrared light source 5 to start. The infrared light passes through the first through hole of the I-beam plate 8, the second through hole of the straight plate 9 and the drug sample in sequence, and is finally received by the infrared detector 6. The infrared detector 6 converts the light signal into infrared spectral data. At the same time, the tactile sensor 11 collects the contact pressure distribution data between the sample and the supporting structure. Both types of data are transmitted to the sensing acquisition module for aggregation.

[0123] Data processing and correction: The data processing module first identifies the three-dimensional micro tilt angle of the sample by collecting pressure data through the tactile sensor 11; then, based on the tilt angle, it corrects the peak position shift and intensity distortion of the infrared spectrum; finally, it loads the preset corrected spectrum-component content correlation model to calculate the true component content of the sample.

[0124] Result Judgment and Linkage: The control module compares the corrected component content and slight tilt angle with the preset component qualification standards and tilt thresholds to determine whether the sample is qualified. The results are displayed in real time on console 2. If the sample is qualified, a green light will illuminate and the data will be recorded. If the tilt exceeds the standard or the component is unqualified, a yellow light will illuminate or an alarm will be triggered. At the same time, instructions are sent to the external production system to achieve linkage between "qualified - continuous production", "tilt exceeding the standard - suspend feeding", and "component unqualified - process adjustment". Unqualified samples will be removed from the testing station.

[0125] Sample replacement: After the test is completed, pull the pull block outside the spring box 10. The pull rod drives the limit block to compress the spring, causing the locking block to disengage from the slot of the plate 9. At this time, the hook-shaped locking block 12 on the bottom of the pull block hooks onto the U-shaped limit block on the bottom of the spring box 10 to achieve state locking. The staff pulls out the plate 9 through the first rectangular slot on the outer wall of the plate 9, replaces the new sample, and enters the next round of testing.

[0126] The formulas involved in this device, such as pressure noise reduction, tilt angle calculation, spectral shift correction, and component content calculation, all adopt dimensionless numerical calculations. These can be achieved through conventional methods such as standardization, which will not be elaborated here. The formulas are obtained by software simulation and fitting based on a large amount of experimental data, which can closely approximate the real detection scenario. The preset parameters in the formulas (such as the filter window length N, the proportional coefficient k, etc.) can be flexibly set by those skilled in the art according to the actual situation such as drug dosage form and detection accuracy requirements.

[0127] This embodiment can be implemented by software, hardware, firmware, or any combination thereof. If implemented by software, it can form a computer program product. After the computer instructions are loaded or executed, they can drive the device to complete functions such as data acquisition, processing, correction, and linkage control. The instructions can be transmitted via wired or wireless means, and the storage medium can be conventional carriers such as USB flash drive, hard disk, ROM, and RAM.

[0128] The execution order of each process in the equipment is determined by the functional logic and is not restricted by the sequence number; the functions of each unit can be implemented through electronic hardware or a combination of hardware and software. The unit division is only a logical functional division. In actual applications, it can be merged or split according to design requirements, and some non-core functions can be selected to be executed as needed.

[0129] The separate components (such as tactile sensor 11, infrared detector 6, etc.) can be physically separated or integrated. The functional units can be integrated into the same processing unit or exist separately. Some or all of the units can be selected to realize the detection function according to actual needs.

[0130] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

[0131] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. An online testing device for pharmaceutical materials, comprising a device housing (1), a control console (2) provided on one side of the top surface of the device housing (1), and a testing chamber opened on the other side of the top surface of the device housing (1), characterized in that, The detection chamber is equipped with an installation status detection component and an infrared detection component. The installation status detection component includes at least one set of tactile sensors (11) for collecting contact pressure distribution data between the drug sample and the sample support structure. The infrared detection component consists of an infrared light source (5) and an infrared detector (6) for collecting infrared spectral data of the drug sample. The control console (2) is equipped with an integrated circuit board, which is equipped with a sensing acquisition module, a data processing module and a control module. The sensing and acquisition module is used to collect contact pressure distribution data and infrared spectral data; The data processing module includes a micro-tilt recognition unit, a spectral shift correction unit, and a component data correction unit; The micro-tilt recognition unit is used to identify the micro-tilt angle of the drug sample; The spectral shift correction unit is used to correct infrared spectral shifts based on a small tilt angle. The component data correction unit is used to calculate the true component content of the drug sample based on the corrected spectral data. The control module is used to control the equipment's operating status based on the corrected component content and slight tilt angle information output by the data processing module, combined with the preset component qualification standards and tilt threshold, and to send control commands to the external production system. The micro-tilt recognition unit includes a pressure feature extraction subunit and an angle calculation subunit; The pressure feature extraction subunit is used to extract the pressure values ​​and pressure differences in different regions of the sample bearing structure after preprocessing and feature quantization of the contact pressure distribution data. The angle calculation subunit converts the extracted pressure difference value into a tiny tilt angle of the sample based on a preset pressure difference and tilt angle mapping model.

2. The online detection device for pharmaceutical materials according to claim 1, characterized in that, The infrared light source (5) in the infrared detection assembly is located on one inner wall of the detection chamber, and the infrared detector (6) is located on the other inner wall of the detection chamber. A base (7) is provided on the bottom surface of the detection chamber between the infrared light source (5) and the infrared detector (6). An I-beam (8) is fixedly connected to the top surface of the base (7). A first through hole coaxial with the infrared detector (6) is opened on the I-beam (8). Two opposing limiting plates are fixedly connected to the four sides of the I-beam (8).

3. The online detection device for pharmaceutical materials according to claim 2, characterized in that, The sample carrying structure includes a plate (9) inserted on the I-beam plate (8). The plate (9) has a second through hole that is coaxial with the first through hole and has the same diameter. The outward end face of the plate (9) has an annular groove for placing the drug sample at the second through hole. A limiting ring (3) for pressing the drug sample is magnetically attached at the annular groove.

4. The online detection device for pharmaceutical materials according to claim 3, characterized in that, The bottom surface of the I-beam (9) is fixed with a rectangular block. The bottom surface of the rectangular block is provided with multiple arc-shaped grooves at equal intervals. The top surface of the base (7) is symmetrically provided with two second rectangular grooves located on both sides of the I-beam (8) and adapted to the rectangular block. The tactile sensor (11) is disposed on the bottom surface of the second rectangular groove and adapted to the arc-shaped groove.

5. The online detection device for pharmaceutical materials according to claim 1, characterized in that, The spectral shift correction unit includes an offset determination subunit and a spectral correction subunit; The offset determination subunit determines the positional offset and intensity distortion of the characteristic peaks in the infrared spectrum based on the tilt angle output by the micro tilt recognition unit and by calling the preset tilt angle-spectral offset correspondence. The spectral correction subunit adjusts the position of the characteristic peaks according to the positional offset and compensates for the spectral intensity according to the intensity distortion to generate corrected spectral data.

6. The online detection device for pharmaceutical materials according to claim 5, characterized in that, The relationship between the tilt angle and the spectral shift is established in the following way: Apply different minute tilt angles to the standard sample, collect infrared spectral data at the corresponding angles, compare the spectral data at the standard angle, record the characteristic peak position offset and intensity distortion, and form a mapping relationship database.

7. The online detection device for pharmaceutical materials according to claim 1, characterized in that, The component data correction unit includes a model calling subunit and a content calculation subunit; The model calls a subunit to load a preset corrected spectral-component content correlation model; The content calculation subunit inputs the corrected spectral data output by the spectral shift correction unit into the correlation model to calculate and output the true component content of the drug sample.

8. The online detection device for pharmaceutical materials according to claim 7, characterized in that, The corrected spectrum-component content correlation model is a mathematical model obtained by training corrected spectral data of multiple samples with known component contents at different tilt angles through partial least squares regression, support vector machine or neural network algorithm.

Images