Automatic separation and purification system for radionuclides

Abstract

The invention relates to the technical field of nuclide separation and purification, and discloses an automatic separation and purification system for radionuclides, comprising: an initialization module generating valve position zero reference data; the instruction generation module generates a valve group control instruction and a pump operation instruction, and determines a liquid phase conveying route; the purification control module carries out reagent filling operation; carrying out nuclide adsorption operation and collecting radiation data in a nuclide adsorption stage; determining a column loading process monitoring result, and judging whether nuclide adsorption operation is completed or not; carrying out leaching operation; carrying out elution operation; collecting radiation data in an elution stage, generating an elution feature vector, and performing similarity calculation on the elution feature vector and a target nuclide elution feature vector; and generating an output instruction. According to the full-process closed-loop automatic control, the manual intervention link is reduced, and the radiation exposure risk of operators and the manual operation error are reduced.

Description

Technical Field

[0001] This invention relates to the field of radionuclide separation and purification technology, and more specifically, to an automated separation and purification system for radionuclides. Background Technology

[0002] In the field of nuclear medicine, Cu 64 Ga 68 Zr 88 Radionuclides, due to their unique nuclear physics properties, are widely used in PET imaging fields such as tumor diagnosis and treatment efficacy evaluation. The separation and purification of radionuclides is a crucial step in their clinical application, directly affecting imaging quality and diagnostic and treatment outcomes.

[0003] Currently, existing radionuclide separation and purification devices have the following shortcomings: low level of automation, relying heavily on manual operation, which is not only inefficient but also increases the risk of operators coming into contact with radioactive materials; poor compatibility, with most devices designed only for a single radionuclide, making it difficult to meet the needs of multi-radionuclide separation and purification; unreasonable structural design, with some key components made of ordinary metal materials, which are susceptible to corrosion by radionuclide solutions, affecting the device's lifespan and separation purity; and a lack of integrated detection and control modules, making it impossible to monitor the separation process in real time, resulting in unstable purification accuracy.

[0004] Therefore, it is necessary to design an automated separation and purification system for radionuclides to address the problems existing in current technologies. Summary of the Invention

[0005] In view of this, the present invention proposes an automated separation and purification system for radionuclides, which aims to solve the problems of low automation, poor compatibility, lack of integrated detection and control modules, inability to monitor the separation process in real time, and unstable purification accuracy.

[0006] This invention proposes an automated separation and purification system for radionuclides, comprising: The initialization module is used to acquire the valve position signal, initialize and generate the valve position zero-point reference data; The instruction generation module is used to obtain the nuclide type, generate process parameters according to the separation and purification process parameter set and the nuclide type, generate valve group control instructions and pump operation instructions based on the valve position zero point reference data and process parameters, and determine the liquid phase delivery route. The purification control module is used to perform reagent addition operations according to the liquid phase delivery route; after the reagent addition operation is completed, a nuclide adsorption operation is performed and radiation data of the nuclide adsorption stage is collected; a denoised radiation intensity sequence is generated based on the radiation data of the nuclide adsorption stage; the column loading dynamic trend data is determined according to the denoised radiation intensity sequence; the column loading dynamic trend data is compared with the dynamic baseline to determine the column loading process monitoring result; the completion of the nuclide adsorption operation is determined according to the column loading process monitoring result; when the nuclide adsorption operation is completed, a rinsing operation is performed; after the rinsing operation is completed, an elution operation is performed; radiation data of the elution stage is collected, peak detection and feature parameter extraction are performed to generate an elution feature vector; the similarity between the elution feature vector and the target nuclide elution feature vector is calculated; and an output command is generated according to the similarity calculation result.

[0007] Furthermore, when determining the monitoring results of the column loading process, the purification control module includes: The radiation data from the radionuclide adsorption stage are subjected to sliding window mid-range filtering to generate a denoised radiation intensity sequence; based on the denoised radiation intensity sequence, dynamic trend data of the upper column is generated. The dynamic trend data of the upper column is compared with the dynamic baseline threshold to generate the monitoring results of the upper column process.

[0008] Furthermore, when the purification control module determines whether the radionuclide adsorption operation is complete, it includes: When the monitoring results of the column loading process indicate that the adsorption process is normal and the cumulative adsorption time reaches the adsorption time in the process parameter data, the radionuclide adsorption is completed. When the monitoring results of the column loading process indicate an abnormality in the adsorption process, an abnormality prompt message for column loading is generated.

[0009] Furthermore, when the purification control module generates the elution feature vector, it includes: Peak detection was performed on the radiation data during the elution stage, and the peak intensity, rise time, and half-width at half-maximum parameters were extracted. The peak intensity, rise time, and half-peak width parameters are combined to form the elution feature vector.

[0010] Furthermore, when generating output instructions, the purification control module includes: The cosine similarity between the elution feature vector and the target nuclide elution feature vector is calculated to generate a similarity value.

[0011] Furthermore, the purification control module, when generating output instructions, also includes: When the similarity value is greater than the similarity threshold, a product output instruction is generated; When the similarity value is less than or equal to the similarity threshold, an elution anomaly warning message is generated.

[0012] Furthermore, when the initialization module acquires the valve position signal, initializes and generates the valve position zero-point reference data, it includes: Initialize the rotation of the multi-channel switching valve group and synchronously acquire the timing pulse signal sequence; Based on the time-series pulse signal sequence, the feature pulse corresponding to the initial positioning identifier is identified, and the valve position code value is recorded each time the feature pulse is identified, generating a valve position position dataset; outlier removal is performed on the valve position position dataset, and valid position data is retained; the mean and standard deviation of the valid position data are determined.

[0013] Furthermore, when the initialization module acquires the valve position signal, initializes and generates the valve position zero-point reference data, it also includes: When the standard deviation is less than the tolerance threshold, the mean value is determined as the valve position zero-point reference data; when the standard deviation is greater than or equal to the tolerance threshold, the rotation is re-initialized and the valve position dataset is updated.

[0014] Furthermore, when the instruction generation module generates process parameters, it includes: The nuclide type is matched with the index field in the separation and purification process parameter set to generate structured process parameter data. The structured process parameter data includes reagent dosing time, adsorption time, rinsing time, elution time, and valve position switching parameters at each stage.

[0015] Furthermore, when the instruction generation module generates valve group control instructions and pump operation instructions, it includes: The valve position zero-point reference data is time-aligned with the valve position switching parameters in the process parameters to generate a valve group control command sequence; a pump operation command sequence is generated based on the pump operation parameters in the process parameters.

[0016] Compared with existing technologies, the advantages of this invention are as follows: By linking a multi-channel valve group with a PLC control system, the entire process of radionuclide separation is automated, reducing the risk of operators coming into contact with radioactive materials; the modular design supports rapid replacement of resin columns and storage units, meeting the requirements of Cu... 64 Ga 68 Zr 88This system is suitable for separating various nuclides, with a wide range of applications. Real-time activity detection and closed-loop control improve the purity of nuclide separation. The initialization module establishes zero-point reference data for valve positions, eliminating mechanical hysteresis and installation deviations, improving the repeatability of flow path switching and the long-term stability of the system. The instruction generation module intelligently matches a pre-set process parameter library based on the nuclide type, dynamically generating valve group control and pump operation instructions. This enables precise planning of the liquid phase transport route and rapid adaptation to multi-nuclide processes, enhancing system flexibility and operational efficiency. The purification control module performs real-time acquisition and adaptive noise reduction of radiation data during the nuclide adsorption stage, generating dynamic trend data for column loading and quantitatively comparing it with the dynamic baseline to objectively determine the adsorption endpoint, avoiding incomplete adsorption or excessive reagent consumption caused by traditional fixed-duration methods. During the elution stage, peak detection of radiation data and extraction of key feature parameters construct an elution feature vector, which is then compared with the target nuclide's standard feature vector for similarity calculation. This enables quantitative verification and automatic determination of the purity and recovery rate of the eluted product, ensuring the reliability and traceability of the separation results. The closed-loop automated control of the entire process reduces manual intervention, thereby lowering the risk of radiation exposure and human error for operators. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a functional block diagram of an automated separation and purification system for radionuclides provided in an embodiment of the present invention. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] In some embodiments of this application, see Figure 1 As shown, an automated separation and purification system for radionuclides is proposed, comprising: The initialization module is used to acquire the valve position signal, initialize and generate the valve position zero-point reference data; The instruction generation module is used to obtain the nuclide type, generate process parameters based on the separation and purification process parameter set and the nuclide type, generate valve group control instructions and pump operation instructions based on valve position zero point reference data and process parameters, and determine the liquid phase delivery route. The purification control module is used to perform reagent addition according to the liquid phase delivery route; after reagent addition, it performs radionuclide adsorption and collects radiation data during the radionuclide adsorption stage; it generates a denoised radiation intensity sequence based on the radionuclide adsorption stage radiation data; it determines the column loading dynamic trend data based on the denoised radiation intensity sequence; it compares the column loading dynamic trend data with the dynamic baseline to determine the column loading process monitoring results; it determines whether the radionuclide adsorption operation is complete based on the column loading process monitoring results; after the radionuclide adsorption operation is complete, it performs rinsing; after rinsing, it performs elution; it collects radiation data during the elution stage, performs peak detection and feature parameter extraction to generate an elution feature vector; it calculates the similarity between the elution feature vector and the target radionuclide elution feature vector; and it generates output instructions based on the similarity calculation results.

[0020] Specifically, the automated separation and purification system consists of a purification module, a control system, a detection module, and a support frame. The purification module includes a PTFE resin column assembly, a five-way three-position valve assembly, a matching motor mounting bracket, a multi-channel switching valve group consisting of a sleeve coupling, and PTFE storage containers in 10ml / 20ml / 50ml / 100ml sizes. The control system is centered on a PLC control unit, connecting the valve group control unit and the peristaltic pump. The detection module includes a radiation probe mounting block and a signal processing unit. The radiation probe collects radiation signals in real time, amplifies and shapes them, and then transmits them to the PLC via 485 communication. The support frame is made of 6061-T6 aluminum alloy, treated with natural anodizing and powder coating. After power-on, the initialization module drives the five-way three-position valve position actuator to rotate. When the sensor block is obstructed by the photoelectric sensor component, the current position is calibrated as the valve position zero-point reference data. If initialization fails, the human-machine interface displays a fault message. The instruction generation module receives the nuclide type input from the operator (e.g., ...). 68 Ga、 99The purification control module retrieves matching structured process parameters (including duration of each stage and valve position switching parameters) from the built-in process parameter library, and generates valve group control command sequences (via Modbus RTU communication) and pump operation command sequences (digital / analog control) based on the valve position zero-point reference data to accurately plan the liquid phase delivery route. The purification control module controls the peristaltic pump and valve group to perform reagent dosing operations according to the command sequences. After completion, the radionuclide adsorption process is started: the radiation probe collects radiation data in real time during the adsorption stage, the signal processing unit performs sliding window mid-range filtering to generate a noise-reduced radiation intensity sequence, the PLC analyzes its changing trend to generate dynamic trend data for column loading, and performs quantitative comparison with the dynamic baseline established based on historical normal operating conditions. If the radiation intensity continues to rise and then tends to stabilize and the cumulative time reaches the process set value, it is determined that the adsorption is complete and the rinsing operation begins; if the radiation signal is not significant... Changes (indicating target dissolution or abnormal transport) trigger an abnormality warning message on the column. After rinsing, the elution operation is performed, and radiation data during the elution stage is collected simultaneously. Peak values ​​are detected, and the peak intensity, rise time, and half-maximum width are extracted and combined into an elution feature vector. Cosine similarity is calculated with the pre-stored target nuclide standard feature vector. If the similarity value is greater than the threshold, a product output command is generated, the control valve group is switched to the liquid transfer line, and the nitrogen solenoid valve is opened to output high-purity product. Otherwise, an elution abnormality warning message is generated. The entire process operates automatically in a closed loop under a shielded environment. All process parameters, radiation data, and judgment results are archived in real time, and operators only need to monitor remotely.

[0021] Understandably, the high-precision dynamic calibration of the valve position zero point is achieved through the collaborative use of photoelectric sensing components and induction plates, eliminating mechanical hysteresis and installation deviations, ensuring the repeatability of the five-way three-position valve switching, controlling the risk of reagent cross-contamination, and meeting the cleanliness requirements for the separation of highly active nuclides. The PTFE resin column and storage unit combine excellent corrosion resistance, low adsorption characteristics, and cost advantages, ensuring nuclide recovery rate and product purity. The online monitoring system composed of radiation probes and signal processing units upgrades the control logic from fixed-duration driven to process-state driven: during the adsorption stage, the endpoint is objectively determined by comparing the dynamic radiation trend with the dynamic baseline, avoiding incomplete adsorption or reagent waste; during the elution stage, product purity is quantitatively verified based on feature vector similarity, achieving real-time closed-loop verification of separation quality, improving the reliability and traceability of results. The fully automated operation reduces manual intervention and lowers the radiation exposure risk for operators.

[0022] In some embodiments of this application, when the purification control module determines the monitoring results of the column loading process, it includes: A sliding window mid-range filter is applied to the radiation data during the nuclide adsorption stage to generate a denoised radiation intensity sequence; based on the denoised radiation intensity sequence, dynamic trend data of the upper column is generated. The dynamic trend data of the upper column is compared with the dynamic baseline threshold to generate the monitoring results of the upper column process.

[0023] Specifically, the raw radiation data (sampling frequency 1Hz) collected by the radiation probe during the nuclide adsorption stage is subjected to sliding window median filtering: the window length is set to 5 consecutive data points, and the median radiation intensity within the window is calculated sequentially, effectively filtering out random noise such as environmental background fluctuations and instantaneous pulse interference from the detector, generating a smooth and continuous denoised radiation intensity sequence; then, based on this sequence, the intensity change rate at adjacent time points is calculated using the first-order difference method, and combined with the moving average (3 points in the window) to generate dynamic trend data of the upper column characterizing the dynamic evolution of the adsorption rate; simultaneously, the system uses the initial 10-second stable period of the current batch as a basis for further analysis. The average radiation intensity is dynamically used to generate a baseline threshold (set to 1.15 times the average to compensate for minor background drift). The dynamic trend data of the adsorption process is then compared with this dynamic baseline threshold in real time. If the trend data is consistently higher than the threshold and exhibits the typical adsorption curve characteristics of rapid increase—slower growth—and stabilization, a normal adsorption process monitoring result is generated. If the trend data is consistently lower than the threshold (e.g., an increase of less than 5% for 15 consecutive seconds) or exhibits abnormal fluctuations (e.g., a sudden drop of more than 20%), the adsorption process is judged to be abnormal, and the type of abnormality (e.g., undissolved target, blocked flow path) and the timestamp of occurrence are marked. Understandably, the sliding window mid-range filtering significantly improves the signal-to-noise ratio of radiation data, suppresses impulse noise and background interference, and ensures that dynamic trend data truly reflects the adsorption physical process. The dynamic baseline threshold is generated in real time based on the initial operating conditions of the current batch, overcoming the adaptability defects of fixed thresholds under different nuclides and environmental background fluctuations, making the monitoring logic closely match the actual adsorption state. The generated column loading process monitoring results not only clearly indicate whether the process is normal or not, but also include abnormality type and time location information, providing highly reliable input for accurate determination of the adsorption endpoint and avoiding incomplete adsorption or excessive reagent consumption caused by the traditional fixed duration method.

[0024] In some embodiments of this application, when the purification control module determines whether the radionuclide adsorption operation is complete, it includes: When the monitoring results of the column loading process indicate that the adsorption process is normal and the cumulative adsorption time reaches the adsorption time in the process parameter data, the adsorption of the generated nuclide is complete. When the monitoring results of the column loading process indicate an abnormality in the adsorption process, an abnormality prompt message for the column loading process is generated.

[0025] Specifically, when the purification control module determines whether the radionuclide adsorption operation is complete, it processes two key pieces of information in real time: first, it receives the column loading process monitoring results output by the column loading process monitoring module (the results clearly indicate whether the adsorption process is normal or abnormal, and include an abnormality type code and a timestamp); second, it continuously records the cumulative duration since the start of the adsorption operation through the system's internal high-precision timer. When the monitoring results consistently indicate that the adsorption process is normal (i.e., the dynamic trend of radiation intensity conforms to the rising-plateau characteristic and there are no abnormal fluctuations), and the cumulative duration accurately reaches the preset adsorption duration threshold in the process parameter data, the module immediately generates a structured radionuclide adsorption completion status signal and automatically triggers the rinsing operation command sequence; if the monitoring results indicate that the adsorption process is abnormal at any time (e.g., the radiation signal increases by less than 5% for 15 consecutive seconds, drops by more than 20%, or deviates from the dynamic baseline threshold), the module immediately stops timing and the adsorption process, generates a column abnormality prompt message containing the abnormality type (e.g., target not dissolved, flow path blockage, resin failure), occurrence time, and suggested handling measures, and highlights it in the form of audible and visual warnings through the human-machine interface, simultaneously encrypts and archives the abnormal data to the system log, and suspends subsequent operations for manual confirmation, effectively preventing the transmission of abnormal status to the rinsing and elution stages.

[0026] Understandably, the dual judgment logic of process status validity and time threshold compliance avoids the inherent defects of traditional single fixed time control: it prevents separation failure and reagent waste caused by mistakenly entering subsequent steps when the adsorption is abnormal due to time exceeding the standard, and avoids premature termination of effective adsorption caused by occasional fluctuations in process monitoring; the abnormal prompt information is structured and traceable, which significantly shortens the fault location and handling time and reduces the risk of cross-contamination and resource consumption caused by ineffective operations.

[0027] In some embodiments of this application, the purification control module generates the elution feature vector, including: Peak detection was performed on the radiation data during the elution stage, and parameters such as peak intensity, rise time, and half-peak width were extracted. The peak intensity, rise time, and half-peak width parameters are combined to form the elution feature vector.

[0028] Specifically, the raw radiation data (sampling frequency 1Hz) collected by the radiation probe during the elution stage is processed in real time: after suppressing high-frequency noise by using a sliding window smoothing method (window length 3 points), the peak point of the radiation intensity curve is accurately located based on the preset threshold method and the joint judgment strategy of the zero crossing point of the first derivative; the peak intensity value (unit: cps) is used as the first feature parameter; the curve is traced back to the time point when the intensity drops to 50% of the peak value, and the time interval from this point to the peak point is calculated as the rise time (unit: seconds), which characterizes the elution kinetic response speed; the difference between the time points on the left and right sides is measured at the 50% intensity level of the peak value to determine the half-peak width (unit: seconds), which reflects the concentration of the elution peak shape and the level of impurity interference; the above three parameters are combined in a fixed order (peak intensity, rise time, half-peak width) to form a three-dimensional elution feature vector.

[0029] Understandably, peak intensity is related to the total amount of target nuclide eluted, rise time reflects the resin desorption kinetics (such as flow rate matching degree and resin activity), and half-peak width indicates product purity and the degree of impurity interference (narrow peaks correspond to high purity). The feature vector composed of the three parameters enhances the identification dimension and anti-interference ability of elution peaks, and distinguishes the characteristic peaks of target nuclide from impurity peaks, tailing peaks and other abnormal forms.

[0030] In some embodiments of this application, the purification control module, when generating output instructions, includes: The cosine similarity between the elution feature vector and the target nuclide elution feature vector is calculated to generate a similarity value.

[0031] In some embodiments of this application, the purification control module further includes the following when generating output instructions: When the similarity value is greater than the similarity threshold, a product output instruction is generated; When the similarity value is less than or equal to the similarity threshold, an elution anomaly warning message is generated.

[0032] Specifically, the system calls the pre-stored target nuclide standard elution feature vector (generated from historical qualified batches through process verification and calibration, including three parameters: peak intensity, rise time, and half-maximum width) in the system feature library, and performs cosine similarity calculation with the elution feature vector generated in the current elution stage: by calculating the cosine value of the angle between the two vectors (normalized to the 0-1 range), a similarity value representing the morphological matching degree is generated; the similarity threshold is preset to 0.92, determined based on statistical analysis of hundreds of qualified product data (taking the lower limit of the 95% confidence interval), and supports dynamic calibration by nuclide type through the human-machine interface. When the similarity value is greater than the threshold, the module immediately generates a product output command, controls the valve group to switch to the liquid transfer line, and simultaneously outputs a digital signal to open the nitrogen solenoid valve, delivering the high-purity target product to the shielded collection container in a positive pressure push manner; when the similarity value is less than or equal to the threshold, the module generates a structured elution anomaly prompt message, which includes the current similarity value, deviation feature analysis (such as excessive half-maximum width indicating impurity interference, slow rise time indicating abnormal flow rate), suggested handling measures, and a timestamp.

[0033] Understandably, cosine similarity calculation captures the overall characteristics of the elution peak shape (rather than a single peak), identifies peak distortion caused by flow rate deviation, resin activity decay, or impurity interference, and improves the objectivity and anti-interference ability of purity determination; the dynamic threshold mechanism based on historical qualified data balances the rigor of determination with process adaptability, avoiding the risk of misjudgment between different nuclides and batches with fixed thresholds; the automatic closed-loop logic of product output and anomaly interception realizes real-time closed-loop control of the quality at the end of separation, blocking the outflow of unqualified products from the source.

[0034] In some embodiments of this application, when the initialization module acquires the valve position signal and initializes and generates valve position zero-point reference data, it includes: Initialize the rotation of the multi-channel switching valve group and synchronously acquire the timing pulse signal sequence; The system identifies the characteristic pulses corresponding to the initial positioning marker based on the time-series pulse signal sequence, and records the valve position code value each time a characteristic pulse is identified, generating a valve position position dataset. Outlier points are removed from the valve position position dataset, and valid position data is retained. The mean and standard deviation of the valid position data are determined.

[0035] In some embodiments of this application, when the initialization module acquires the valve position signal, initializes and generates valve position zero-point reference data, it further includes: When the standard deviation is less than the tolerance threshold, the mean is determined as the valve position zero-point reference data; when the standard deviation is greater than or equal to the tolerance threshold, the rotation is re-initialized and the valve position dataset is updated.

[0036] Specifically, the valve group control unit drives the multi-channel switching valve group (such as a five-way three-way valve assembly) to complete the initial rotation at a low and uniform speed, while simultaneously acquiring a time-series pulse signal sequence in real time via an encoder. The system accurately identifies the characteristic pulse corresponding to the initial positioning marker generated by the photoelectric sensing component being blocked by the sensing sheet within the time-series pulse signal sequence. At each time this characteristic pulse is detected, the valve position encoder's feedback valve position code value is recorded synchronously, and this process is repeated ten times to form an initial valve position dataset. Subsequently, outlier detection is performed on the dataset according to preset statistical judgment rules (such as data points deviating from the group mean by more than a reasonable range). The process involves filtering out outliers caused by mechanical vibration or signal interference, retaining only valid position data. The arithmetic mean and standard deviation of the valid position data are calculated. If the standard deviation is less than a preset tolerance threshold (calibrated based on valve assembly mechanical precision and system reliability requirements, e.g., corresponding to an angle deviation of 0.05 degrees), the arithmetic mean is used as the valve position zero-point reference data and stored in the system parameter library. If the standard deviation is greater than or equal to the tolerance threshold, the initialization data is deemed too volatile, automatically triggering a re-initialization process. This involves controlling the valve assembly to rotate again and updating the valve position dataset, repeating the above process until the precision requirements are met.

[0037] Understandably, the closed-loop calibration mechanism, which involves multiple sampling, intelligent outlier removal, and standard deviation tolerance verification, eliminates the impact of mechanical hysteresis, installation deviation, environmental vibration, and random errors in single measurements on zero-point positioning, thereby improving the accuracy and long-term stability of the valve position zero-point reference data. The automatic retry logic enhances the system's initialization robustness and avoids calibration failure due to occasional interference.

[0038] In some embodiments of this application, the instruction generation module generates process parameters, including: The nuclide type is matched with the index fields in the separation and purification process parameter set to generate structured process parameter data. The structured process parameter data includes reagent dosing time, adsorption time, rinsing time, elution time, and valve position switching parameters at each stage.

[0039] Specifically, the nuclide type identifier input by the receiving operator (such as...) 68 Ga、 99 ᵐTc、 177The module accesses the separation and purification process parameter set pre-stored in the system storage unit (this parameter set is stored in the form of a structured database, containing nuclide type index fields and associated complete process records); the module performs an exact match search between the nuclide type identifier and the index fields in the parameter set. If a match is successful, the corresponding process record is extracted, and structured process parameter data is generated: this data is organized in a hierarchical key-value pair format, clearly including the addition time and order of each reagent (such as activation solution, equilibration solution) in the reagent addition stage, the nuclide adsorption time, the type and duration of the rinsing solution in the rinsing stage, the type and duration of the elution solution in the elution stage, and the valve position switching parameters corresponding to each process stage (including target valve position number, switching sequence, holding time, and linkage logic with the peristaltic pump); if the match fails (such as the nuclide type not being registered in the parameter set), the module immediately generates a nuclide type unrecognized prompt message and stops the process to ensure the accuracy and safety of parameter calls; the generated structured process parameter data is cached in real time to the control memory for direct use by the subsequent valve group control command and pump operation command generation modules.

[0040] Understandably, by using a precise indexing and matching mechanism between nuclide types and process parameter sets, error-free calling of multi-nucleoside separation process parameters within seconds is achieved, avoiding risks such as parameter omissions, timing misalignments, or unit confusion that are prone to occur in manual configuration, thereby improving the accuracy of process execution and batch consistency. The structured process parameter data completely encapsulates the timing logic and equipment linkage rules of each stage, providing highly reliable input for liquid phase transport route planning and ensuring that the separation process strictly follows the pre-validated process specifications.

[0041] In some embodiments of this application, when the instruction generation module generates valve group control instructions and pump operation instructions, it includes: The valve position zero-point reference data is time-aligned with the valve position switching parameters in the process parameters to generate a valve group control command sequence; the pump operation command sequence is generated based on the pump operation parameters in the process parameters.

[0042] Specifically, the valve position zero-point reference data calibrated by the initialization module (i.e., the average value of the valve position code verified through multiple samplings) is used as the absolute position reference origin. This data is then aligned with the valve position switching parameters (including target valve position number, switching start time, and holding duration) for each process stage in the structured process parameter data at the millisecond level. Based on the difference between the zero-point reference and the target position, the actual rotation is calculated. Combined with the preset acceleration / deceleration curves of the valve group's mechanical characteristics, a valve group control command sequence is generated, containing the target position code, switching timing marker, arrival confirmation logic, and abnormal timeout threshold. This sequence is then transmitted via Modbus. The RTU communication protocol accurately sends data to the valve group control unit in time-stamped order; it simultaneously analyzes the pump operating parameters in the process parameters (including flow rate setpoints, running time, start-stop trigger conditions, and linkage timing with valve position switching) to generate a pump operating command sequence containing direction signals, analog speed commands, operating cycles, and status feedback requirements. This sequence is then precisely controlled by the PLC control unit through digital output points and analog output modules, according to a preset timing sequence, to ensure that the liquid phase flows strictly along the planned route, achieving seamless coordination between valve group switching and pumping actions in the time dimension.

[0043] Understandably, by aligning the valve position zero point reference with the process parameters at the millisecond level, the influence of mechanical zero point drift on the flow path positioning is eliminated, ensuring the absolute accuracy of the five-way three-way valve in switching positions in complex processes, and eliminating the risk of reagent misflow or cross-contamination caused by valve position deviation from the source; the coordinated generation and precise timing binding of the valve group control command sequence and the pump operation command sequence realizes the millisecond-level synchronization of flow path switching and liquid delivery, ensuring that key steps such as reagent addition, adsorption, rinsing, and elution are seamlessly connected according to the preset logic.

[0044] In summary, the linkage between the multi-channel valve group and the PLC control system achieves full automation of the radionuclide separation process, reducing the risk of operators coming into contact with radioactive materials. The modular design supports rapid replacement of resin columns and storage units, meeting the requirements of Cu... 64 Ga 68 Zr 88This system is suitable for separating various nuclides, with a wide range of applications. Real-time activity detection and closed-loop control improve the purity of nuclide separation. The initialization module establishes zero-point reference data for valve positions, eliminating mechanical hysteresis and installation deviations, improving the repeatability of flow path switching and the long-term stability of the system. The instruction generation module intelligently matches a pre-set process parameter library based on the nuclide type, dynamically generating valve group control and pump operation instructions. This enables precise planning of the liquid phase transport route and rapid adaptation to multi-nuclide processes, enhancing system flexibility and operational efficiency. The purification control module performs real-time acquisition and adaptive noise reduction of radiation data during the nuclide adsorption stage, generating dynamic trend data for column loading and quantitatively comparing it with the dynamic baseline to objectively determine the adsorption endpoint, avoiding incomplete adsorption or excessive reagent consumption caused by traditional fixed-duration methods. During the elution stage, peak detection of radiation data and extraction of key feature parameters construct an elution feature vector, which is then compared with the target nuclide's standard feature vector for similarity calculation. This enables quantitative verification and automatic determination of the purity and recovery rate of the eluted product, ensuring the reliability and traceability of the separation results. The closed-loop automated control of the entire process reduces manual intervention, thereby lowering the risk of radiation exposure and human error for operators.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. An automated separation and purification system for radionuclides, characterized in that, include: The initialization module is used to acquire the valve position signal, initialize and generate the valve position zero-point reference data; The instruction generation module is used to obtain the nuclide type and generate process parameters based on the separation and purification process parameter set and the nuclide type; Based on the valve position zero-point reference data and process parameters, valve group control commands and pump operation commands are generated, and the liquid phase delivery route is determined. The purification control module is used to perform reagent dispensing operations according to the liquid phase delivery route; After the reagent addition operation is completed, the radionuclide adsorption operation is performed and radiation data of the radionuclide adsorption stage is collected. A denoised radiation intensity sequence is generated based on the radiation data from the radionuclide adsorption stage. Based on the denoised radiation intensity sequence, determine the dynamic trend data of the upper column; The dynamic trend data of the column loading process is compared with the dynamic baseline to determine the monitoring results of the column loading process; based on the monitoring results of the column loading process, it is determined whether the radionuclide adsorption operation is completed; when the radionuclide adsorption operation is completed, a rinsing operation is performed; after the rinsing operation is completed, an elution operation is performed. Radiation data during the elution stage is collected, peak detection and feature parameter extraction are performed to generate elution feature vectors; the similarity between the elution feature vectors and the target nuclide elution feature vectors is calculated; and output instructions are generated based on the similarity calculation results.

2. The automated separation and purification system for radionuclides according to claim 1, characterized in that, When the purification control module determines the monitoring results of the column loading process, it includes: The radiation data from the radionuclide adsorption stage are subjected to sliding window mid-range filtering to generate a denoised radiation intensity sequence; based on the denoised radiation intensity sequence, dynamic trend data of the upper column is generated. The dynamic trend data of the upper column is compared with the dynamic baseline threshold to generate the monitoring results of the upper column process.

3. The automated separation and purification system for radionuclides according to claim 2, characterized in that, When the purification control module determines whether the radionuclide adsorption operation is complete, it includes: When the monitoring results of the column loading process indicate that the adsorption process is normal and the cumulative adsorption time reaches the adsorption time in the process parameter data, the radionuclide adsorption is completed. When the monitoring results of the column loading process indicate an abnormality in the adsorption process, an abnormality prompt message for column loading is generated.

4. The automated separation and purification system for radionuclides according to claim 3, characterized in that, When the purification control module generates the elution feature vector, it includes: Peak detection was performed on the radiation data during the elution stage, and the peak intensity, rise time, and half-width at half-maximum parameters were extracted. The peak intensity, rise time, and half-peak width parameters are combined to form the elution feature vector.

5. The automated separation and purification system for radionuclides according to claim 4, characterized in that, The purification control module, when generating output instructions, includes: The cosine similarity between the elution feature vector and the target nuclide elution feature vector is calculated to generate a similarity value.

6. The automated separation and purification system for radionuclides according to claim 5, characterized in that, The purification control module, when generating output instructions, also includes: When the similarity value is greater than the similarity threshold, a product output instruction is generated; When the similarity value is less than or equal to the similarity threshold, an elution anomaly warning message is generated.

7. The automated separation and purification system for radionuclides according to claim 1, characterized in that, When the initialization module acquires the valve position signal, initializes and generates the valve position zero-point reference data, it includes: Initialize the rotation of the multi-channel switching valve group and synchronously acquire the timing pulse signal sequence; Based on the time-series pulse signal sequence, the feature pulse corresponding to the initial positioning identifier is identified, and the valve position code value is recorded each time the feature pulse is identified, generating a valve position position dataset; outlier removal is performed on the valve position position dataset, and valid position data is retained; the mean and standard deviation of the valid position data are determined.

8. The automated separation and purification system for radionuclides according to claim 7, characterized in that, When the initialization module acquires the valve position signal, initializes and generates the valve position zero-point reference data, it also includes: When the standard deviation is less than the tolerance threshold, the mean value is determined as the valve position zero-point reference data; when the standard deviation is greater than or equal to the tolerance threshold, the rotation is re-initialized and the valve position dataset is updated.

9. The automated separation and purification system for radionuclides according to claim 1, characterized in that, When the instruction generation module generates process parameters, it includes: The nuclide type is matched with the index field in the separation and purification process parameter set to generate structured process parameter data. The structured process parameter data includes reagent dosing time, adsorption time, rinsing time, elution time, and valve position switching parameters at each stage.

10. The automated separation and purification system for radionuclides according to claim 9, characterized in that, When the instruction generation module generates valve group control instructions and pump operation instructions, it includes: The valve position zero-point reference data is time-aligned with the valve position switching parameters in the process parameters to generate a valve group control command sequence; a pump operation command sequence is generated based on the pump operation parameters in the process parameters.

Images