CN-121972003-A - Filter membrane degerming filter equipment and system

CN121972003ACN 121972003 ACN121972003 ACN 121972003ACN-121972003-A

Abstract

The invention relates to the technical field of terminal sterilization and filtration in a radioactive drug preparation process, in particular to a filter membrane sterilization and filtration device and a filter membrane sterilization and filtration system. The method comprises the steps of collecting liquid medicine radioactivity data and filter membrane structure strain data by a state sensing module and generating real-time data flow, constructing a filter membrane digital twin body by a twin modeling module, outputting accumulated radiation damage prediction results through data assimilation, generating a filtering parameter optimization sequence by a prediction decision module through a deep reinforcement learning intelligent body, controlling and adjusting liquid medicine flow rate by a control execution module through fuzzy self-adaptive PID (proportion integration differentiation) control and feedforward compensation, verifying that an integrity test is carried out by a feedback module, and returning the data to the twin modeling module to realize model parameter correction. Each module forms a closed-loop control system through a high-speed data bus, so that damage of radiation to a filter membrane material is effectively inhibited, and the structural integrity and sterility assurance level of the membrane in the filtering process are maintained.

Inventors

XIE XIONGJIE
MAO SHENG
Gan Renfeng
ZHANG FAGUANG

Assignees

福州原子高科医药有限公司

Dates

Publication Date: 20260505
Application Date: 20260126

Claims (10)

1. The filter membrane sterilization and filtration device is characterized by comprising a state sensing module, a twin modeling module, a prediction decision module, a control execution module and a verification feedback module; the state sensing module is used for collecting the radioactivity data of the liquid medicine flowing through the sterilizing filter membrane and the structural strain data of the sterilizing filter membrane to generate a real-time data stream; The twin modeling module receives the real-time data stream, constructs a digital twin body of the sterilizing filter membrane, dynamically updates the digital twin body by assimilating the structural strain data, and outputs a cumulative radiation damage prediction result of the sterilizing filter membrane; The prediction decision module comprises a decision agent, and is used for receiving the accumulated radiation damage prediction result and the radioactivity data, and generating a filtering parameter optimization sequence through the decision agent, wherein the filtering parameter optimization sequence comprises a flow rate set point instruction; the control execution module receives the flow rate set point instruction, receives the accumulated radiation damage prediction result, and adjusts the operation parameters of the precision metering pump to control the liquid medicine filtering flow rate; And the verification feedback module applies test conditions to the degerming filter membrane after the filtration is finished to obtain integrity test data, and returns the integrity test data to the twin modeling module to correct the model parameters of the digital twin.
2. The filter membrane sterilization and filtration device according to claim 1, wherein the state sensing module comprises a radioactivity sensor, a strain sensor network and a signal processing unit; The radioactive sensor captures radioactive particles and converts the radioactive particles into photon signals, the photon signals generate electric pulse signals through the photoelectric conversion unit, the electric pulse signals reject background noise pulses through the discriminator and enter the counter, and the counter counts and outputs activity time sequence electric signals in a fixed time window; The strain sensor network is composed of optical fibers written into a grating array, the optical fibers are packaged in the sterilization filter membrane in a preset topology, light emitted by a light source is reflected by the grating array, and formed optical signals are received and analyzed by a spectrum demodulator to obtain wavelength drift amount; The signal processing unit synchronizes the activity time sequence electric signal with the wavelength drift amount, converts the wavelength drift amount into structural strain data according to a preset calibration coefficient, adds a time stamp to the activity time sequence electric signal and the structural strain data, and packages the activity time sequence electric signal and the structural strain data to generate the real-time data stream.
3. The filter membrane sterilization filter device according to claim 1, wherein the twin modeling module performs the following data processing: Receiving the real-time data stream, extracting radioactivity data, and calculating the radiant energy deposition rate of each grid unit of the sterilizing filter membrane by adopting a finite element analysis method based on the geometry and material intrinsic parameters of the sterilizing filter membrane; inputting the calculated energy deposition rate into a polymer radiation damage kinetic equation, solving a time-varying field of the porosity and the average pore diameter of the degerming filter membrane through numerical integration, and outputting a predicted strain field; comparing the predicted strain field with an actually measured structure strain field extracted from the real-time data stream, adopting an aggregate Kalman filtering algorithm, reversely optimizing and adjusting local damage rate parameters in the polymer radiation damage kinetic equation by taking the actually measured structure strain field as an observation value, and outputting the updated accumulated radiation damage prediction result based on the optimized parameters.
4. The filter membrane sterilization and filtration device according to claim 3, wherein the decision agent in the predictive decision module is a deep reinforcement learning agent; the state vector of the deep reinforcement learning agent consists of a characteristic value of the accumulated radiation damage prediction result, a history and predicted activity attenuation curve, the volume of the residual liquid medicine and accumulated filtering time; The deep reinforcement learning agent finishes pre-training in an offline simulation environment by taking maximized filter membrane integrity maintenance probability as a rewarding target, and receives the state vector and outputs a preliminary flow velocity adjustment suggestion action during online operation; the prediction decision module further comprises a model prediction controller, the model prediction controller receives the preliminary flow velocity adjustment recommended action, in the current control period, the filter membrane accumulated damage minimization in the future limited time domain is used as an objective function, a group of future flow velocity sequences are encoded and iteratively optimized through a genetic algorithm, and the future flow velocity sequence with the lowest total damage cost obtained by optimizing is output as the filter parameter optimization sequence.
5. The filter membrane sterilization and filtration device according to claim 1, wherein the control execution module comprises a fuzzy adaptive PID controller and a feedforward compensator; The fuzzy self-adaptive PID controller receives the flow speed set point instruction and an actual flow speed value fed back by the electromagnetic flowmeter, calculates deviation and deviation change rate, dynamically adjusts PID parameters on line through a fuzzy rule table, and outputs basic control quantity to a driver of the precise metering pump; the feedforward compensator receives the accumulated radiation damage prediction result output by the twin modeling module, performs differential operation on the accumulated radiation damage prediction result to obtain a damage change rate, and multiplies the damage change rate by a preset feedforward gain coefficient to generate feedforward compensation quantity; the basic control quantity and the feedforward compensation quantity are overlapped in an adder, and an overlapped signal output by the adder is used as a final control signal to drive the precise metering pump to act so as to control the filtering flow rate of the liquid medicine.
6. The filter membrane sterilization and filtration device according to claim 5, wherein the control execution module is further connected with a pressure transmitter, and the pressure transmitter monitors the pressure difference between the upstream and downstream of the sterilization filter membrane and outputs pressure difference data; A pressure protection sub-loop is arranged in the control execution module, and the pressure protection sub-loop receives the pressure difference data and compares the pressure difference data with a preset safety threshold value; When the pressure difference data exceeds the preset safety threshold, the pressure protection sub-loop generates a limiting signal, the limiting signal is inserted into the output of the fuzzy self-adaptive PID controller, a control quantity softening logic is activated, and the change rate of the control signal output to the precise metering pump is limited.
7. The filter membrane sterilization and filtration device according to claim 1, wherein when the filtration is finished, the verification feedback module switches the pneumatic valve to enable the clean compressed nitrogen source to be communicated to the upstream cavity of the sterilization filter membrane, and the pressure is raised to a preset test pressure and maintained through the pressure regulating valve; The verification feedback module records a pressure decay time sequence curve through a high sampling rate pressure sensor, and synchronously records gas diffusion flow time sequence data through a downstream mass flowmeter; The verification feedback module carries out low-pass filtering on the pressure decay time sequence curve and the gas diffusion flow time sequence data to remove noise, then adopts an unsteady diffusion physical model to fit the filtered pressure decay time sequence curve and the gas diffusion flow time sequence data, calculates to obtain a standardized diffusion current value, and uses the standardized diffusion current value as a core integrity index; and the verification feedback module packages the time sequence data of the whole process of the filtering task and the core integrity index into a batch feedback data set, and transmits the batch feedback data set to the database of the twin modeling module.
8. The filter membrane sterilization and filtration device according to claim 7, wherein the twinning modeling module extracts a core integrity index in the batch feedback data set as a real state observation value of the sterilization filter membrane at the end of the present filtration cycle after receiving the batch feedback data set; the twinning modeling module compares the real state observed value with a series of predicted damage state values corresponding to the time generated by the digital twins in the current filtering period; the twin modeling module reversely adjusts key model coefficients affecting long-term prediction accuracy in the polymer radiation damage kinetic equation through a maximum likelihood estimation algorithm; and after the twin modeling module completes coefficient adjustment, optimizing the initial state of the digital twin body entering the next working period.
9. The filter membrane sterilization filter device according to claim 8, wherein the state sensing module, the twin modeling module, the predictive decision module, the control execution module and the verification feedback module are connected by a high-speed data bus; the real-time data stream output by the state sensing module is input to the twin modeling module; The accumulated radiation damage prediction result output by the twin modeling module is input to the prediction decision module; The filtering parameter optimization sequence output by the prediction decision module is input to the control execution module; The control signal output by the control execution module drives the precise metering pump; The integrity test data output by the verification feedback module is input to the twin modeling module; The twin modeling module receives the integrity test data and corrects model parameters to form a closed loop data stream.
10. The filter membrane sterilization and filtration system is applied to the filter membrane sterilization and filtration device according to any one of claims 1-9, and is characterized by comprising a state sensing module, a twin modeling module, a prediction decision module, a control execution module and a verification feedback module; The state sensing module is configured to collect the radioactivity data of the liquid medicine flowing through the sterilizing filter membrane and the structural strain data of the sterilizing filter membrane and generate a real-time data stream; the twin modeling module is configured to receive the real-time data stream, construct a digital twin body of the sterilizing filter membrane, dynamically update the digital twin body by assimilating the structural strain data, and output a cumulative radiation damage prediction result of the sterilizing filter membrane; the prediction decision module is configured to comprise a decision-making agent, receive the accumulated radiation damage prediction result and the radioactivity data, and generate a filtering parameter optimization sequence through the decision-making agent, wherein the filtering parameter optimization sequence comprises a flow rate set point instruction; the control execution module is configured to receive the flow rate set point instruction and the accumulated radiation damage prediction result, and adjust the operation parameters of the precision metering pump to control the liquid medicine filtering flow rate; And the verification feedback module is configured to apply test conditions to the degerming filter membrane after the filtering is finished to obtain integrity test data, and transmit the integrity test data back to the twin modeling module so as to correct the model parameters of the digital twin body.

Description

Filter membrane degerming filter equipment and system Technical Field The invention relates to the technical field of terminal sterilization and filtration in a radioactive drug preparation process, in particular to a filter membrane sterilization and filtration device and a filter membrane sterilization and filtration system. Background The conventional terminal sterilization and filtration technology for the electropositive radiopharmaceuticals refers to a process for removing microorganisms by adopting a microporous filter membrane in the final filling stage in the production process of the electropositive radiopharmaceuticals, wherein the electropositive radiopharmaceuticals generally comprise short half-life radionuclides, the rapid decay characteristic of the electropositive radiopharmaceuticals requires that the preparation process must be completed within a strictly controlled time window, and meanwhile, the sterility of the radiopharmaceuticals is maintained, so that bacteria and particulate matters are efficiently removed by the terminal sterilization and filtration through a physical interception principle, so that the standard of the aseptic medicines is met, and the terminal sterilization and filtration technology is suitable for rapid and closed production line operation, thereby guaranteeing the microbial safety of the medicines before clinical use. The existing terminal sterilization and filtration technology of positive electrode radiopharmaceuticals has the technical pain that a filter membrane for sterilization is generally made of a polymer material, and the radiant energy continuously released by positron nuclides such as fluorine 18 in the decay process can directly act on the molecular structure of the filter membrane material. The irradiation energy may cause the polymer main chain to undergo fracture or crosslinking reaction, and microscopically shows increased brittleness of the material and irreversible change in pore morphology. Since the drug filtration operation must be completed immediately at higher radioactivity, the membrane continues to accumulate absorbed radiation doses during contact, potentially causing a recessive impairment of membrane integrity. For example, in the production of fluorine 18 labeled deoxyglucose injection, radiation damage causes submicron defects to form on or in the membrane surface as the high activity liquid drug flows through the filter membrane. Such defects are not necessarily immediately detectable in conventional bubble point tests, but may be a way for microorganisms to penetrate, ultimately affecting the sterility assurance level of the drug. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a filter membrane sterilization and filtration device and a system, which solve the technical problem of potential failure of membrane integrity in the sterilization and filtration process caused by damage of a microstructure of a filter membrane polymer material due to continuous radiation irradiation generated by radionuclide decay. In order to solve the technical problems, the specific technical scheme of the invention is as follows: in a first aspect, the filter membrane sterilization and filtration device provided by the invention comprises a state sensing module, a twin modeling module, a prediction decision module, a control execution module and a verification feedback module; the state sensing module is used for collecting the radioactivity data of the liquid medicine flowing through the sterilizing filter membrane and the structural strain data of the sterilizing filter membrane to generate a real-time data stream; The twin modeling module receives the real-time data stream, constructs a digital twin body of the sterilizing filter membrane, dynamically updates the digital twin body by assimilating the structural strain data, and outputs a cumulative radiation damage prediction result of the sterilizing filter membrane; The prediction decision module comprises a decision agent, and is used for receiving the accumulated radiation damage prediction result and the radioactivity data, and generating a filtering parameter optimization sequence through the decision agent, wherein the filtering parameter optimization sequence comprises a flow rate set point instruction; the control execution module receives the flow rate set point instruction, receives the accumulated radiation damage prediction result, and adjusts the operation parameters of the precision metering pump to control the liquid medicine filtering flow rate; And the verification feedback module applies test conditions to the degerming filter membrane after the filtration is finished to obtain integrity test data, and returns the integrity test data to the twin modeling module to correct the model parameters of the digital twin. Further, the filter membrane sterilization and filtration device provided by the invention comprises a state sensing module, a