A method and apparatus for integrated control of blow-filling-seal of liquid medicine plastic bottles

Abstract

The present application relates to the technical field of liquid medicine packaging production control, and discloses a blow-filling-seal integrated production control method and equipment for plastic bottles of liquid medicine, which is executed by a central control system and comprises the following steps: controlling an extrusion unit to extrude a parison, which is cut and blow molded into a sterile container by a blow-filling-seal main machine; injecting a predetermined dose of liquid medicine into the container and performing heat sealing; and conveying the finished product to an online monitoring unit for detection and judgment of whether it is qualified or not. The central control system collects process parameters and quality attribute data in real time, establishes a mapping relationship, and performs online prediction on the quality attribute of the subsequent finished product; when the prediction result deviates from the quality target interval, the target optimal process parameter adjustment amount is solved within the preset safety constraint range, and is issued to the corresponding equipment controller for closed-loop adaptive adjustment. The present application changes post-removal into pre-prediction intervention, effectively reduces the system waste rate, and improves the consistency of finished product quality and the execution precision of the production process.

Description

Technical Field

[0001] This invention relates to the field of pharmaceutical packaging production control technology, specifically to an integrated production control method and equipment for blow-fill-seal pharmaceutical plastic bottles. Background Technology

[0002] Blow-fill-seal technology is widely used in the production of sterile pharmaceutical packaging, where plastic bottles are extruded, filled, and sealed in a continuous circulating system.

[0003] Current blow-fill-seal production control systems mostly adopt fixed process parameter operation modes, and quality control usually relies on post-production sampling or simple end-of-line rejection. They cannot detect and compensate for the impact of material characteristic fluctuations or external environmental interference in a timely manner during operation. When unqualified products are found at the end, continuous scrap has already been generated, resulting in a high overall scrap rate and difficulty in ensuring the consistency of finished product quality.

[0004] Blow-fill-seal involves multiple complex sub-processes. Existing equipment lacks independent dynamic closed-loop adjustment mechanisms at these critical nodes. During the extrusion and blow molding stages, the screw speed and blowing pressure are mostly set to fixed outputs, making it difficult to adapt to instantaneous changes in melt pressure. This can easily lead to unstable tubular preform dimensions and uneven wall thickness of the formed container. In subsequent processes, the control of filling dosage and heat sealing temperature and pressure also tends to be static, lacking real-time feedback compensation, which affects the reliability of the encapsulation and reduces the overall process execution accuracy of the equipment.

[0005] In the online quality inspection of finished products, existing monitoring systems usually adopt single-dimensional detection methods, such as relying solely on manual random sampling or basic mechanical flaw detection. This cannot simultaneously assess both the integrity of the seal and the consistency of the appearance. In particular, the detection rate of minor leaks and subtle surface defects commonly found in pharmaceutical plastic bottles is low, increasing the risk of substandard products entering the circulation process and making it difficult to meet the stringent requirements for sterility and safety in pharmaceutical packaging. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an integrated production control method and equipment for blow-fill-seal plastic bottles for pharmaceutical liquids. This solves the problems of high scrap rates, low process execution accuracy, and low detection rate of minor defects caused by the lack of pre-prediction intervention and dynamic closed-loop control of each sub-process in existing blow-fill-seal production.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an integrated production control method and equipment for blow-fill-seal of pharmaceutical liquid plastic bottles, comprising the following steps:

[0008] The extrusion unit is controlled to extrude a continuous tubular preform, the blow-fill-seal main unit is controlled to close the mold to cut the tubular preform, and sterile compressed gas is introduced to blow-form the preform into a sterile container.

[0009] A predetermined dose of medicine solution, constrained by product specifications, is injected into the sterile container using an aseptic filling system. The neck of the sterile container is then sealed using a heat-sealing mechanism to obtain the finished container.

[0010] The finished container is removed from the mold cavity and conveyed to the online quality monitoring unit for non-contact testing. The central control system determines whether it is qualified or not based on the test data. The qualified finished container is conveyed backward and the unqualified finished container is physically removed.

[0011] The central control system collects the process parameters of the above modules and the key quality attribute data obtained by detection in real time, establishes a mapping relationship between the two, and makes online prediction of the quality attributes of the subsequently generated finished containers based on the mapping relationship.

[0012] When the prediction result deviates from the preset quality target range representing the boundary of qualified product attributes, the optimal process parameter adjustment amount is solved within the preset safety constraints limited by equipment operating limits and process robustness, and then sent to the equipment controller of the corresponding module for closed-loop adaptive adjustment.

[0013] Preferably, the specific steps for the central control system to establish the mapping relationship and perform online prediction include: using a timestamp synchronization mechanism and combining position sensors to compensate for delay changes during material conveying, synchronizing and associating the real-time collected process parameters with the key quality attribute data of each corresponding finished product container; training and verifying based on the data obtained from the experimental design, establishing a process quality prediction model composed of mathematical functions, wherein the process quality prediction model receives the current process parameters online and outputs the predicted key quality attribute vector.

[0014] Preferably, the step of solving for the target optimal process parameter adjustment amount within the preset safety constraints limited by equipment operating limits and process robustness includes: comparing the predicted key quality attribute vector with the preset quality target range; when a deviation trend is detected, actively triggering optimization control calculation, using preset upper and lower limits of parameter adjustment as constraints, superimposing the process parameter adjustment vector to be solved on the current process parameter setpoint vector using the process quality prediction model to calculate the predicted value vector, and taking the minimization of the norm of the overall deviation between the predicted value vector and the preset quality attribute target center value vector as the optimization objective to solve for the target optimal process parameter adjustment amount; wherein, the upper and lower limits of parameter adjustment constitute the specific numerical boundaries of the preset safety constraint range, and are jointly determined by the robust process range determined in the process development stage, the physical performance limits of materials, and the mechanical and electrical safety limits of equipment.

[0015] Preferably, the central control system executes the distribution and closed-loop adaptive adjustment through a dual closed-loop structure, specifically including: a first closed loop, which, based on the deviation between the predicted key quality attribute vector and the target center value vector of the quality attribute, distributes and executes the target optimal process parameter adjustment to the equipment controller in real time; and a second closed loop, which, after the adjustment is executed, continuously collects the actual key quality attribute data output by the online quality monitoring unit, compares the actual key quality attribute data with the predicted key quality attribute vector, and triggers relearning and correction of the parameters of the process quality prediction model itself when a systematic deviation occurs.

[0016] Preferably, during the extrusion of the continuous tubular preform, the central control system dynamically adjusts the dimensional stability of the tubular preform by combining feedforward and feedback: the central control system adds the preset base speed and speed adjustment amount according to the product specifications and material grade to calculate the instantaneous target speed of the extruder screw and outputs control; the speed adjustment amount is calculated by multiplying the difference between the preset melt pressure target setting value and the real-time pressure value measured by the melt pressure sensor installed near the die head by the proportional gain coefficient of the pressure-speed adjustment, and the preset melt pressure target setting value is limited to a range that ensures uniform preform extrusion and prevents melt fracture.

[0017] Preferably, during the process of introducing sterile compressed gas to blow-mold the preform into a sterile container, multi-stage pressure control is adopted: the central control system sends continuously changing control signals to the high-speed gas proportional valve, so that the gas pressure injected into the cut preform reproduces a preset pressure-time curve. The curve successively includes a pre-blowing stage with lower pressure and a main blowing stage with higher pressure. The preset pressure-time curve is obtained by establishing a correlation model through experimental data optimization, using curve characteristics as input variables and wall thickness uniformity as output response. In the pressure holding and cooling stage after molding, real-time data is acquired by collecting temperature and flow sensors in the cooling pipeline, and the power of the cooling unit or the opening of the control valve is actively adjusted in a closed loop.

[0018] Preferably, during the process of injecting a predetermined dose of medication according to product specifications into the aseptic container using the aseptic filling system and sealing the neck of the aseptic container using a heat-sealing mechanism, the central control system performs closed-loop control on the filling dose and the sealing temperature, sealing pressure, and sealing time: It uses real-time pressure values ​​measured by pressure sensors in the pipeline for feedback adjustment to stabilize the medication pressure before the terminal filter at a preset pressure value, and sends an opening pulse signal to the filling valve to control the valve to open for a preset filling time, ensuring that the medication injected into the aseptic container reaches the predetermined dose; the preset pressure value is limited to a range that ensures filling speed and avoids medication foaming, and the preset filling time is determined by a trial filling. The weight of the obtained product is converted into volume and then corrected and calibrated. Real-time data from the internal temperature sensor of the heat sealing mechanism is collected, and the electrical power supplied to the heating element inside the heat sealing mechanism is adjusted according to the deviation between the real-time data and the preset temperature setting value. The output force of the actuator is adjusted in a closed loop through feedback from the pressure sensor in the cylinder air supply line or the current or torque feedback of the servo motor itself. After the pressure reaches the preset pressure setting value, the internal timer is started to maintain the clamping state under the action of the output force of the actuator until the preset duration ends. The preset temperature setting value, the preset pressure setting value, and the preset duration are all taken from the process parameter constraints determined by experimental design methods and verified to be qualified.

[0019] Preferably, when removing the finished product container from the mold cavity, the process includes the steps of finished product demolding and waste separation: the central control system controls the ejection system integrated with the mold to apply mechanical force to push the finished product container out of the mold cavity, while simultaneously controlling the gripper of the picking robot to grasp the finished product container; under stable gripping conditions, the picking robot moves along a preset trajectory that defines a spatial anti-interference coordinate range, and uses relative motion to generate stress concentration at the pre-designed weak connection point between the finished product container and the waste, causing fracture and separating the finished product container from the waste.

[0020] Preferably, in the process of the central control system determining whether a product is qualified based on the detection data, a discrimination logic function is executed to achieve the following: real-time acquisition of a multi-dimensional set of measurement values, including the impedance value measured by the high-voltage discharge system and the surface feature quantification index extracted by the machine vision system; inputting the multi-dimensional set of measurement values ​​into the discrimination logic function; comparing the multi-dimensional set of measurement values ​​with a preset set of qualification standards; and outputting a binary result indicating that the product is qualified when all measurement values ​​meet the corresponding acceptable standard or threshold in the preset set of qualification standards. The threshold range in the preset set of qualification standards is set by repeatedly testing known qualified products and samples containing typical boundary defects, combined with statistical analysis.

[0021] A second aspect of the present invention provides an integrated production control device for blow-fill-seal plastic bottles for pharmaceutical liquids, comprising: an extrusion unit, a blow-fill-seal main unit, an online quality monitoring unit, and a central control system;

[0022] The central control system runs a preset production control program and communicates bidirectionally with the extrusion unit, blow-fill-seal host and online quality monitoring unit via an industrial data bus.

[0023] This invention provides an integrated production control method and equipment for blow-fill-seal manufacturing of pharmaceutical liquid plastic bottles. It has the following beneficial effects:

[0024] 1. This invention establishes a mapping relationship between process parameters and key quality attributes through a central control system and performs online prediction. When the prediction result deviates from the target range, it actively solves for the optimal process parameter adjustment within the safety constraints and sends it to the equipment for closed-loop adaptive adjustment. This transforms the traditional post-event sampling and rejection into pre-event prediction and intervention, which can promptly compensate for the impact of material fluctuations or environmental interference in production, reduce the scrap rate of the system, and improve the consistency of finished product quality.

[0025] 2. This invention adopts an independent dynamic closed-loop control strategy for each sub-process in the blow-fill-seal continuous cycle. By adjusting the screw speed through a combination of feedforward and feedback in the extrusion stage, and by using multi-segment pressure control to reproduce the set curve in the blow molding stage, as well as providing real-time feedback compensation for the filling dosage and the temperature and pressure of the heat seal, the dimensional stability of the tubular preform, the uniformity of the container wall thickness, and the reliability of the sealing are ensured, thereby improving the overall process execution accuracy of the production equipment.

[0026] 3. This invention introduces non-contact online quality monitoring after the finished product is removed from the mold cavity. By collecting the impedance value of the high-voltage discharge system and the quantitative indicators of surface features extracted by the machine vision system in real time, the multi-dimensional measurement values ​​are compared with the qualified standard threshold set by statistical analysis to determine the quality. This avoids the limitations of a single detection method, improves the detection rate of minor leaks and appearance defects, and ensures the sterility and safety standards of the finished drug liquid plastic bottles. Attached Figure Description

[0027] Figure 1 This is an overall flow chart of the blow-fill-seal process of the present invention;

[0028] Figure 2 This is a schematic diagram of the intelligent blow-fill-seal production equipment of the present invention.

[0029] The components include: 1. Extrusion unit; 2. Blow-fill-seal main unit; 3. Central control system. Detailed Implementation

[0030] The technical solutions in 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.

[0031] See attached document Figure 1 The present invention provides an integrated production control device for blow-fill-seal of pharmaceutical plastic bottles, which is physically composed of multiple functional modular units and realizes data communication and collaborative work through a central control system 3.

[0032] The equipment includes: an extrusion unit 1, a blow-fill-seal main unit 2, an online quality monitoring unit, and a central control system 3. These units are integrated into a closed equipment frame, and its core operating area is in a Class A clean environment maintained by a sterile air supply system.

[0033] Extrusion unit 1 is used to heat and melt solid medical-grade polymer particles and extrude them through a die to form a continuous tubular preform.

[0034] The blow-fill-seal unit 2 is the unit that performs the core process. It receives the preform from the extrusion unit 1 and, within its internal closed aseptic mold system, sequentially completes the blow molding of the container, the aseptic filling of the liquid medicine, and the melt sealing of the container.

[0035] An online quality monitoring unit is installed at the finished product outlet of the blow-fill-seal main unit 2 to conduct continuous and comprehensive quality inspections on each sealed plastic bottle of medicine.

[0036] The central control system 3 serves as the control core of the entire equipment. It performs two-way data communication with the extrusion unit 1, blow-fill-tube main machine, on-line quality monitoring unit and other auxiliary systems (such as sterile air supply system) via an industrial data bus. A preset production control program runs inside the central control system 3, which is responsible for coordinating the operation timing of each unit and performing closed-loop feedback adjustment based on real-time collected process parameters.

[0037] Referring to the drawing Figure 2 , the present invention provides an integrated blow-fill-seal production control method for medicinal liquid plastic bottles, which is scheduled and executed by the central control system 3 in one continuous production cycle, comprising the following steps:

[0038] S10: Feeding and plasticizing. The central control system 3 controls the heating system in the extrusion unit 1 to melt and plasticize polymer particles, and simultaneously drives the screw of the extruder to stably extrude a tubular parison from the molten material through an annular die.

[0039] S20: Parison cutting and molding. When the parison sags to a predetermined length, the mold in the blow-fill-seal main machine 2 closes and cuts the parison synchronously. Subsequently, the core rod mechanism introduces sterile compressed gas to blow and expand the parison in the mold cavity to form a sterile container with a set shape.

[0040] S30: Aseptic filling. After the container is molded, the central control system 3 immediately instructs the aseptic filling system in the blow-fill-seal main machine 2 to start, and precisely injects a predetermined dose of medicinal liquid into the container through an independent channel integrated with the core rod.

[0041] S40: Melt sealing. After filling is completed, a heat sealing mechanism in the blow-fill-seal main machine 2 performs hot-press sealing on the neck of the container to form a complete and airtight package.

[0042] S50: Finished product demolding and conveying. After sealing is completed, the mold is opened, the finished product is removed from the mold cavity, and sent to the on-line quality monitoring unit by a conveying mechanism.

[0043] S60: On-line detection and decision-making. The on-line quality monitoring unit detects the sealing integrity and weight of each finished product conveyed. Detection data are sent to the central control system 3 in real time. The central control system 3 judges whether the product is qualified or unqualified according to preset quality standards.

[0044] S70: Product processing. For products judged as qualified, they are allowed to continue to the subsequent packaging process. For products judged as unqualified, the central control system 3 issues an instruction to drive a rejecting device in the on-line quality monitoring unit to remove the unqualified product from the production line.

[0045] In the method of this invention, the feeding and plasticizing step S10 aims to stably process medical-grade polymer particles into tubular preforms with uniform dimensions and suitable temperature under controlled sterile conditions, providing a qualified substrate for subsequent processes. This step can be broken down into the following sequentially coordinated sub-steps.

[0046] S101, Multi-segment Temperature Closed-Loop Control. To achieve precise thermal management of polymer particles from solid to molten state, the barrel of extrusion unit 1 is divided into multiple independent temperature zones along the material flow direction. Each temperature zone is equipped with an independent heating element and temperature sensor, such as a sheathed thermocouple. The central control system 3 implements independent closed-loop control for each temperature zone. The basic principle is that the central control system 3 collects the real-time process temperature measured by the temperature sensors in each temperature zone and compares it with the preset target temperature for that zone. It then uses a PID (Proportional-Integral-Derivative) control algorithm to calculate the control output signal to adjust the power of the heating element. In this algorithm, the proportional term responds to the current temperature deviation, the integral term eliminates long-term steady-state deviations, and the derivative term predicts the trend of deviation changes to prevent overshoot. The goal of this control process is to stabilize the temperature of each zone at the set value, ensuring that the polymer undergoes a sufficient but not excessive thermal process, guaranteeing complete plasticization while simultaneously sterilizing the material itself using high temperatures.

[0047] A temperature zone The PID control output model can be expressed as:

[0048] ;

[0049] in, For the controller in time Temperature range The output signal of the heating power; Temperature zone In time The temperature deviation, whose value is the preset target temperature. With real-time process temperature difference; , , These are the proportional, integral, and derivative gain coefficients of the PID controller.

[0050] The implementation of PID control algorithms and the tuning of their parameters can be accomplished by those skilled in the art using conventional methods, which are well-known techniques in the field and will not be elaborated upon here. It should be noted that... , , The specific values ​​of these three gain coefficients were determined experimentally during the equipment commissioning phase, using empirical methods or engineering tuning methods such as Ziegler-Nichols, for specific polymer materials and production rates, in order to achieve optimal temperature control response characteristics.

[0051] S102, Parison Dimensional Stability Control. To ensure that the extruded tubular parison has a uniform wall thickness and outer diameter along its length, the central control system 3 dynamically adjusts the screw speed of the extruder. This adjustment employs a control strategy combining feedforward and feedback. The feedforward control section presets a base screw speed based on product specifications and material grade. The feedback control section operates on the principle that the extrusion volume is directly related to the melt pressure at the die; increased pressure leads to increased extrusion volume, and vice versa. Therefore, a melt pressure sensor is installed near the extruder die to monitor the pressure of the molten material before extrusion in real time. The central control system 3 uses this pressure signal to compare with the set pressure target value, calculates the speed adjustment amount, and adds it to the base speed, controlling the speed of the extruder main motor via a frequency converter.

[0052] The screw speed adjustment model can be expressed as:

[0053] ;

[0054] in, For the screw in time The instantaneous target rotational speed; The preset base speed; This is the proportional gain coefficient for pressure and speed regulation; The target setpoint for melt pressure; For melt pressure sensor in time The measured real-time pressure value.

[0055] In this way, the system can actively compensate for changes in melt flowability caused by batch variations in raw materials or slight changes in ambient temperature, thereby maintaining a stable extrusion volume and ensuring the consistency of parison dimensions. Similarly, the proportional gain coefficient... The value of needs to be determined through process verification experiments in order to avoid system oscillation while ensuring adjustment sensitivity.

[0056] S103, Aseptic Protection During Extrusion. To prevent contamination of the tubular parison's surface by microorganisms or suspended particles from the environment during its vertical descent from the extrusion die, a Class A unidirectional flow clean air curtain system is installed directly below the die. This system continuously blows clean air filtered by a high-efficiency particulate air (HEPA) filter downwards, creating a dynamic positive pressure environment with a Class A cleanliness level around the parison. This positive pressure environment effectively isolates external unclean air, ensuring that the parison's inner and outer surfaces remain sterile before entering the mold of the blow-fill-seal unit 2. The central control system 3 continuously monitors the airflow velocity of this clean air curtain and its pressure difference relative to the background environment using wind speed sensors and differential pressure gauges. This ensures that these key environmental parameters remain within the acceptable ranges required by process validation and GMP (Good Manufacturing Practice) requirements, thereby guaranteeing the continuous effectiveness of the aseptic protection barrier.

[0057] In the method of the present invention, the implementation of the preform cutting and forming step S20 aims to precisely cut the tubular preform conveyed from the extrusion unit 1 within a closed sterile mold and blow-form it into a sterile container with a predetermined shape and wall thickness distribution. The specific implementation of this step can be decomposed into the following sequentially coordinated sub-steps.

[0058] S201, Synchronous Cutting and Mold Closing. The mold (usually consisting of two main mold halves) inside the blow-fill-seal main unit 2 closes according to preset production cycle times under the command of the central control system 3. The precise timing of mold closure is a key process parameter, which, combined with the parison extrusion rate set in the preceding step S10, determines the length of the parison being cut. Simultaneously with mold closure, a cutting mechanism linked to the mold (e.g., a hot cutter) performs a transverse cut on the parison, separating it from the continuous parison above. The mold closure and cutting actions are precisely synchronized through the equipment's mechanical cam system or multi-axis servo motor system to ensure a high degree of consistency in the parison length cut in each production cycle.

[0059] S202, multi-stage pressure blow molding. After the mold is fully closed, the upper mandrel mechanism descends, its tip piercing and sealing the open end of the preform. Subsequently, the central control system 3 instructs the sterile gas supply system to inject sterile compressed gas, filtered through a terminal sterilization filter (e.g., a 0.22μm pore size filter), into the sealed preform through channels inside the mandrel. To optimize the wall thickness distribution of the container and avoid localized thinning or thickening, the gas injection pressure is not constant but follows a preset pressure-time curve. From a physical molding principle perspective, this curve typically includes a low-pressure pre-blowing stage, which allows the preform to expand uniformly without excessive tensile stress, forming a centrally symmetrical bubble; and a high-pressure main blow stage, which allows the initially expanded preform to quickly and completely conform to the inner wall of the mold cavity to replicate the fine structure of the mold cavity.

[0060] The central control system 3 precisely controls the gas pressure injected into the preform by sending continuously varying control signals to the high-speed gas proportional valve, thereby reproducing the preset curve. This pressure curve function can be expressed as:

[0061] ;

[0062] in, For blow molding gas in time Instantaneous pressure; As a preset pressure objective function, it defines the complete pressure change path from pre-blowing to main-blowing; This represents the total duration of the blow molding process.

[0063] The pressure objective function The specific form is determined through systematic methods such as Design of Experiments (DoE) during the process development and validation phases. This process uses key features of the function curve (such as the magnitude and duration of pre-blowing pressure, the rate of transition to main blowing pressure, the peak value of main blowing pressure, and the holding time) as input variables, and key quality properties of the final vessel (especially the uniformity of wall thickness in various parts) as output responses. By establishing a correlation model through experimental data, the optimal pressure curve is obtained.

[0064] S203, Cooling and Solidification. After the main blowing stage, the container is pressed tightly against the inner wall of the mold with cooling water channels. The central control system 3 maintains a preset pressure holding and cooling time. During this period, the cooling medium circulating inside the mold (e.g., cooling water) rapidly removes heat from the container, causing it to transform from a molten state to a solid state and solidify. The temperature and flow rate of the mold cooling system are not merely passively monitored, but actively controlled in a closed loop by the central control system 3. The system acquires real-time data through temperature and flow sensors installed in the cooling pipes and compares it with the set target values, thereby adjusting the power of the cooling unit or controlling the opening of the valves to ensure that the mold temperature is maintained within a stable and narrow process window. This ensures that each container receives consistent and sufficient cooling, resulting in stable mechanical strength and dimensions after demolding, without deformation.

[0065] In the method of this invention, the aseptic filling step S30, which follows immediately after the container forming step S20, aims to precisely inject a predetermined dose of sterile medication into the freshly formed sterile container within a very short time. This process is the core link in ensuring the sterility of the final product and the accuracy of its contents. The specific implementation of this step can be broken down into the following sequentially coordinated sub-steps.

[0066] S301, Establishment and Maintenance of the Aseptic Drug Pathway. Before production begins, the entire drug flow path from the external aseptic storage tank to the filling channel inside the core rod mechanism of the blow-fill-seal main unit must undergo in-line cleaning (CIP) and in-line sterilization (SIP) processes. The specific procedures for CIP and SIP can be performed by those skilled in the art using conventional methods; these are well-known techniques and will not be elaborated upon here. During production, the drug solution flows through the aseptic pipeline under the action of the delivery pump, and at its final end before entering the core rod, it passes through a sterile-grade filter (typically a 0.22μm pore size filter element) that has undergone integrity testing to ensure that the drug solution injected into the container is sterile.

[0067] S302, Implementation of High-Precision Dosing Control. To achieve accurate filling dosage, this method can employ various lower-level control methods, one typical and widely used implementation being time-pressure filling. Its physical principle is that for a given flow path and fluid, when the upstream pressure remains constant, the volume of fluid flowing per unit time (i.e., the flow rate) is also constant; therefore, the total filling volume is linearly proportional to the filling time. Specifically, the central control system 3 uses an independent closed-loop control loop, typically employing a pressure sensor installed on the pipeline as feedback, and uses a PID algorithm to control the speed or power of the delivery pump, stabilizing the drug pressure before the terminal filter at a preset value. During filling, the central control system sends a precise opening pulse signal to the high-speed solenoid valve or pneumatic valve integrated with the mandrel, controlling the valve to open for an extremely short time. .

[0068] Filling volume The relationship with these parameters can be approximated as follows:

[0069] ;

[0070] in, This refers to the volume of the liquid medicine dispensed in a single filling. For the pipeline pressure is The steady-state flow rate of the liquid medicine through the filling channel. This flow rate value assumes that factors affecting flow resistance, such as the viscosity and temperature of the liquid medicine, remain stable during the production process. This refers to the opening time of the filling valve.

[0071] In practical applications, the key control parameter is filling time. Pressure setpoint and filling time Both parameters were optimized and calibrated during the process validation phase. The setting needs to be considered comprehensively. It should be high enough to ensure filling speed and stability, but it should also avoid being too high, which could cause liquid shearing, foaming, or impact on the equipment. The calibration is in Once determined, a trial filling is conducted at an initial time, and the weight of the resulting product is accurately measured. The volume is then calculated based on the drug solution density and compared with the target volume. Make corrections until the filling weight stabilizes within the allowable error range of the target value.

[0072] As an alternative implementation, dosage control can also be achieved using a servo motor-driven volumetric pump (e.g., a ceramic pump or a peristaltic pump). In this method, the filling volume is directly proportional to the rotation angle or number of revolutions of the servo motor, and the central control system 3 controls the motor rotation by sending precise pulse commands to the servo driver, thereby achieving high-precision dosage dispensing.

[0073] S303, Timing of Molding and Filling. To minimize the exposure time of the sterile container opening, the central control system 3 strictly manages the execution sequence of the molding and filling steps. At the moment the pressure-holding and cooling phase of step S20 ends, but before the mold opens, the central control system 3 immediately triggers the filling command in S30. At this point, the filling needle (i.e., the tip of the mandrel) is already inside the container, and the medication is directly injected. The entire process is completed within the sealed mold cavity, ensuring a seamless transition from molding to filling and providing further assurance of the product's sterility.

[0074] In the method of this invention, the melt-sealing step S40 is performed immediately after aseptic filling is completed. Its purpose is to use heat and pressure to fuse the open end of the container to form a complete and permanent seal, thereby establishing and maintaining the sterile barrier of the product and the integrity of its contents. The specific implementation of this step can be decomposed into the following sequential and coordinated sub-steps.

[0075] S401, Positioning and Action of the Sealing Mechanism. After the filling step S30 is completed, the mandrel mechanism that performed the filling retracts from the container. Subsequently, the central control system 3 commands a set of independent sealing mechanisms to operate. This sealing mechanism typically includes one or more pairs of heated sealing heads or sealing clamps. Upon command, the sealing heads, driven by a pneumatic or servo motor, rapidly move to a predetermined position, clamping the neck area of ​​the container that has just been filled from the outside.

[0076] S402, Closed-Loop Control of Key Sealing Parameters. The physical principle of fusion sealing lies in applying sufficient heat to melt the inner and outer surfaces of the container neck material, while simultaneously applying sufficient mechanical pressure to force them to adhere tightly and diffuse and entangle at the molecular level. Finally, under maintained pressure, the material cools and solidifies, forming a homogeneous whole. To achieve reliable and repeatable sealing quality, this physical process requires precise closed-loop control of three key process parameters (CPPs)—sealing temperature, sealing pressure, and sealing time.

[0077] Specifically, for sealing temperature, each sealing head incorporates a heating element (e.g., a resistance heating rod) and a temperature sensor (e.g., a thermocouple). The central control system 3 continuously collects real-time data from the temperature sensor and uses PID control algorithms to adjust the electrical power supplied to the heating element, thereby ensuring the surface temperature of the sealing head is precisely stabilized at a preset target value. For sealing pressure, the output force of the actuator (e.g., a cylinder or servo motor) driving the sealing head clamping is precisely controlled. In one embodiment, if pneumatic drive is used, a pressure sensor is installed in the cylinder's air supply line as feedback, and the central control system 3 instructs an electro-proportional valve for adjustment; if servo drive is used, closed-loop force control is achieved through the servo motor's own current or torque feedback. This ensures that the pressure applied to the container neck remains constant at a preset value throughout the sealing time. For sealing time, once the sealing head is clamped in place and the pressure reaches the set value, the central control system 3 starts an internal timer to maintain the clamping state for a preset duration. After the timer expires, the system instructs the sealing mechanism to release and reset.

[0078] S403, Verification and Setting of Process Parameters. The specific set values ​​for the sealing temperature, pressure, and time mentioned above are not arbitrarily selected, but determined through systematic experiments during the process development and verification phase. This process typically employs the Design of Experiments (DoE) method to study the impact of different combinations of these three parameters on seal quality. Seal quality is evaluated through a series of physical tests, such as seal integrity tests (e.g., vacuum decay method or high-voltage discharge method), burst pressure tests, and visual inspections. Through these experiments, a range of process parameters that can stably produce qualified sealing products is determined, i.e., a verified acceptable range. In daily production, the central control system 3 uses specific values ​​within this range as the target set values ​​for each parameter.

[0079] In the method of this invention, the demolding and conveying step S50, as the final step in the blow-fill-seal process cycle, aims to completely remove the finished container, which has been filled and sealed, from the mold and separate it from waste materials generated during the production process (such as flash and excess material), and finally convey it stably to the subsequent process. The specific implementation of this step can be decomposed into the following continuous and coordinated sub-steps.

[0080] S501, Synchronous Mold Opening and Ejection. After the pressure holding and cooling timer in the melt sealing step S40 ends, the central control system 3 instructs the mold opening and closing mechanism (e.g., a hydraulic cylinder or servo motor) to actuate, causing the two halves of the mold to separate smoothly along a predetermined trajectory. After the mold opens, the finished container adheres to one side of the mold cavity due to cooling shrinkage. To achieve reliable demolding, an ejection system is integrated into the mold. The principle is that although the container has solidified and shaped at this time, it still retains a certain residual temperature, and its rigidity has not yet reached its maximum value, so it is relatively sensitive to mechanical stress. After the mold opens to the predetermined position, the central control system 3 immediately drives the ejection system, for example through ejector rods or ejector plates, to apply a precisely controlled mechanical force to push the finished container out of the mold cavity. The timing, speed, and magnitude of the ejection force of the entire mold opening and ejection action are precisely programmed by the central control system 3 to ensure that the force is evenly distributed, which can overcome the shrinkage clamping force without exceeding the yield strength of the container at that temperature, thereby avoiding whitening, stress marks, or permanent deformation of the container when it leaves the mold cavity.

[0081] S502, Finished Product Grabbing and Waste Separation. As the container is ejected, a servo motor-driven or cylinder-driven robotic arm (also called a picking arm) moves to a predetermined position above or to the side of the mold. The gripper at the front of this robotic arm, its shape matching the local contour of the finished product container, can firmly hold a whole plate or a single container. After the gripper secures the finished product, another waste handling mechanism, usually integrated with the mold, operates simultaneously. For example, a pressure plate or lever mechanism fixes or pushes away the waste portion attached to the finished product. Subsequently, the picking robotic arm moves backward or upward along a preset trajectory. This relative motion creates stress concentration at pre-designed weak connection points between the finished product and the waste (usually thinner neck or bottom connecting ribs). When the stress exceeds the tear strength of the material, a controlled fracture occurs, separating the finished product container from the waste. The separated waste then falls into the waste collection channel below under gravity or by a compressed air stream.

[0082] S503, Conveying and Process Confirmation. After waste separation, the picking robot continues its movement, precisely placing the gripped finished product container onto the conveyor belt below. The conveyor belt transports the finished product to subsequent online quality inspection, printing, and packaging stations. To ensure the continuity and stability of the production process, sensors are installed at key nodes in the demolding and conveying paths for process confirmation. For example, photoelectric sensors are installed in the mold cavity or a machine vision system is used to confirm that the finished product from the previous cycle has been completely ejected. Simultaneously, sensors can also be installed at the beginning of the conveyor belt to confirm that the finished product has been successfully placed onto the conveyor belt. The signals from these sensors are sent back to the central control system 3 and serve as necessary permitting conditions or process interlock signals for executing the next production cycle. Only after confirming that the mold has been emptied and the finished product has been successfully transferred will the system initiate the next preform extrusion and mold closing action, thus forming a closed-loop monitoring system that ensures equipment safety and production reliability.

[0083] In the method of this invention, the implementation of the online detection and decision-making step S60 is a process quality control link integrated at the end of the production line. Its purpose is to perform 100% non-destructive testing on each finished container that has just come off the line, determine in real time whether it meets the preset quality standards, and execute corresponding handling actions based on the judgment results, while providing data support for the dynamic adjustment of upstream processes. The specific implementation of this step can be broken down into the following sub-steps.

[0084] S601, Data Acquisition of Multiple Key Quality Attributes. Finished containers flow from a conveyor belt through one or more integrated inspection stations. The technology selection for these inspection stations directly addresses the core requirements for ensuring the safety of sterile pharmaceuticals: Container Seal System Integrity (CCSI) and absence of visible foreign matter. At these stations, online sensor systems targeting different key quality attributes are deployed to acquire physical data for each product in a non-contact, high-speed manner.

[0085] In a typical embodiment, the system integrates a high-voltage discharge leak detection system (HVLD) to detect seal integrity. This system applies a high-frequency, high-voltage electric field to both sides of the container and identifies the presence of micropores or cracks invisible to the naked eye by measuring the impedance or current value passing through the container (especially the sealed area).

[0086] To detect appearance defects, the system integrates a machine vision inspection system. This system consists of multiple industrial cameras, specially designed lighting sources (such as backlights, ring lights, or dome lights), and an image acquisition card. The cameras capture high-resolution images of the container from different angles to identify surface and shape defects such as black spots, foreign objects, material defects, container deformation, and residual burrs.

[0087] In other embodiments, other detection units may be integrated according to product requirements, such as an online weighing unit for detecting filling volume, or a barcode reader for identifying product identity.

[0088] S602, Data Processing and Acceptance Determination. Raw data (e.g., impedance values, image matrix) collected by each sensor system is transmitted in real time to the central control system 3 or a dedicated industrial computer. An internal decision-making algorithm processes and analyzes this data. This decision-making process can be formally represented as a discriminant function:

[0089] ;

[0090] in, The result is usually a binary value (e.g., 1 represents pass, 0 represents fail). This is the set of quality attribute data of the current product that has been measured. ,in It is the first The measured values ​​of each test item; For a pre-defined set of qualified standards, ,in Is with The corresponding acceptable standards or thresholds; To determine the logic function, it measures the value With standard The comparison is performed, and a final qualification conclusion is output. . When processing data from a machine vision system, the function F internally contains a series of image processing algorithms, such as image preprocessing, defect segmentation, feature extraction, etc., and finally converts the image information into something that can be compared with a threshold (such as the maximum allowable defect area) quantifiable indicators for comparison .

[0091] The set of qualification criteria is determined scientifically in the process verification stage through repeated tests on a large number of samples of known qualified products and samples containing typical boundary defects, combined with statistical analysis (such as receiver operating characteristic curve analysis).

[0092] S603, decision execution and physical separation. The central control system 3, based on the determination result immediately generates and sends an instruction to a downstream actuator. If the product is qualified, no instruction is issued, and the product continues to flow to the packaging process along the main conveyor belt. If the product is unqualified, the system will accurately time and trigger a physical separation device when the unqualified product reaches the rejection station. The device may be a blowing nozzle driven by a high-speed air valve, which blows the unqualified product away from the main conveyor belt through an instantaneous strong airflow; it may also be a push rod or swing arm driven by a servo motor or cylinder, which pushes or pushes the unqualified product into the waste channel. To ensure the reliability of the rejection action, a confirmation sensor is usually installed at the entrance of the waste channel to verify that the unqualified product has been successfully removed.

[0093] S604, statistical analysis and feedback of process data. In addition to real-time determination of individual products, the central control system 3 also processes all collected quality data for continuous recording and statistical process control (SPC) analysis. This embodies a quality control strategy from post-inspection rejection to active process prevention. The system monitors the long-term trends of statistical indicators such as the mean and standard deviation of key quality attributes. When it is found that an indicator drifts beyond the control limit or presents an adverse trend, the system can automatically trigger an alarm to prompt the operator to intervene, or in a more advanced implementation, use this trend information as a feedback signal to automatically fine-tune one or more relevant process parameters in upstream steps (such as S10 or S20), so as to realize closed-loop adaptive optimization of the production process. The ultimate goal is to actively correct process drift before the process produces out-of-spec products, and maintain the process capability at a stable and acceptable level.

[0094] In the method of this invention, the product processing step S70 follows immediately after the online detection and decision-making step S60. Its purpose is to perform a series of orderly post-processing steps on all qualified finished products until they are transferred to the final packaging process. This step is a crucial bridge connecting primary and secondary packaging, ensuring the continuity, orderliness, and traceability of the product flow. The specific implementation of this step can be broken down into the following sub-steps.

[0095] S701, Receiving and Buffering of Qualified Products. After step S60, qualified products continue along the main conveyor belt into the product processing area. The principle is to establish a dynamic material storage area to absorb the mismatch between the constant cycle time of the upstream blow-fill-seal machine 2 and the speed fluctuations or brief stops that may occur in downstream equipment (such as cartoning machines), thereby achieving flexible connections between equipment and improving the overall equipment efficiency (OEE) of the entire line. For this purpose, a buffer system is typically installed at the beginning of this area. This buffer system can be a simple cumulative conveyor belt or a buffer device following the first-in, first-out (FIFO) principle, such as a serpentine conveyor belt or a multi-layer vertical buffer tower, to ensure the sequential integrity of product batches.

[0096] S702, Product Individualization and Orientation. Products produced by the blow-fill-seal process are typically connected together in a slab of multiple units (also known as a stack or cassette). Before proceeding to subsequent processes, this slab of products needs to be separated into individual containers; this process is called individualization. Individualization can be accomplished through a dedicated cutting station, which can employ various methods, such as punching with a mold matching the shape of the connecting ribs, or using rotary cutters or ultrasonic cutters for precision cutting. After cutting, the individual containers need to be aligned to a uniform orientation for subsequent printing or boxing. This orientation process can be achieved using mechanical guides, star wheels, or a small robotic gripper guided by machine vision.

[0097] S703, Online Secondary Processing and Verification. Before being conveyed to the final packaging equipment, individual product containers undergo one or more online secondary processing units. A typical processing unit is a coding system used to print variable information, such as production batch number, expiration date, and serial number, at designated locations on each container. This coding system can be non-contact inkjet or laser coding technology. To ensure coding quality and the integrity of product traceability, a preferred implementation is to place an online visual verification system immediately after the coding unit. This system uses an industrial camera to read the newly printed characters or barcodes and compares them with theoretical values ​​issued by the central control system 3 to confirm the correctness and clarity of the printed content in real time. In another embodiment, the secondary processing unit also includes an online labeling machine and its corresponding label presence, position, and information verification system.

[0098] S704, Transfer to subsequent processes. After all necessary processing is completed, qualified and secondary-processed individual products are systematically transferred to the final secondary packaging equipment, such as an automatic cartoning machine. This transfer process is accomplished by a dedicated material handling system. In a simple embodiment, this system can be a conveyor belt synchronized with the cartoning machine's inlet. In more complex production line layouts, this system can be a multi-axis robot. The control of this transfer system communicates with the cartoning machine's control system, achieving seamless material handling between the two independent devices through a series of predefined digital I / O signals or handshake signals such as request and ready based on industrial Ethernet protocols (such as Profinet or EtherNet / IP).

[0099] One of the core innovations of this invention lies in establishing a cross-module collaborative feedback control system, which transcends the traditional single-point, post-processing quality control model. This system does not simply reject defective products, but rather integrates and analyzes data from different production modules to achieve predictive and adaptive adjustment of key upstream process parameters, thereby proactively preventing quality deviations. The specific implementation of this system can be broken down into the following sub-steps.

[0100] S801, Distributed Data Acquisition and Synchronization. The foundation of this system is establishing a data link covering the entire process. The central control system 3 can acquire heterogeneous data from different steps in real time, including process parameters and key quality attribute data. To achieve effective correlation of cross-module data, the system employs a high-precision timestamp synchronization mechanism, combined with position sensors such as conveyor belt encoders, to compensate for minor delays during material transport. In this way, the central control system 3 can accurately trace back the entire set of specific process parameters experienced by each final detected finished product in each upstream process, thereby establishing a high-fidelity causal relationship dataset from process to result for each individual product—a complete dataset of process fingerprint and quality result.

[0101] S802, Establishment and Online Execution of Process Quality Prediction Model. To achieve the shift from passive detection to proactive prediction, the system internally constructs and runs one or more process quality prediction models. The core of this model is a mathematical function that describes the complex mapping relationship between upstream process parameters and final product quality attributes. This model can be expressed as:

[0102] ;

[0103] in, The key quality attribute vector predicted by the model; This is the input vector of key upstream process parameters; This is the function for the prediction model.

[0104] This model function The model was not established based on a simple linear assumption, but rather through a systematic design of experiments (DoE) to acquire a large amount of data, and then trained and validated using multivariate statistical analysis or machine learning algorithms. During production, the model receives process parameters collected in real time by the S801. And the quality of the products to be produced. Conduct online predictions.

[0105] S803, optimized control decision based on prediction bias. The central control system 3 continuously updates the model's prediction results. Compared with the preset quality target range The system performs a comparison. When it detects a persistent trend of the predicted value deviating from the target center, it proactively triggers an optimization control calculation. The goal of this calculation is to find an optimal adjustment amount for the process parameters. Its optimization problem can be formalized as:

[0106] ;

[0107] in, It is the process parameter adjustment vector to be solved; This is the current vector of process parameter settings; It is the target center value vector of the quality attribute; This represents the norm used to measure the overall deviation between the predicted and target values; and It is the constraint vector for parameter adjustment.

[0108] Determining the aforementioned constraint vector is crucial for ensuring the safe and stable operation of the system. The value of each component is not determined empirically, but rather based on a deep understanding of the process window: the upper limit... and lower limit The robust process range determined during the process development phase, the physical property limits of materials (such as the temperature sensitivity of melt flow index), and the mechanical and electrical safety limits of the equipment are all combined to ensure that any automatic adjustments will not lead to serious defects, damage to equipment, or safety problems.

[0109] S804, control command issuance and dual closed-loop verification. After solving the optimization problem, the central control system 3 will obtain the optimal adjustment amount. It is decomposed into instructions for specific device controllers and executed. The control architecture here is a dual closed-loop structure:

[0110] First closed loop (fast inner loop): based on model prediction With the goal The deviation is calculated and the process parameters are fine-tuned in real time. This is a loop with a fast response speed and small adjustment range, and its goal is to quickly suppress process drift caused by minor fluctuations in raw materials, environmental changes, etc.

[0111] The second closed loop (slow outer loop): After the adjustment is implemented, the system will continue to utilize the online detection results of S60. To verify the long-term effects of the adjustment measures, and With model prediction A comparison is made. When a systematic deviation occurs between the two, the system triggers an adjustment to the prediction model. The model itself undergoes relearning or parameter correction. This is a slower but more fundamental loop, aiming to continuously optimize the model's prediction accuracy and generalization ability, enabling it to adapt to a wider range of operating conditions.

Claims

1. An integrated production control method for blow-fill-seal manufacturing of pharmaceutical liquid plastic bottles, characterized in that, Includes the following steps: Control the extrusion unit (1) to extrude a continuous tubular preform, control the blow-fill-seal host (2) to close the mold to cut the tubular preform, and introduce sterile compressed gas to blow-form the preform into a sterile container; A predetermined dose of medicine solution, constrained by product specifications, is injected into the sterile container using an aseptic filling system. The neck of the sterile container is then sealed using a heat-sealing mechanism to obtain the finished container. The finished container is removed from the mold cavity and sent to the online quality monitoring unit for non-contact testing. The central control system (3) determines whether it is qualified or not based on the test data. The qualified finished container is sent backward and the unqualified finished container is physically removed. The central control system (3) collects the process parameters of the extrusion unit (1), blow-fill-seal host (2), aseptic filling system, heat sealing mechanism and online quality monitoring unit in real time, as well as the key quality attribute data obtained by the detection, establishes a mapping relationship between the two, and makes online predictions on the quality attributes of the subsequently generated finished containers based on the mapping relationship. When the prediction result deviates from the preset quality target range representing the boundary of qualified product attributes, the optimal process parameter adjustment amount is solved within the preset safety constraints limited by equipment operating limits and process robustness, and then sent to the corresponding equipment controller for closed-loop adaptive adjustment.

2. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 1, characterized in that, The specific steps for the central control system (3) to establish mapping relationships and perform online predictions include: A timestamp synchronization mechanism is adopted, and combined with position sensors to compensate for delays in the material conveying process, the real-time collected process parameters are synchronized and associated with the key quality attribute data of each corresponding finished product container. Based on the data obtained from the experimental design, a process quality prediction model composed of mathematical functions is established through training and verification. The process quality prediction model receives the current process parameters online and outputs the predicted key quality attribute vector.

3. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 2, characterized in that, The step of solving for the target optimal process parameter adjustment within the preset safety constraints limited by equipment operating limits and process robustness includes: Compare the predicted key quality attribute vector with the preset quality target range; When a deviation from the trend is detected, the optimization control calculation is actively triggered. Based on the preset upper limit and lower limit of parameter adjustment as constraints, the process quality prediction model is used to calculate the predicted value vector by superimposing the process parameter adjustment vector to be solved on the current process parameter set value vector. The optimization objective is to minimize the norm of the overall deviation between the predicted value vector and the preset quality attribute target center value vector, and solve for the target optimal process parameter adjustment amount. The upper and lower limits of parameter adjustment constitute the specific numerical boundaries of the preset safety constraint range, which are jointly determined by the robust process range determined during the process development stage, the physical performance limits of materials, and the mechanical and electrical safety limits of equipment.

4. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 3, characterized in that, The central control system (3) performs the distribution and closed-loop adaptive adjustment through a dual closed-loop structure, specifically including: The first closed loop involves, based on the deviation between the predicted key quality attribute vector and the target center value vector of the quality attribute, issuing and executing the target optimal process parameter adjustment amount to the equipment controller in real time; The second closed loop involves continuously collecting the actual key quality attribute data output by the online quality monitoring unit after the adjustment is performed, comparing the actual key quality attribute data with the predicted key quality attribute vector, and triggering the relearning and correction of the parameters of the process quality prediction model itself when a systematic deviation occurs.

5. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 1, characterized in that, During the extrusion of the continuous tubular preform, the central control system (3) dynamically adjusts the dimensional stability of the tubular preform by combining feedforward and feedback: The central control system (3) will add the preset base speed and speed adjustment amount according to the product specifications and material grade to calculate the instantaneous target speed of the extruder screw and output the control. The speed adjustment amount is calculated by multiplying the difference between the preset melt pressure target setting value and the real-time pressure value measured by the melt pressure sensor installed near the die head by the proportional gain coefficient of the pressure speed adjustment. The preset melt pressure target setting value is limited to a range that ensures the consistency of the blank size and avoids system oscillation.

6. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 1, characterized in that, In the process of introducing sterile compressed gas to blow-mold the preform into a sterile container, multi-stage pressure control is employed: The central control system (3) sends a continuously changing control signal to the high-speed gas proportional valve to make the gas pressure injected into the cut billet reproduce a preset pressure-time curve. The curve successively includes a pre-blowing stage with lower pressure and a main-blowing stage with higher pressure. The preset pressure-time curve is obtained by establishing a correlation model through experimental data with curve characteristics as input variables and wall thickness uniformity as output response. During the pressure holding and cooling stage after molding, real-time data is collected through temperature and flow sensors in the cooling pipeline, and the power of the cooling unit or the opening of the control valve is actively adjusted in a closed loop.

7. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 1, characterized in that, During the process of injecting a predetermined dose of medicine solution according to product specifications into the sterile container using an aseptic filling system and sealing the neck of the sterile container using a heat-sealing mechanism, the central control system (3) performs closed-loop control on the filling dose and the sealing temperature, sealing pressure and sealing time respectively: The pressure value measured by the pressure sensor in the pipeline is used for feedback adjustment to stabilize the pressure of the medicine before the terminal filter at the preset pressure value, and an opening pulse signal is sent to the filling valve to control the opening of the filling valve for a preset filling time, so as to ensure that the medicine injected into the sterile container reaches the predetermined dose. The preset pressure value is limited to a range that ensures the filling speed and avoids foaming of the liquid. The preset filling time is obtained by converting the weight of the product obtained from the trial filling into the volume and then correcting and calibrating it. By collecting real-time data from the temperature sensor inside the heat sealing mechanism, the electrical power supplied to the heating element inside the heat sealing mechanism is adjusted according to the deviation between the real-time data and the preset temperature setting value. The output force of the heat sealing mechanism is adjusted in a closed loop by feedback from the pressure sensor installed in the cylinder air supply line or from the current or torque feedback of the servo motor itself. Once the pressure reaches the preset pressure setting value, the internal timer is activated to maintain the clamping state under the output force of the heat sealing mechanism until the preset duration ends. The preset temperature setting, preset pressure setting, and preset duration are all taken from the process parameter constraints determined by experimental design methods and verified to be qualified.

8. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 1, characterized in that, The process of removing the finished container from the mold cavity includes the steps of finished product demolding and waste separation: The central control system (3) controls the ejection system of the mold integration to apply mechanical force to push the finished product container out of the mold cavity, and at the same time controls the gripper of the picking robot to grab the finished product container; When the gripper is in a stable state, the picking robot moves backward or upward along a preset trajectory. The relative motion causes stress concentration at the pre-designed weak connection point between the finished product container and the waste, resulting in fracture and separation of the finished product container from the waste.

9. The integrated production control method for blow-fill-seal of pharmaceutical liquid plastic bottles according to claim 1, characterized in that, In the process of the central control system (3) determining whether the data is qualified or not based on the detection data, the determination is achieved by executing a discrimination logic function. The specific steps include: Real-time acquisition of a multi-dimensional set of measurement values, including the impedance value measured by the high-voltage discharge system and the surface feature quantification index extracted by the machine vision system; The multidimensional measurement value set is input into the discrimination logic function, and the multidimensional measurement value set is compared with the preset qualified standard set. When all measurement values ​​meet the corresponding acceptable standard or threshold in the preset qualified standard set, a binary result of qualified is output. The threshold range in the preset set of qualified standards is set by repeatedly testing known qualified products and samples containing typical boundary defects, combined with statistical analysis.

10. An integrated production control device for blow-fill-seal manufacturing of pharmaceutical liquid plastic bottles, characterized in that, An integrated production control method for blow-fill-seal plastic bottles for performing any one of claims 1-9 includes an extrusion unit (1), a blow-fill-seal host (2), an online quality monitoring unit, and a central control system (3). The central control system (3) has a preset production control program running inside it. The central control system (3) communicates bidirectionally with the extrusion unit (1), the blow-fill-seal host (2) and the online quality monitoring unit through the industrial data bus.

Images