A sensor-based dynamic calibration method and system for fluid loading

Abstract

The present application relates to the technical field of automated fluid processing, and provides a sensor-based fluid loading dynamic calibration method and system, which comprises the following steps: after performing a calibration operation on a bioreactor, obtaining a first sensing signal output in a transition fluid disturbance observation stage by controlling a sensor of a fluid loading unit; obtaining a second sensing signal output in a stable sample physical property evaluation stage by controlling the sensor of the fluid loading unit; comparing features of the first sensing signal and the second sensing signal, and determining a disturbance compensation amount according to a comparison result; comparing features of the second sensing signal and a historical reference sensing signal, and determining a drift compensation amount according to a comparison result; and differentially compensating and adjusting driving parameters or sensor interpretation logic of the fluid loading unit according to the disturbance compensation amount and the drift compensation amount. The present application has the effect of providing a continuously accurate sample volume reflecting an undisturbed sample state.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automated fluid handling, and in particular to a sensor-based dynamic calibration method and system for fluid sampling. BACKGROUND

[0002] In applications such as biological manufacturing or long-term environmental monitoring, automated fluid sampling systems, as a key component connecting bioreactors and subsequent analysis units, directly affect the accuracy and reliability of the entire system. Such systems often need to extract micro samples for analysis at a high accuracy and repeatability at regular intervals during a long continuous operation period. The sampling unit is composed of a precision pump, a sampling pipeline, and process monitoring sensors. Although the system will perform initial flow calibration at startup, as the biological process progresses, the physicochemical properties of the culture medium will change, and components such as pumps and pipelines will also experience performance drift due to long-term operation.

[0003] Therefore, the existing sensor-based dynamic calibration method for fluid sampling cannot effectively identify and compensate for the disturbance of the calibration operation itself to the sample microenvironment and its impact on the sensor response, which may become a source of introducing new volumetric measurement bias, and cannot continuously provide sampling volumes that accurately reflect the state of undisturbed samples.

[0004] In view of the above problems, the existing technology needs to be improved. SUMMARY

[0005] The purpose of the present application is to provide a sensor-based dynamic calibration method and system for fluid sampling, which can effectively identify and compensate for the disturbance of the calibration operation itself to the sample microenvironment and its impact on the sensor response, thereby providing a sampling volume that continuously and accurately reflects the state of undisturbed samples.

[0006] The present application provides a sensor-based dynamic calibration method for fluid sampling, and the technical solution is as follows:

[0007] After performing the calibration operation on the bioreactor, the first sensor signal output during the transition fluid disturbance observation stage is obtained by controlling the sensor of the fluid sampling unit;

[0008] The second sensor signal output during the stable sample physical property evaluation stage is obtained by controlling the sensor of the fluid sampling unit;

[0009] The first sensor signal and the second sensor signal are compared in terms of features, and the disturbance compensation amount is determined according to the comparison result;

[0010] The second sensor signal and the historical reference sensor signal are compared in terms of features, and the drift compensation amount is determined according to the comparison result;

[0011] According to the disturbance compensation amount and the drift compensation amount, the driving parameters or sensor interpretation logic of the fluid loading unit are compensated and adjusted differently to complete the dynamic calibration of the fluid loading.

[0012] Through the above scheme, the disturbance caused by the calibration operation itself to the sample microenvironment and its influence on the sensor response can be effectively identified and compensated, thereby providing a loading volume that continuously and accurately reflects the state of the undisturbed sample.

[0013] Further, the present application also proposes that, according to the above-mentioned one kind based on sensor's fluid loading dynamic calibration method, according to the disturbance compensation amount and the drift compensation amount, the driving parameters or sensor interpretation logic of the fluid loading unit are compensated and adjusted differently to complete the dynamic calibration of the fluid loading.

[0014] Extracting the time feature and the morphological feature of the first sensing signal; extracting the time feature and the morphological feature of the second sensing signal;

[0015] Based on the morphological feature of the second sensing signal, a reference morphological feature is determined; the morphological feature of the first sensing signal is compared with the reference morphological feature to obtain a comparison result;

[0016] According to the comparison result, the time feature of the first sensing signal is corrected, or the proportion of the time feature of the first sensing signal in the calculation is modified;

[0017] According to the modified time feature of the first sensing signal or the proportion of the time feature in the calculation, the determination process of the disturbance compensation amount and the determination process of the drift compensation amount are re-performed;

[0018] According to the re-determined disturbance compensation amount and the drift compensation amount, the driving parameters or sensor interpretation logic of the fluid loading unit are compensated and adjusted differently to complete the dynamic calibration of the fluid loading.

[0019] Through the above scheme, the accuracy of the disturbance compensation and the drift compensation is further improved.

[0020] Further, the present application also proposes that, according to the above-mentioned one kind based on sensor's fluid loading dynamic calibration method, after performing the calibration operation on the bioreactor, the first sensing signal output in the transition fluid disturbance observation stage is obtained by controlling the sensor of the fluid loading unit.

[0021] After completing the calibration operation on the bioreactor, the standard disturbance liquid is injected into the bioreactor inlet by the fluid loading unit to form a transition fluid disturbance observation stage;

[0022] In the transition fluid disturbance observation stage, the pressure sensor and the flow rate sensor through the fluid sample loading passage collect continuous output data, and the data is subjected to time sequence filtering and window smoothing processing to obtain a first sensing signal.

[0023] Through the above scheme, the acquisition method of the disturbance signal is determined, and the signal quality is improved.

[0024] Further, the application also proposes that, according to the above-mentioned one kind based on sensor's fluid sample loading dynamic calibration method, according to the comparison result, the step of correcting the time characteristic of the first sensing signal or modifying the proportion of the time characteristic of the first sensing signal in the calculation includes:

[0025] Extracting interference mode features from the morphological features of the first sensing signal;

[0026] Comparing the interference mode features with the preset reference mode to determine whether the change of the morphological features is due to the executed calibration operation;

[0027] When it is determined that the change of the morphological features is due to the executed calibration operation, the time characteristic of the first sensing signal is corrected, or the proportion of the time characteristic of the first sensing signal in the calculation is modified.

[0028] Through the above scheme, it can be determined whether the change of the morphological features is due to the calibration operation, and misjudgment is avoided.

[0029] Further, the application also proposes that, according to the above-mentioned one kind based on sensor's fluid sample loading dynamic calibration method, the step of comparing the interference mode features with the preset reference mode to determine whether the change of the morphological features is due to the executed calibration operation includes:

[0030] Comparing the interference mode features with the preset reference mode to determine whether the change of the morphological features is due to the executed calibration operation;

[0031] When the determination result is due to the calibration operation, the time characteristic of the first sensing signal is corrected, or the proportion of the time characteristic of the first sensing signal in the calculation is modified, and the interference mode feature corresponding to the calibration operation is obtained as the deviation appearance;

[0032] According to the deviation appearance, the reference mode is updated.

[0033] Through the above scheme, the reference mode can dynamically adapt to the actual situation, and the determination accuracy is improved.

[0034] Further, the application also proposes that, according to the above-mentioned one kind based on sensor's fluid sample loading dynamic calibration method, the step of updating the reference mode according to the deviation appearance includes:

[0035] The deviation sample is combined with the reference pattern before updating by weighting to generate an updated reference pattern.

[0036] Through the above scheme, a specific reference pattern updating mechanism is provided.

[0037] Further, the application also proposes that, according to the above-mentioned sensor-based fluid sample dynamic calibration method, the step of combining the deviation sample with the reference pattern before updating by weighting to generate an updated reference pattern comprises:

[0038] Obtaining a historical deviation sample sequence composed of a plurality of deviation samples;

[0039] Generating a trend sample based on the historical deviation sample sequence;

[0040] Determining the difference between the trend sample and the reference pattern before updating as a difference degree;

[0041] Adjusting the weight for weighting combination according to the difference degree;

[0042] Combining the deviation sample with the reference pattern before updating by weighting according to the adjusted weight for weighting combination to generate an updated reference pattern.

[0043] Through the above scheme, the reference pattern updating is more robust and intelligent.

[0044] Further, the application also proposes that, according to the above-mentioned sensor-based fluid sample dynamic calibration method, the step of generating a trend sample based on the historical deviation sample sequence comprises:

[0045] Receiving the historical deviation sample sequence;

[0046] Smoothing the historical deviation sample sequence;

[0047] Generating a trend sample based on the smoothed historical deviation sample sequence.

[0048] Through the above scheme, the accuracy of the trend sample is improved.

[0049] Further, the application also proposes that, according to the above-mentioned sensor-based fluid sample dynamic calibration method, after the step of obtaining a historical deviation sample sequence composed of a plurality of deviation samples is performed, further comprising:

[0050] For each deviation sample in the historical deviation sample sequence, comparing it with the reference pattern;

[0051] According to the comparison result, determining whether the deviation sample is an abnormal data point;

[0052] Excluding the abnormal data point in the historical deviation sample sequence.

[0053] Through the above scheme, the reliability of the historical deviation profile sequence is improved.

[0054] The application also provides a sensor-based fluid sampling dynamic calibration system, and the technical scheme is as follows:

[0055] A sensor-based fluid sampling dynamic calibration system for performing sensor-based fluid sampling dynamic calibration, comprising:

[0056] The first sensing signal acquisition module is configured to, after performing the calibration operation on the bioreactor, acquire a first sensing signal output in the transition fluid disturbance observation stage by controlling the sensor of the fluid sampling unit;

[0057] The second sensing signal acquisition module is configured to acquire a second sensing signal output in the stable sample physical property evaluation stage by controlling the sensor of the fluid sampling unit;

[0058] The disturbance compensation amount acquisition module is configured to compare the features of the first sensing signal and the second sensing signal, and determine a disturbance compensation amount according to the comparison result;

[0059] The drift compensation amount acquisition module is configured to compare the features of the second sensing signal and the historical reference sensing signal, and determine a drift compensation amount according to the comparison result;

[0060] The dynamic calibration execution module is configured to differentially compensate and adjust the driving parameters or the sensor interpretation logic of the fluid sampling unit according to the disturbance compensation amount and the drift compensation amount, so as to complete the dynamic calibration of the fluid sampling.

[0061] Through the above scheme, a hardware / software carrier for implementing the above method is provided.

[0062] As can be seen from the above, the sensor-based fluid sampling dynamic calibration method and system provided by the application can effectively identify and compensate the disturbance caused by the calibration operation itself to the sample microenvironment and the influence of the disturbance on the sensor response, thereby providing a sampling volume that continuously and accurately reflects the state of the undisturbed sample. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 The method flowchart of the sensor-based fluid sampling dynamic calibration method in one embodiment of the application;

[0064] Figure 2 The method flowchart of the sensor-based fluid sampling dynamic calibration method in one embodiment of the application;

[0065] Figure 3 Figure 2 is a flow chart of a method for a sensor-based dynamic calibration method for fluid loading according to another embodiment of the present application;

[0066] Figure 4 Figure 3 is a flow chart of a method for a sensor-based dynamic calibration method for fluid loading according to another embodiment of the present application;

[0067] Figure 5 Figure 4 is a flow chart of a method for a sensor-based dynamic calibration method for fluid loading according to another embodiment of the present application;

[0068] Figure 6 Figure 5 is a flow chart of a method for a sensor-based dynamic calibration method for fluid loading according to another embodiment of the present application;

[0069] Figure 7 Figure 6 is a flow chart of a method for a sensor-based dynamic calibration method for fluid loading according to another embodiment of the present application;

[0070] Figure 8 Figure 7 is a flow chart of a method for a sensor-based dynamic calibration method for fluid loading according to another embodiment of the present application;

[0071] Figure 9 Figure 8 is a system block diagram of a sensor-based dynamic calibration system for fluid loading according to another embodiment of the present application;

[0072] BRIEF DESCRIPTION OF DRAWINGS

[0073] 1. A sensor-based dynamic calibration system for fluid loading; 11. A first sensing signal acquisition module; 12. A second sensing signal acquisition module; 13. A disturbance compensation amount acquisition module; 14. A drift compensation amount acquisition module; 15. A dynamic calibration execution module. DETAILED DESCRIPTION

[0074] The technical solutions in the present application will be described clearly and completely below in conjunction with the accompanying drawings in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. The components of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0075] It should be noted that like reference numerals and characters refer to like elements throughout the following description and the claims, and that, unless otherwise indicated, like reference numerals and characters in different figures refer to like elements. Also, in the description, terms "first", "second", etc. are used only for distinguishing between similar elements, and do not imply a sequential or chronological order, unless otherwise indicated.

[0076] The conventional existing automated fluid sampling system, when performing long-term, high-precision micro-sampling analysis on a bioreactor of an engineered strain that has physiological characteristics showing significant changes in response to external stimuli, if the calibration process introduces a calibration liquid with subtle differences from the environment in the reactor to ensure the clarity of the operation, causes a temporary physiological or chemical state change in the micro-environment of the sample near the sampling port that is difficult to directly perceive by conventional sensors, thereby affecting the representativeness of the subsequent sample or the response of the sensor, and there is a problem that the calibration process becomes the source of introducing new volume metering deviation, making it difficult to ensure that dynamic calibration can continuously provide the sampling volume that accurately reflects the undisturbed sample state in the reactor under the analysis requirements of long-time continuous operation and narrow range of concentration change of the detected substance.

[0077] For example, assume that in a continuous flow culture tank for producing a specific protein drug, an engineered strain is cultured that shows significant physiological response to osmotic pressure value or trace chemical substance concentration value. The automated fluid sampling system needs to extract nanoliter to microliter range of sample from the reactor for analysis at regular intervals. The system uses a sensor-based dynamic calibration method, using sterile physiological saline as the standard calibration liquid. In the calibration step, in order to ensure that the sampling port is not blocked, a very small volume of calibration liquid will be discharged from the end of the sampling needle and then reabsorbed. Although the volume of the calibration liquid is small, its osmotic pressure value differs from the culture liquid by several orders of magnitude in parts per million, or lacks some trace nutrients present in the culture liquid. This small difference causes a temporary local osmotic pressure change or nutrient concentration decrease in the micro-environment near the sampling port. The first sample extracted immediately after the calibration operation, the cells near the sampling port of which may have a temporary physiological state change due to the local osmotic pressure change, such as a slight shift in cell membrane permeability or metabolic product release rate. The sensor used for dynamic calibration, such as a sensor based on conductivity to detect the liquid interface, may not be able to directly identify this change in cell physiological state, because the difference in conductivity between the calibration liquid and the culture liquid is not large, or the sensor only focuses on the time of liquid column passing. The system performs dynamic calibration calculation according to this normal sensor signal, adjusting the driving parameters of the pump. This adjustment aims to compensate for the physical drift of the pump or pipeline, but since the sensor signal is indirectly affected by the sample micro-environment change introduced by the calibration, the dynamic calibration calculation may unconsciously also compensate for this slight deviation from the true state of the sample.

[0078] To this end, the application proposes a sensor-based dynamic calibration method for fluid loading, which combines Figure 1 as shown, comprising:

[0079] S1. After performing the calibration operation on the bioreactor, the first sensing signal output in the transition fluid disturbance observation stage is obtained by controlling the sensor of the fluid loading unit;

[0080] S2. The second sensing signal output in the stable sample physical property evaluation stage is obtained by controlling the sensor of the fluid loading unit;

[0081] S3. The first sensing signal and the second sensing signal are compared in terms of features, and the disturbance compensation amount is determined according to the comparison result;

[0082] S4. The second sensing signal and the historical reference sensing signal are compared in terms of features, and the drift compensation amount is determined according to the comparison result;

[0083] S5. According to the disturbance compensation amount and the drift compensation amount, the driving parameters or the sensor interpretation logic of the fluid loading unit are compensated and adjusted differently to complete the dynamic calibration of the fluid loading.

[0084] Wherein, after performing the calibration operation on the bioreactor, the first sensing signal outputted in the transition fluid disturbance observation stage is acquired by controlling the sensor of the fluid sampling unit, the transition fluid disturbance observation stage refers to the period after the completion of the calibration operation, the fluid state near the sampling port is temporarily changed due to the interaction between the calibration liquid and the fluid in the reactor, and the purpose is to capture the influence of the instantaneous or short-time disturbance caused by the calibration operation itself on the sensing signal; the second sensing signal outputted in the stable sample physical property evaluation stage is acquired by controlling the sensor of the fluid sampling unit, the stable sample physical property evaluation stage refers to the period after the influence of the transition fluid disturbance basically subsides, the fluid state representing the fluid state not directly affected by the calibration operation is acquired from the main area of the reactor, and the purpose is to acquire a relatively stable signal for evaluating the drift of the device itself and as a disturbance comparison reference; the first sensing signal is a signal collected in the transition fluid disturbance observation stage, reflects the comprehensive response of the device physical state after superimposing the fluid disturbance introduced by the calibration operation, and can be collected by a pressure sensor, a flow rate sensor, an optical sensor, a conductivity sensor and the like; the second sensing signal is a signal collected in the stable sample physical property evaluation stage, mainly reflects the response of the device physical state to the stable fluid state, and can be collected by a pressure sensor, a flow rate sensor, an optical sensor, a conductivity sensor and the like; the historical reference sensing signal refers to the sensing signal collected and stored by the system in a previous known stable state, and the purpose is to serve as a reference for evaluating the physical drift of the device; the feature comparison refers to analyzing the sensing signal, extracting its key features, and comparing the same or different features of different signals, for example, the rising time, peak value, integral area, waveform shape and the like of the signal waveform can be compared, and the purpose is to quantify the difference between the signals; the disturbance compensation amount is a value or parameter determined according to the feature comparison result of the first sensing signal and the second sensing signal, and represents the influence degree of the fluid disturbance introduced by the calibration operation on the sensing signal or the sampling process; the drift compensation amount is a value or parameter determined according to the feature comparison result of the second sensing signal and the historical reference sensing signal, and represents the influence degree of the drift of the physical components of the device over time on the sensing signal or the sampling process; the differentiated compensation adjustment refers to the targeted and different degrees of modification of the driving parameters or sensor interpretation logic of the fluid sampling unit according to the determined disturbance compensation amount and drift compensation amount, and the purpose is to simultaneously offset the influence of the calibration disturbance and the device drift.

[0085] The solution of the present application first controls the sensor of the fluid sampling unit to obtain a first sensing signal output in a transition fluid disturbance observation stage after performing the calibration operation on the bioreactor. The signal of this stage contains the physical drift of the device itself and the temporary disturbance information caused by the calibration operation to the micro-environment of the fluid near the sampling port. Then, after the disturbance effect basically subsides, the sensor of the fluid sampling unit is controlled again to obtain a second sensing signal output in a stable sample physical property evaluation stage. The signal of this stage mainly reflects the response of the physical drift of the device itself to the stable fluid state. Then, the first sensing signal and the second sensing signal are compared in features. By comparing the difference between the two signals, the influence of the disturbance introduced by the calibration operation on the sensing signal can be identified and quantified, so as to determine the disturbance compensation amount. At the same time, the second sensing signal and the historical reference sensing signal are compared in features. By comparing the difference between the signal under the current stable state and the historical reference signal, the drift of the device performance over time can be identified and quantified, so as to determine the drift compensation amount. Finally, according to the determined disturbance compensation amount and drift compensation amount, the driving parameters or sensor interpretation logic of the fluid sampling unit are compensated and adjusted differently. This differential adjustment can specifically offset the disturbance caused by the calibration operation and the drift of the device itself, so as to ensure that the subsequent sampling process can accurately obtain the target volume of fluid. The whole process forms a closed dynamic calibration mechanism, which can continuously monitor and correct various deviation sources in the sampling process.

[0086] In some preferred embodiments, the application is implemented as follows: assuming that the fluid loading unit uses a micro pump for fluid delivery, and is equipped with an optical sensor for detecting fluid passing and a pressure sensor for monitoring the pressure of the pipeline. After performing a calibration operation on the bioreactor, for example, injecting a small amount of standard calibration liquid, immediately start the micro pump for a short time to simulate the loading process, and collect the output data of the optical sensor and the pressure sensor, which after preprocessing form the first sensing signal. Wait for a period of time to ensure that the influence of the calibration liquid subsides, and then start the micro pump again to simulate the loading process, and collect the output data of the optical sensor and the pressure sensor to form the second sensing signal. Extract the features of the first sensing signal and the second sensing signal, such as the rise time of the pressure waveform, the peak pressure, the liquid passing time of the optical signal, etc. Compare the features of the first sensing signal and the second sensing signal, calculate the difference between them, and determine the disturbance compensation amount according to the difference. For example, if the pressure peak of the first signal is lower than that of the second signal, it may indicate that the calibration liquid has reduced the local viscosity, and the driving force of the pump needs to be increased. At the same time, extract the features of the second sensing signal and compare them with the corresponding features of the historical reference sensing signal stored in the system, calculate the difference, and determine the drift compensation amount according to the difference. For example, if the optical passing time of the second signal is longer than the historical reference, it may indicate that the pump efficiency has decreased, and the running time of the pump needs to be increased. Finally, according to the calculated disturbance compensation amount and drift compensation amount, adjust the driving voltage, pulse width or running time of the micro pump, or modify the threshold or calibration curve of the sensor to determine the liquid level position, to achieve dynamic calibration of the loading volume.

[0087] Optionally, in combination with Figure 2 As shown, S5 includes the step of differentially compensating and adjusting the driving parameters or sensor interpretation logic of the fluid loading unit according to the disturbance compensation amount and the drift compensation amount to complete the dynamic calibration of the fluid loading.

[0088] S51, extract the time features and morphological features of the first sensing signal; extract the time features and morphological features of the second sensing signal;

[0089] S52, determine the reference morphological features based on the morphological features of the second sensing signal; compare the morphological features of the first sensing signal with the reference morphological features to obtain a comparison result;

[0090] S53, according to the comparison result, correct the time features of the first sensing signal, or modify the proportion of the time features of the first sensing signal in the calculation;

[0091] S54, according to the modified time features of the first sensing signal or the proportion of the time features in the calculation, re-perform the determination process of the disturbance compensation amount and the determination process of the drift compensation amount;

[0092] S55, differentially compensating and adjusting the driving parameters of the fluid sample loading unit or the sensor interpretation logic according to the re-determined disturbance compensation amount and drift compensation amount, to complete dynamic calibration of fluid sample loading.

[0093] wherein the time characteristic refers to a regularity description of the change of the sensing signal with time, which can be represented by parameters such as the time point of signal occurrence, duration, rise time, fall time, etc.; the shape characteristic refers to the shape and structural attributes of the waveform of the sensing signal, which can be represented by parameters such as the peak value, width, slope, waveform symmetry, etc. of the signal; the reference shape characteristic refers to the shape characteristic serving as a comparison standard, which can be determined by statistically averaging, modeling or selecting a typical value of the shape characteristics of the second sensing signals obtained in the stable sample physical property evaluation stage; the correction of the time characteristic of the first sensing signal refers to adjusting the time axis or key time point data of the first sensing signal through an algorithm or model, to correct the time deviation introduced by external disturbance; the proportion of the correction of the time characteristic of the first sensing signal in the calculation refers to adjusting the weight coefficient of the time characteristic parameter of the first sensing signal in the algorithm model for calculating the disturbance compensation amount and the drift compensation amount, to reduce or increase its influence on the calculation result; the determination process of the disturbance compensation amount refers to a series of algorithm steps for calculating the value or adjustment amount of the instantaneous disturbance caused by the calibration operation that needs to be compensated based on the sensing signal characteristics; the determination process of the drift compensation amount refers to a series of algorithm steps for calculating the value or adjustment amount of the performance drift caused by long-term operation of the device that needs to be compensated based on the sensing signal characteristics; the differential compensation adjustment refers to making targeted and differentiated modifications to the driving parameters (such as the driving voltage, pulse number, running time of the pump) or the sensor interpretation logic (such as the signal threshold value, gain, interpretation algorithm) of the fluid sample loading unit according to the calculated disturbance compensation amount and drift compensation amount.

[0094] In some preferred embodiments, the implementation is as follows. First, the time features of the first sensing signals outputted in the transition fluid disturbance observation stage, such as the time point of the signal peak, and the morphological features, such as the width of the signal waveform and the height of the peak, are extracted. At the same time, the time features and morphological features of the second sensing signals outputted in the stable sample physical property evaluation stage are extracted. Based on the morphological features of the second sensing signals, such as the average waveform width and peak height, the reference morphological features are determined. The morphological features (waveform width and peak height) of the first sensing signals currently acquired are compared with the reference morphological features, such as calculating the Euclidean distance or correlation coefficient between them. If the comparison result shows that there is a significant difference between the morphological features of the first sensing signals and the reference morphological features, then according to the difference degree, the peak time point of the first sensing signals is corrected, or the weight of the peak time point in the calculation model is reduced when calculating the disturbance compensation amount. Then, using the corrected peak time point or the modified weight, and other signal features, the disturbance compensation amount and the drift compensation amount are recalculated. Finally, according to the recalculated disturbance compensation amount and drift compensation amount, such as if the disturbance compensation amount indicates that the sample volume needs to be increased, the number of driving pulses of the pump is increased; if the drift compensation amount indicates that the sensor reading threshold needs to be adjusted, the reading threshold is modified, to complete the dynamic calibration of the fluid sampling.

[0095] Optionally, in combination with Figure 3 As shown in FIG. 1, after S1 performs the calibration operation on the bioreactor, the step of acquiring the first sensing signals outputted in the transition fluid disturbance observation stage by controlling the sensors of the fluid sampling unit includes:

[0096] S11, after completing the calibration operation on the bioreactor, controlling the fluid sampling unit to inject a standard disturbance liquid into the inlet of the bioreactor to form a transition fluid disturbance observation stage;

[0097] S12, in the transition fluid disturbance observation stage, acquiring continuous output data through the pressure sensor and the flow rate sensor of the fluid sampling channel, and performing time series filtering and window smoothing processing on the data to obtain the first sensing signals.

[0098] The standard disturbance liquid refers to a liquid with known physical and chemical properties, such as physiological saline, a specific buffer solution, or pure water, which is used to simulate or amplify the influence of the calibration operation on the fluid environment, and the purpose is to provide a controllable fluid sample that reflects the instantaneous state after calibration; the transition fluid disturbance observation stage refers to a short time window before the sample fluid is completely stable after the completion of the calibration operation, during which the influence of the calibration operation on the local fluid environment may not have completely dissipated, and the standard disturbance liquid is injected to define and extend this stage for observation, and the purpose is to isolate and capture the instantaneous influence of the calibration operation on the fluid state; the pressure sensor and the flow rate sensor of the fluid loading channel refer to devices for monitoring the changes of pressure and flow rate of the fluid when flowing in the pipeline in real time, the pressure sensor can detect the fluid resistance or driving pressure, and the flow rate sensor can detect the speed or flow of the fluid passing through, and the purpose is to comprehensively obtain the dynamic physical information in the fluid loading process; the continuous output data refers to the data stream collected and output by the sensor without interruption within a period of time at a set sampling frequency, rather than discrete or single-point data, and the purpose is to capture the complete process of the change of the fluid state over time; the time sequence filtering and window smoothing processing refers to applying digital signal processing technology to the time sequence data collected continuously, the time sequence filtering is used to remove high-frequency noise or interference of specific frequency, and the window smoothing processing (such as moving average) is used to reduce data fluctuations and highlight data trends, and the purpose is to improve the signal-to-noise ratio and stability of the sensing signal, so that it more accurately reflects the real change of the fluid state.

[0099] In some preferred embodiments, specifically, after completing the calibration operation on the bioreactor, the precision pump of the fluid loading unit injects a specific volume of physiological saline into the bioreactor inlet as a standard disturbance liquid, forming a transition fluid disturbance observation stage, which lasts for a specific time. Within this specific time, the piezoresistive pressure sensor and the ultrasonic flow rate sensor installed on the fluid loading channel continuously collect pressure and flow rate data at a set sampling frequency. The collected data stream is input to the processing unit, first subjected to Kalman filtering to remove high-frequency noise, and then subjected to moving average smoothing processing of a specific window size, finally obtaining the first sensing signal.

[0100] Optionally, in combination with Figure 4 As shown in S53, the step of correcting the time characteristics of the first sensing signal or modifying the proportion of the time characteristics of the first sensing signal in the calculation according to the comparison result includes:

[0101] S531, extracting the interference mode feature from the morphological feature of the first sensing signal;

[0102] S532, comparing the interference mode feature with the preset reference mode to determine whether the change of the morphological feature is due to the executed calibration operation;

[0103] S533, when determining that the change in the shape feature is due to the performed calibration operation, modifying the time feature of the first sensing signal, or adjusting the proportion of the time feature of the first sensing signal in the calculation.

[0104] wherein the interference pattern feature refers to a pattern extracted from the shape feature of the first sensing signal that reflects the signal change caused by the calibration operation, which can be achieved by signal decomposition, feature extraction algorithm or pattern recognition technology, the purpose of which is to decompose the change in the shape feature into different components, so as to more easily distinguish the change caused by the calibration operation from the change caused by other factors; the preset reference pattern refers to a pattern representing the typical change in the shape feature caused by the calibration operation, which can be achieved by historical calibration operation data statistics, theoretical model construction or expert experience setting, the purpose of which is to provide a reference standard for judging whether the current change in the shape feature is consistent with the expected result of the calibration operation; comparing the interference pattern feature with the preset reference pattern to determine whether the change in the shape feature is due to the performed calibration operation refers to comparing the similarity or difference between the interference pattern feature and the reference pattern, and determining whether the current change in the shape feature is consistent with the typical change caused by the calibration operation by setting a threshold or a classifier, which can be achieved by distance calculation, correlation analysis or machine learning classification, the purpose of which is to determine whether the change in the shape feature is indeed due to the calibration operation before modifying the time feature of the first sensing signal.

[0105] In some preferred embodiments, the interference pattern feature can be extracted from the shape feature of the first sensing signal using wavelet analysis method, the shape feature curve of the first sensing signal is decomposed at multiple scales, and the high-frequency components at a certain scale are extracted as the interference pattern feature, which can reflect the instantaneous disturbance or noise in the signal. The preset reference pattern can be obtained by performing multiple calibration operations under standard conditions when the system is first installed or regularly maintained, recording the shape feature of the first sensing signal obtained after each calibration operation, extracting the corresponding wavelet high-frequency components as a historical interference pattern feature sample set, and then calculating the average pattern or median pattern of the sample set as the initial reference pattern. When comparing the interference pattern feature with the preset reference pattern to determine whether the change in the shape feature is due to the performed calibration operation, the correlation coefficient between the currently extracted interference pattern feature and the reference pattern can be calculated, and if the correlation coefficient is greater than a preset threshold, it is determined that the change in the shape feature is due to the performed calibration operation. When determining that the change in the shape feature is due to the performed calibration operation, the amplitude of modifying the time feature of the first sensing signal or the weight coefficient of the time feature of the first sensing signal in calculating the disturbance compensation amount can be adjusted according to the difference between the interference pattern feature and the reference pattern, for example, the inverse of the correlation coefficient. For example, the higher the correlation coefficient, the greater the weight.

[0106] Optionally, in combination with Figure 5 As shown, S532 comprises the step of comparing the interference pattern feature with a preset reference pattern to determine whether the change in the morphology feature is due to the performed calibration operation.

[0107] S5321, comparing the interference pattern feature with a preset reference pattern to determine whether the change in the morphology feature is due to the performed calibration operation.

[0108] S5322, when the determination result is due to the calibration operation, correcting the time feature of the first sensing signal, or modifying the proportion of the time feature of the first sensing signal in the calculation, and obtaining the interference pattern feature corresponding to the calibration operation as the deviation appearance.

[0109] S5323, updating the reference pattern according to the deviation appearance.

[0110] Wherein, the interference pattern feature refers to the specific waveform or pattern extracted from the morphology feature of the first sensing signal, which reflects the possible influence of the calibration operation, which can be extracted by waveform analysis, frequency spectrum analysis or machine learning model, and the purpose is to capture the influence of the calibration operation on the morphology of the sensing signal; the preset reference pattern refers to the reference pattern used for comparison with the interference pattern feature, which represents the morphology feature or morphology feature in the ideal state that the system should have under normal calibration operation, which can be based on historical data statistical average, modeling or expert experience setting, and the purpose is to provide a reference standard for judging the influence of the calibration operation; the deviation appearance refers to the interference pattern feature corresponding to the calibration operation when the change in the morphology feature is due to the calibration operation, which records the specific influence mode of the specific calibration operation on the morphology of the sensing signal, and the purpose is to provide data for subsequent reference pattern update; updating the reference pattern refers to the process of adjusting or correcting the preset reference pattern according to the obtained deviation appearance, which can adopt weighted average, exponential smoothing or trend analysis based on historical sequence, and the purpose is to make the reference pattern adapt to the change of the system over time and the cumulative effect that the calibration operation itself may bring, and improve the accuracy of subsequent identification.

[0111] In some preferred embodiments, the comparison between the interference pattern feature and the preset reference pattern can be specifically implemented by calculating the Euclidean distance between the two, and a distance threshold is set, when the calculated distance is greater than the threshold, it is determined that the change in the shape feature is due to the calibration operation performed. When the determination result is due to the calibration operation, the modification of the time feature of the first sensing signal can be to stretch or compress the time sequence of the first sensing signal linearly, so that it is aligned with the reference pattern on the time axis. Alternatively, the proportion of the modified time feature of the first sensing signal in the calculation can be to multiply the time feature by a preset attenuation factor, for example 0.7, when calculating the disturbance compensation amount. Obtaining the interference pattern feature corresponding to the calibration operation as the deviation morphology can be to store the currently extracted interference pattern feature data in a historical record list. Updating the reference pattern according to the deviation morphology can adopt a moving average method, averaging the latest deviation morphology with the last N deviation morphologies, and then performing a weighted average of the average value and the old reference pattern to generate a new reference pattern.

[0112] Optionally, the step of updating the reference pattern according to the deviation morphology in S5323 comprises:

[0113] Combining the deviation morphology and the reference pattern before updating by weighted combination to generate the updated reference pattern.

[0114] Wherein, the deviation morphology refers to the interference pattern feature extracted from the shape feature of the first sensing signal after the calibration operation, which is different from the preset reference pattern, and reflects the specific change pattern that the calibration operation may introduce to the shape of the sensor signal; the reference pattern refers to the preset pattern used for comparison with the interference pattern feature, which represents the sensor signal shape feature of the system in an ideal or normal state, and can be established by historical data averaging, statistical model or expert experience, etc.; the weighted combination refers to a mathematical operation of linearly superimposing two or more values or vectors according to a preset weight ratio, which can be implemented in the form of formula W1*A+W2*B, wherein A and B are the quantities to be combined, W1 and W2 are the corresponding weights, and usually satisfy W1+W2=1, and the purpose is to give different influence according to the reliability or importance when fusing different information sources.

[0115] In some preferred embodiments, the deviation profile is combined with the reference pattern before update by weighting to generate the reference pattern after update, which can be implemented in a linear weighting manner. For example, a weight factor alpha can be set, which is in the range of 0 to 1. The reference pattern after update can be calculated by the formula "reference pattern_new = alpha * deviation profile + (1-alpha) * reference pattern_old". Wherein, "reference pattern_old" is the reference pattern before update, "deviation profile" is the interference pattern feature obtained by the current calibration operation, and "reference pattern_new" is the reference pattern after update. The value of the weight factor alpha can be determined according to the characteristics of the system, the noise level of the data or the desired update speed. For example, if it is desired that the reference pattern be more sensitive to new deviation profiles, the value of alpha can be increased; if it is desired that the reference pattern be more stable and not sensitive to single data, the value of alpha can be reduced. In this way, the influence of new information on the update of the reference pattern can be flexibly controlled.

[0116] Optionally, in combination with Figure 6 As shown in the figure, the step of combining the deviation profile with the reference pattern before update by weighting to generate the reference pattern after update comprises:

[0117] A1, obtaining a historical deviation profile sequence composed of a plurality of deviation profiles;

[0118] A2, generating a trend profile based on the historical deviation profile sequence;

[0119] A3, determining the difference between the trend profile and the reference pattern before update as a difference degree;

[0120] A4, adjusting the weight for weighting combination according to the difference degree;

[0121] A5, combining the deviation profile with the reference pattern before update by weighting according to the adjusted weight for weighting combination to generate the reference pattern after update.

[0122] The historical deviation profile sequence refers to a set of deviation profiles generated by a plurality of calibration operations performed at different time points arranged in chronological order, which can be realized by a data list stored in a memory or a database, and aims to provide historical data for analyzing long-term impact trends of the calibration operations; the trend profile refers to a long-term and stable change pattern reflecting the impact of the calibration operations on the sample microenvironment obtained by analyzing or modeling the historical deviation profile sequence, which can be realized by statistical methods such as moving average, exponential smoothing or regression analysis, and aims to extract a representative change trend from discrete deviation profiles; the difference degree refers to the deviation of the trend profile from the reference mode before updating in terms of features, which can be realized by calculating the distance, similarity or difference of a specific statistical indicator between the two, and aims to quantify the gap between the cumulative impact of the calibration operation and the current reference mode; and the weight for weighted combination refers to dynamically changing the proportion of the deviation profile and the reference mode before updating in the weighted combination according to the size of the difference degree, which can be realized by a preset weight adjustment function or lookup table, and aims to enable the updated reference mode to more reasonably integrate the latest deviation information and historical trend information.

[0123] In some preferred embodiments, the historical deviation profile sequence can be stored in a ring buffer recording deviation profiles generated by the last N calibration operations. When generating the trend profile, a moving average calculation can be performed on the historical deviation profile sequence in the ring buffer to obtain a smoothed trend profile. When determining the difference degree, the Euclidean distance between the trend profile and the reference mode before updating in the key feature dimension can be calculated. According to the Euclidean distance as the difference degree, the weights of the deviation profile and the reference mode in the weighted combination can be adjusted by a preset nonlinear function, for example, when the difference degree is small, the weight of the deviation profile can be relatively high to quickly respond to the latest changes; when the difference degree is large, the weight of the deviation profile can be relatively low, and the weight of the reference mode can be relatively high to avoid abnormal deviation profiles causing excessive impact on the reference mode, while ensuring that the updating direction is consistent with the long-term trend. Finally, according to the adjusted weights, the deviation profile and the reference mode before updating are combined by a linear weighted formula to obtain the updated reference mode.

[0124] Optionally, in combination with Figure 7 As shown in the figure, the step of A2 generating the trend profile based on the historical deviation profile sequence includes:

[0125] A21, receiving the historical deviation profile sequence;

[0126] A22, smoothing the historical deviation profile sequence;

[0127] A23, generating the trend profile based on the smoothed historical deviation profile sequence.

[0128] wherein the historical deviation profile sequence refers to a collection of deviation profiles obtained in multiple calibration operations arranged in chronological order, which can be stored in system memory, database or file; the smoothing process refers to an algorithm applied to time series data to reduce noise and short-term fluctuations, revealing underlying trends or periodicity, which can be achieved by moving average, exponential smoothing, Savitzky-Golay filtering, etc., the purpose of which is to filter out random errors and highlight long-term trends; the trend profile refers to a representative pattern or vector extracted from the historical deviation profile sequence, reflecting the long-term evolution direction and amplitude of the deviation pattern, the purpose of which is to provide a more stable reference than a single deviation profile, which can better represent the long-term behavior of the system.

[0129] In some preferred embodiments, the application is implemented as follows: after each calibration operation, the system stores the obtained deviation profile in a ring buffer or database, forming a time series containing several recent deviation profiles, i.e. the historical deviation profile sequence. The historical deviation profile sequence is smoothed, which can be achieved by moving average, for example, for each data point in the historical deviation profile sequence, the average value of the data point and several adjacent data points is calculated as the smoothed value of the data point. Alternatively, exponential smoothing can be used, giving higher weight to recent data. Based on the smoothed historical deviation profile sequence, the trend profile is generated, which can be based on the last point of the smoothed sequence as the current trend profile, or a linear regression is performed on the smoothed sequence, and the slope and intercept of the regression line are combined into the characteristic vector of the trend profile.

[0130] Optionally, in combination with Figure 8 As shown in FIG. 1, after the step of obtaining a historical deviation profile sequence composed of several deviation profiles, the step A1 further includes:

[0131] A11, for each deviation profile in the historical deviation profile sequence, compare it with the reference pattern;

[0132] A12, according to the comparison result, determine whether the deviation profile is an abnormal data point;

[0133] A13, exclude the abnormal data point from the historical deviation profile sequence.

[0134] wherein comparing each of the historical deviation profiles to the reference pattern means performing a comparison operation between each data point, i.e. each deviation profile, in the historical deviation profile sequence and the currently used reference pattern, which can be implemented by calculating a difference measure between the deviation profile and the reference pattern, such as the Euclidean distance, correlation coefficient or structural similarity index between them, with the purpose of quantifying the degree of deviation between each deviation profile and the reference pattern representing the normal or expected pattern; determining whether the deviation profile is an abnormal data point means applying a pre-set rule or statistical method to decide whether the deviation profile significantly deviates from the reference pattern based on the difference measure obtained from the comparison, so as to be marked as abnormal, which can be implemented by setting a fixed threshold, when the difference measure exceeds the threshold, the deviation profile is determined as abnormal, or by statistical analysis based on historical data, such as calculating the mean and standard deviation of historical difference measures, points exceeding a certain standard deviation range are determined as abnormal, with the purpose of identifying outliers that do not conform to the normal pattern, which can be caused by accidental factors or errors; excluding abnormal data points in the historical deviation profile sequence means removing or marking the deviation profiles determined as abnormal data points from the historical deviation profile sequence, so that they do not participate in the subsequent trend profile generation process, which can be implemented by creating a new sequence that does not contain abnormal data points, or skipping the data points marked as abnormal during the processing process, with the purpose of purifying historical data and improving the data quality used to generate the trend profile.

[0135] In some preferred embodiments, in particular, after obtaining the historical deviation profile sequence consisting of several deviation profiles, the system can traverse each of the deviation profiles in the sequence. For the currently processed deviation profile, the Euclidean distance between it and the currently stored reference pattern is calculated, which can be used as the comparison result to measure the degree of deviation. Then, the system can determine whether the deviation profile is an abnormal data point according to a pre-set abnormality determination rule, for example, if the calculated Euclidean distance exceeds a certain fixed multiple (e.g. three times the standard deviation) of the average Euclidean distance of historical data, the deviation profile is determined as an abnormal data point. Finally, when constructing the sequence used to generate the trend profile, the system can skip all deviation profiles determined as abnormal data points and only use non-abnormal deviation profiles to perform subsequent smoothing processing and trend profile generation steps.

[0136] By the above technical solutions, abnormal data points in the historical deviation profile sequence can be effectively identified and removed, thereby improving the data quality used to generate the trend profile, so that the generated trend profile can more accurately reflect the actual drift trend of the system, thereby optimizing the updating process of the reference pattern, enhancing the accuracy and stability of dynamic calibration, and reducing the adverse effects of occasional abnormal data on long-term calibration results.

[0137] A sensor-based fluid loading dynamic calibration system for performing sensor-based fluid loading dynamic calibration, comprising Figure 9 As shown in the figure, the sensor-based fluid loading dynamic calibration system 1 comprises:

[0138] A first sensor signal acquisition module 11, configured to acquire a first sensor signal output in a transition fluid disturbance observation stage after performing a calibration operation on a bioreactor by controlling a sensor of a fluid loading unit;

[0139] A second sensor signal acquisition module 12, configured to acquire a second sensor signal output in a stable sample physical property evaluation stage by controlling the sensor of the fluid loading unit;

[0140] A disturbance compensation amount acquisition module 13, configured to compare features of the first sensor signal and the second sensor signal, and determine a disturbance compensation amount according to a comparison result;

[0141] A drift compensation amount acquisition module 14, configured to compare features of the second sensor signal and a historical reference sensor signal, and determine a drift compensation amount according to a comparison result;

[0142] A dynamic calibration execution module 15, configured to perform differential compensation adjustment on a driving parameter or a sensor interpretation logic of the fluid loading unit according to the disturbance compensation amount and the drift compensation amount, so as to complete dynamic calibration of fluid loading.

[0143] The first sensing signal acquisition module refers to a processing unit for collecting sensor data at a specific time point or stage, which can be implemented by an integrated circuit, a microprocessor or a dedicated signal processing hardware. The second sensing signal acquisition module refers to a processing unit for collecting sensor data at another specific time point or stage, which can be implemented by a similar hardware or software structure as the first sensing signal acquisition module. The disturbance compensation amount acquisition module refers to a processing unit for analyzing and comparing the features of the two sets of sensing signals and calculating a value reflecting the deviation introduced by the calibration operation, which can be implemented by a software program running a specific algorithm or a digital signal processor. The drift compensation amount acquisition module refers to a processing unit for analyzing and comparing the current sensing signal with historical data and calculating a value reflecting the performance degradation of the device, which can be implemented by a combination of a data storage unit and a calculation unit. The dynamic calibration execution module refers to a control unit for adjusting the working parameters of the fluid sampling unit or the data processing method according to the calculated compensation value, which can be implemented by a microcontroller or a programmable logic device that outputs control signals. The fluid sampling unit refers to the actuator responsible for the actual extraction or delivery of fluid, which can be implemented by a precision syringe pump, a peristaltic pump or a microfluidic pump. The sensor refers to a device for sensing the physical or chemical properties of the fluid and outputting corresponding signals, which can be implemented by a pressure sensor, a flow rate sensor, a conductivity sensor or an optical sensor. The transition fluid disturbance observation stage refers to the short period after the completion of the calibration operation, during which the calibration liquid and the fluid in the reactor mix or interact with each other. The stable sample physical property evaluation stage refers to the period after the end of the transition stage, during which the fluid near the sampling port returns to a relatively stable state. The historical reference sensing signal refers to the sensing signal data collected and stored by the system under previous normal operation or standard conditions, which is used as a reference for evaluating the drift of the device. The driving parameter refers to the input signal or configuration value that controls the action of the fluid sampling unit, which can be implemented by the driving voltage, pulse width, step frequency or running time of the pump. The sensor interpretation logic refers to the algorithm or rule used by the system to interpret the original signal of the sensor to determine the state or volume of the fluid, which can be implemented by threshold judgment, signal integration or pattern recognition algorithm. The differentiated compensation adjustment refers to the differentiated and refined modification of the fluid sampling unit according to two different compensation amounts (disturbance compensation amount and drift compensation amount).

[0144] The scheme of the present application realizes dynamic calibration of fluid sampling process through modular system design, solves the problem that calibration process may introduce new volumetric measurement bias. Specifically, after performing calibration operation on the bioreactor, the system first acquires the sensor signal in the transition fluid disturbance observation stage through the first sensor signal acquisition module, captures the instantaneous influence of the calibration operation on the local fluid state. Subsequently, the second sensor signal acquisition module acquires the sensor signal in the stable sample physical property evaluation stage, reflecting the physical properties of the sample in the relatively stable state. The disturbance compensation amount acquisition module quantifies the influence of the disturbance introduced by the calibration operation on the sensor response by comparing the sensor signal characteristics of the two stages, and calculates the value that needs to be compensated. At the same time, the drift compensation amount acquisition module compares the sensor signal in the stable stage with the pre-stored historical reference sensor signal, evaluates the performance drift caused by long-term operation of the device, and calculates the corresponding drift compensation amount. Finally, the dynamic calibration execution module comprehensively utilizes the two independent compensation amounts to make differential adjustment to the driving parameters of the fluid sampling unit or the sensor interpretation logic. This double compensation mechanism enables the system to correct both the physical drift of the device itself and the potential disturbance to the sample microenvironment caused by the calibration operation, ensuring that the final sampling volume accurately reflects the fluid state in the reactor unaffected by the calibration. In this way, the system not only compensates for the errors caused by device aging, but also avoids the calibration process itself becoming a new error source, improving the accuracy and reliability of the fluid sampling volume in long-term operation.

[0145] In some preferred embodiments, the application is implemented as follows. The fluid loading unit can be a precision syringe pump driven by a stepper motor, with driving parameters including the number of pulses and the frequency of the stepper motor. The sensors can be micro pressure sensors and flow rate sensors installed in the syringe pump pipeline. The first and second sensing signal acquisition modules can be integrated on a microcontroller, which acquires the output voltage signals of the pressure and flow rate sensors at different time windows after calibration through a control analog-to-digital converter (ADC), and converts the analog signals into digital signals. The disturbance compensation amount acquisition module and the drift compensation amount acquisition module can be implemented as software programs running on the microcontroller. The disturbance compensation amount acquisition module can analyze the pressure and flow rate signal waveform characteristics (such as peak value, falling edge slope, stable value) in the transition phase (for example, 0-5 seconds after calibration) and the stable phase (for example, 10-15 seconds after calibration), and calculate the difference between the two as the disturbance compensation amount. The drift compensation amount acquisition module can compare the signal stable value in the stable phase with the historical average stable value stored in the non-volatile memory, and calculate the drift compensation amount reflecting the change in pump or pipeline efficiency. The dynamic calibration execution module can also be a software program on the microcontroller, which adjusts the number of pulses or frequency sent to the stepper motor driver, or modifies the pressure / flow rate threshold for judging the liquid interface, according to the calculated disturbance compensation amount and drift compensation amount. For example, if it is calculated that there is a temporary pressure rise (disturbance) caused by calibration and a flow rate drop (drift) caused by device aging, the system can simultaneously fine-tune the pulse number to compensate for the drift, and temporarily adjust the sensor reading threshold to adapt to the signal characteristics in the disturbance phase, or make a weighted correction to the data in the disturbance phase when calculating the final loading volume.

[0146] Through the above technical solutions, the application provides a fluid loading dynamic calibration system based on sensors, which can effectively identify and compensate for the disturbance to the sample microenvironment caused by the calibration operation itself and the loading volume drift caused by the aging of the physical components of the device. This enables the system to continuously provide a loading volume that accurately reflects the state of the undisturbed sample in the reactor during long-term continuous operation, avoiding the calibration process from becoming a new source of volume measurement deviation. Especially for biological systems that are sensitive to environmental changes, this system can ensure that the collected samples have higher representativeness, thereby improving the reliability of subsequent analysis results and supporting more accurate biological process monitoring and control.

[0147] The above only describes embodiments of the application and is not intended to limit the protection scope of the application. For those skilled in the art, the application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the application shall be included in the protection scope of the application.

Claims

1. A sensor-based dynamic calibration method for fluid loading, characterized by, The method comprises the following steps: After performing the calibration operation on the bioreactor, a first sensing signal output by a transition fluid disturbance observation stage is obtained by controlling the sensor of the fluid sampling unit; A second sensing signal output by a stable sample physical property evaluation stage is obtained by controlling the sensor of the fluid sampling unit; The first sensing signal and the second sensing signal are compared in terms of features, and a disturbance compensation amount is determined according to the comparison result; The second sensing signal and a historical reference sensing signal are compared in terms of features, and a drift compensation amount is determined according to the comparison result; The driving parameters or sensor interpretation logic of the fluid sampling unit are differentially compensated and adjusted according to the disturbance compensation amount and the drift compensation amount, so as to complete dynamic calibration of fluid sampling. The transition fluid disturbance observation stage refers to a period in which, after the completion of the calibration operation, the fluid state near the sampling port is temporarily changed due to the interaction between the calibration liquid and the fluid in the bioreactor. The stable sample physical property evaluation stage refers to a period in which, after the influence of the transition fluid disturbance is basically eliminated, the fluid state representing the fluid state not directly affected by the calibration operation is obtained from the main area of the bioreactor.

2. The sensor-based dynamic calibration method for fluid loading according to claim 1, wherein, The step of differentially compensating and adjusting the driving parameters or sensor interpretation logic of the fluid sampling unit according to the disturbance compensation amount and the drift compensation amount to complete dynamic calibration of fluid sampling comprises the following steps: Time features and shape features of the first sensing signal are extracted; time features and shape features of the second sensing signal are extracted; Based on the shape features of the second sensing signal, reference shape features are determined; the shape features of the first sensing signal are compared with the reference shape features to obtain a comparison result; According to the comparison result, the time features of the first sensing signal are corrected, or the proportion of the time features of the first sensing signal in calculation is modified; According to the modified time features of the first sensing signal or the proportion of the time features of the first sensing signal in calculation, the determination process of the disturbance compensation amount and the determination process of the drift compensation amount are re-performed; According to the re-determined disturbance compensation amount and drift compensation amount, the driving parameters or sensor interpretation logic of the fluid sampling unit are differentially compensated and adjusted to complete dynamic calibration of fluid sampling.

3. The sensor-based dynamic calibration method for fluid loading according to claim 1, wherein, The step of obtaining the first sensing signal output by the transition fluid disturbance observation stage after performing the calibration operation on the bioreactor by controlling the sensor of the fluid sampling unit comprises the following steps: After completing the calibration operation on the bioreactor, a standard disturbance liquid is injected into the bioreactor inlet of the fluid sampling unit to form a transition fluid disturbance observation stage; During the transition fluid disturbance observation stage, continuous output data are collected by the pressure sensor and the flow rate sensor of the fluid sampling channel, and the data are subjected to time series filtering and window smoothing processing to obtain the first sensing signal.

4. The sensor-based dynamic calibration method for fluid loading according to claim 2, wherein, The step of correcting the time features of the first sensing signal or modifying the proportion of the time features of the first sensing signal in calculation according to the comparison result comprises the following steps: Interference mode features are extracted from the shape features of the first sensing signal; The interference pattern feature is compared with a preset reference pattern to determine whether the change in the morphology feature is due to the performed calibration operation; The interference pattern feature is compared with a preset reference pattern to determine whether the change in the morphology feature is due to the performed calibration operation refers to comparing the similarity or difference between the interference pattern feature and the reference pattern, and determining whether the current change in the morphology feature is consistent with the typical change caused by the calibration operation by setting a threshold or classifier; When it is determined that the change in the morphology feature is due to the performed calibration operation, the time feature of the first sensing signal is corrected, or the proportion of the time feature of the first sensing signal in the calculation is modified; and the interference pattern feature corresponding to the calibration operation is obtained as a deviation profile; According to the deviation profile, the reference pattern is updated; The interference pattern feature refers to a pattern reflecting the signal change caused by the calibration operation extracted from the morphology feature of the first sensing signal.

5. The sensor-based dynamic calibration method for fluid loading according to claim 4, wherein, The step of updating the reference pattern according to the deviation profile comprises: The deviation profile is combined with the reference pattern before updating to generate the updated reference pattern.

6. A sensor-based fluid loading dynamic calibration system for performing sensor-based fluid loading dynamic calibration, the system comprising: Comprise: A first sensing signal acquisition module is configured to acquire a first sensing signal output in a transition fluid disturbance observation stage by controlling a sensor of a fluid loading unit after performing a calibration operation on a bioreactor; A second sensing signal acquisition module is configured to acquire a second sensing signal output in a stable sample physical property evaluation stage by controlling the sensor of the fluid loading unit; A disturbance compensation amount acquisition module is configured to compare the features of the first sensing signal and the second sensing signal, and determine a disturbance compensation amount according to the comparison result; A drift compensation amount acquisition module is configured to compare the features of the second sensing signal and a historical reference sensing signal, and determine a drift compensation amount according to the comparison result; A dynamic calibration execution module is configured to differentially compensate and adjust the driving parameters or sensor interpretation logic of the fluid loading unit according to the disturbance compensation amount and the drift compensation amount, so as to complete the dynamic calibration of the fluid loading. The transition fluid disturbance observation stage refers to a period when the fluid state near the sampling port changes temporarily due to the interaction between the calibration liquid and the fluid in the bioreactor after the completion of the calibration operation; The stable sample physical property evaluation stage refers to a period when the fluid state in the main area of the bioreactor, which is representative of the fluid state not directly affected by the calibration operation, is obtained after the influence of the transition fluid disturbance subsides.

Images