Microfluidic chip for early diagnosis of sepsis and evaluation of mitochondrial damage based on PCT-IL6 rapid detection

Abstract

The invention discloses a micro-fluidic chip for early diagnosis of sepsis and evaluation of mitochondrial damage based on PCT-IL6 rapid detection, and relates to the technical field of in-vitro diagnosis. Comprising an initial flow velocity control unit, a monitoring point setting and data acquisition unit, a feature extraction and input feature set construction unit, a deep learning model analysis unit, a flow period division unit, an efficient flow velocity control unit and a turbulence period flow velocity regulation and control unit. And the initial flow speed control unit is used for enabling the trace liquid to flow in the micron-sized channel by adopting the initially set maximum flow speed when the detection is started. By collecting flow information in real time and using a deep learning model to identify turbulence, the system can intelligently divide a flow time period, maintain an efficient flow speed in a normal time period, and dynamically regulate and control the flow speed in a turbulence time period to avoid bubbles and blockage, such accurate management improves the mixing uniformity and reaction integrity of a sample and a reagent, reduces misdiagnosis risks, and improves the detection accuracy. Reliability of detection results is ensured, and timeliness of patient diagnosis is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of in vitro diagnosis, and in particular to a microfluidic chip for early diagnosis of sepsis and evaluation of mitochondrial damage based on PCT-IL6 rapid detection. Background Art

[0002] The PCT-IL6 rapid test microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage is an innovative in vitro diagnostic device. Its core is to achieve rapid screening and early warning of sepsis by detecting interleukin-6 (IL-6) levels and mitochondrial damage markers in the blood. Microfluidic chip technology reduces sample requirements and speeds up analysis. At the same time, the multi-channel integrated design supports the simultaneous detection of IL-6 and multiple mitochondrial damage parameters, improving the accuracy and efficiency of diagnosis. The chip helps clinicians to understand the patient's condition in a timely manner and improve treatment strategies, thereby reducing the mortality rate and risk of complications of sepsis.

[0003] Microfluidic chips rely on precise control of the flow of trace amounts of liquid in sepsis detection, as their core function is to perform biochemical reactions and analyses in micron-scale channels. This requires the liquid to flow under controlled conditions to ensure that the reactants are accurately mixed, reacted, and detected at a specified time and location. To achieve this, the detection process in the prior art usually uses a preset flow rate to control the flow of liquid to maintain a constant delivery rate. Such precise control helps avoid reaction times that are too long or too short, prevents incomplete or over-reactions, ensures the stability, repeatability, and accuracy of the analysis results, and improves the reliability and consistency of the detection process.

[0004] The prior art has the following deficiencies:

[0005] When conducting sepsis testing, in order to improve the detection efficiency, the existing technology usually tries to increase the flow rate of trace liquids as much as possible while ensuring accurate detection. However, if the trace liquid generates turbulence due to the fast flow rate in the micron-scale channel, the turbulence will greatly increase the probability of bubble generation, because the turbulence makes the flow of the liquid no longer smooth, accompanied by chaotic and irregular flow states. In this environment, air is more likely to be drawn into the liquid to form bubbles. If the flow of trace liquids continues to be controlled at this flow rate to generate bubbles, the bubbles will block the micron-scale channel or interfere with the normal flow of the liquid, resulting in uneven mixing of samples and reagents and incomplete reactions. This will lead to false negative or false positive results, so that patients may not receive timely and correct diagnosis, and inaccurate test results will delay the timely diagnosis and treatment of sepsis, which is a rapidly progressive and fatal disease. Delays in treatment will greatly increase the risk of worsening of the disease, increase the mortality rate of patients and the incidence of serious complications.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0007] The purpose of the present invention is to provide a microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection. By collecting the flow information of trace liquid in real time during the detection process and using a deep learning model to identify and analyze potential turbulence, the system can intelligently divide the flow process into a turbulence period and a normal period. For the normal period, maintaining the initial maximum flow rate ensures an efficient detection speed. During the turbulence period, the flow rate is dynamically controlled to avoid problems such as bubble generation and channel blockage caused by turbulence. This precise flow rate management and turbulence control mechanism not only improves the mixing uniformity and reaction integrity of samples and reagents in micron-level channels, but also reduces the probability of false negative and false positive results, ensuring the reliability of the test results and the timeliness of patient diagnosis, thereby effectively reducing the risk of misdiagnosis and delayed treatment of sepsis, and ultimately helping to improve the survival rate and treatment effect of patients, so as to solve the problems in the above-mentioned background technology.

[0008] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: a microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection, comprising an initial flow rate control unit, a monitoring point setting and data acquisition unit, a feature extraction and input feature set construction unit, a deep learning model analysis unit, a flow period division unit, an efficient flow rate control unit, and a turbulent period flow rate control unit;

[0009] An initial flow rate control unit, at the beginning of the detection, uses an initially set maximum flow rate to make a trace amount of liquid flow in the micrometer-sized channel;

[0010] The monitoring point setting and data acquisition unit sets monitoring points at locations where turbulence is likely to occur in the micron-level channel to collect flow information of trace liquids in real time;

[0011] The feature extraction and input feature set construction unit extracts key features that can reflect the occurrence of turbulence from the real-time data collected at the monitoring point, analyzes the extracted features under the monitoring window, and establishes an input feature set to provide high-quality data input for the subsequent deep learning model;

[0012] A deep learning model analysis unit inputs the analyzed features in the input feature set into a pre-trained deep learning model, and identifies potential turbulence phenomena in the micron-scale channel through the deep learning model;

[0013] The flow period division unit divides the liquid flow process under the monitoring window into a turbulent period and a normal period based on the analysis results of the deep learning model;

[0014] The efficient flow rate control unit controls the flow of trace liquid in the micron-scale channel at the initially set maximum flow rate according to the flow of trace liquid in normal time periods, so as to maintain high detection efficiency;

[0015] The flow rate control unit during the turbulence period adjusts the initially set maximum flow rate based on the trace liquid flow during the turbulence period through the real-time turbulence conditions of the trace liquid flow predicted under the monitoring window, and controls the flow of trace liquid in the micron-level channel with the actual flow rate after adjustment to avoid the generation of bubbles and improve the detection accuracy.

[0016] Preferably, the initially set maximum flow rate is precisely set according to the structural design of the microfluidic chip, the channel size, the physical properties of the sample and the reagent, and the required biochemical reaction time; its setting should meet the following requirements:

[0017] Ensure that samples and reagents are fully mixed and reacted within the specified time to achieve an efficient testing process;

[0018] Maintain uniform flow of samples and reagents to prevent low flow rates from causing prolonged reaction time and reduced detection efficiency;

[0019] The flow rate setting needs to match the response speed and control capability of the detection equipment to ensure the stability and consistency of the detection, meet the clinical needs for fast and reliable detection, and optimize the detection process.

[0020] Preferably, in a micron-sized channel, locations where turbulence is likely to occur generally include sharp turns of the channel, contraction or expansion locations, areas where geometric structures suddenly change, and any areas where obstacles or foreign matter exist.

[0021] Preferably, key features that can reflect the occurrence of turbulence are extracted from the real-time data collected at the monitoring point, wherein the extracted key features include the local pressure change of the trace liquid in the micron-sized channel and the velocity fluctuation in the flow of the trace liquid. After extraction, the local pressure change of the trace liquid in the micron-sized channel and the velocity fluctuation in the flow of the trace liquid are analyzed under the monitoring window, and a pressure fluctuation quantification factor and a flow velocity disturbance quantification factor are generated respectively, the frequency of the local pressure change of the trace liquid in the micron-sized channel is quantified by the pressure fluctuation quantification factor, and the amplitude and frequency of the velocity fluctuation in the flow of the trace liquid are quantified by the flow velocity disturbance quantification factor;

[0022] After obtaining the pressure fluctuation quantification factor and flow velocity disturbance quantification factor generated by analyzing the trace liquid flow information, the pressure fluctuation quantification factor and the flow velocity disturbance quantification factor are used to establish an input feature set, and the analyzed features in the input feature set are input into a pre-trained deep learning model to generate a turbulence generation coefficient. The potential turbulence phenomenon in the micron-scale channel is identified through the turbulence generation coefficient.

[0023] Preferably, the turbulence generation coefficient generated when predicting the flow of a trace amount of liquid in a micron-sized channel by a pre-trained deep learning model is compared and analyzed with a pre-set turbulence generation coefficient reference threshold, and the liquid flow process under the monitoring window is divided. The specific division steps are as follows:

[0024] If the turbulence generation coefficient is greater than the turbulence generation coefficient reference threshold, the liquid flow process under the monitoring window is divided into a turbulence generation period;

[0025] If the turbulence generation coefficient is less than or equal to the turbulence generation coefficient reference threshold, the liquid flow process under the monitoring window is divided into a normal period.

[0026] Preferably, under the monitoring window, the specific steps of generating a pressure fluctuation quantification factor after analyzing the local pressure change of a trace amount of liquid in a micron-sized channel are as follows:

[0027] Under the monitoring window, the local pressure change data of trace liquid in the micron-scale channel is collected in real time, and the time series t1, t2, ..., t n Indicates that at the moment of collection within the monitoring window, the collected instantaneous pressure values ​​are P(t1), P(t2), ..., P(t n ), on this basis, calculate the instantaneous pressure change rate ΔP at every two consecutive time points i , expressed as:

[0028] ΔP i =|P(t i+1 )-P(t i )| / (t i+1 -t i )

[0029] Where i = 1, 2, ..., n-1, i represents the index of the sampling time point, and n represents the total number of sampling time points in the monitoring window;

[0030] Introducing nonlinear variation factor φ i , Where k is an adjustment parameter used to amplify small pressure changes in order to have a higher sensitivity to turbulence precursors, and finally obtain a change sequence Φ = {φ1, φ2, ..., φ n-1}, used for subsequent calculations;

[0031] Calculation of pressure fluctuation quantification factor PWI through the change sequence x , the pressure fluctuation quantization factor is defined as:

[0032]

[0033] , where α and β are adjustment parameters used to control the sensitivity of the fluctuation index to different pressure change rates.

[0034] Preferably, in the monitoring window, the specific steps of generating the flow velocity disturbance quantification factor after analyzing the velocity fluctuation in the flow of a trace amount of liquid are as follows:

[0035] The real-time velocity data collected from the monitoring point is recorded as v(t), where t represents time. The root mean square value of the real-time velocity fluctuation is calculated within the monitoring window to quantify the velocity fluctuation amplitude. The calculation expression is:

[0036]

[0037] in, is the average flow rate in the monitoring window, T represents the total duration of the monitoring window, and RMS(v) represents the root mean square value of the real-time flow rate fluctuation;

[0038] The velocity disturbance quantification factor D is calculated using the RMS value of real-time velocity fluctuations. index , quantify the intensity and frequency of flow velocity changes, flow velocity disturbance quantification factor D index The calculation expression is:

[0039]

[0040] In the formula, f v It is the fluctuation frequency obtained by counting the zero-crossing points of flow velocity changes within the monitoring window, reflecting the frequency of flow velocity changes.

[0041] Preferably, according to the flow of the trace liquid during the turbulence period, the initially set maximum flow rate is regulated, and the trace liquid is controlled to flow in the micron-sized channel at the actual flow rate after regulation. The specific steps are as follows:

[0042] During the detection process, the deep learning model is used to analyze the flow information in the monitoring window in real time to generate the current turbulence generation coefficient Turbu Ge , turbulence generation coefficient Turbu Ge Used to quantify the possibility of turbulence occurring in a trace amount of liquid in a micrometer-sized channel;

[0043] Introduce the correction coefficient ψ to adjust the initial maximum flow rate V max , the calculation expression of the correction coefficient is:

[0044]

[0045] , where T ref is the reference threshold of turbulence generation coefficient, δ is an adjustment factor used to control the sensitivity of flow velocity adjustment;

[0046] Based on the calculated correction coefficient ψ, the actual flow rate V after regulation adjusted , actual flow rate V adjusted The calculation expression is: V adjusted =V max ·ψ, the actual flow rate after this regulation V adjusted Used to control the flow of trace liquids in micron-sized channels.

[0047] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0048] The present invention collects the flow information of trace liquid in real time during the detection process and uses a deep learning model to identify and analyze potential turbulence. The system can intelligently divide the flow process into a turbulence occurrence period and a normal period. For the normal period, maintaining the initially set maximum flow rate ensures an efficient detection speed. During the turbulence occurrence period, the flow rate is dynamically controlled to avoid problems such as bubble generation and channel blockage caused by turbulence. This precise flow rate management and turbulence control mechanism not only improves the mixing uniformity and reaction integrity of samples and reagents in the micron-level channel, but also reduces the probability of false negative and false positive results, ensuring the reliability of the test results and the timeliness of patient diagnosis, thereby effectively reducing the risk of misdiagnosis and delayed treatment of sepsis, and ultimately helping to improve the survival rate and treatment effect of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0050] Figure 1 Schematic diagram of the overall structural unit of the microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection of the present invention. DETAILED DESCRIPTION

[0051] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0052] The present invention provides Figure 1 The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection shown includes an initial flow rate control unit, a monitoring point setting and data acquisition unit, a feature extraction and input feature set construction unit, a deep learning model analysis unit, a flow period division unit, an efficient flow rate control unit, and a turbulent period flow rate control unit;

[0053] An initial flow rate control unit, which, at the beginning of the detection, uses an initially set maximum flow rate to make a trace amount of liquid (including biological samples and reagents) flow in the micron-scale channel;

[0054] The maximum flow rate initially set is to improve the detection efficiency while ensuring that the sample is fully tested. This high flow rate helps promote rapid mixing and efficient reaction of samples and reagents in the micron-scale channel, shortens the reaction time and accelerates the completion of biochemical reactions, improves the overall detection speed and throughput, and thus meets the clinical needs for rapid diagnosis and timely treatment. Fast flow can effectively reduce the residence time of samples and reagents in the channel, reduce analysis delays, and enable laboratories to process more samples in a shorter time, thereby significantly improving diagnostic efficiency and responsiveness. For patients with acute and critical illnesses such as sepsis, early detection and diagnosis are key. Shortening the detection time means that intervention and treatment can be carried out earlier, significantly improving the patient's prognosis and reducing mortality. In addition, the application of high flow rates can also optimize resource utilization, reduce manual operations and waiting time in the detection process, and ensure maximum utilization of laboratory resources.

[0055] The maximum flow rate set initially should be precisely set according to the structural design of the microfluidic chip, channel size, physical properties of samples and reagents (such as viscosity and density), and the required biochemical reaction time. Its setting should meet the following requirements: ensure that the sample and reagent can be fully mixed and reacted within the specified time to achieve an efficient detection process; at the same time, maintain a uniform flow of samples and reagents to prevent the flow rate from being too low, resulting in prolonged reaction time and reduced detection efficiency. The flow rate setting must also match the response speed and control capability of the detection equipment to ensure the stability and consistency of the detection, thereby meeting the clinical demand for rapid and reliable detection and optimizing the detection process.

[0056] The monitoring point setting and data acquisition unit sets monitoring points at locations where turbulence is likely to occur in the micron-level channel to collect flow information of trace liquids in real time;

[0057] The feature extraction and input feature set construction unit extracts key features that can reflect the occurrence of turbulence from the real-time data collected at the monitoring point, analyzes the extracted features under the monitoring window, and establishes an input feature set to provide high-quality data input for the subsequent deep learning model;

[0058] A deep learning model analysis unit inputs the analyzed features in the input feature set into a pre-trained deep learning model, and identifies potential turbulence phenomena in the micron-scale channel through the deep learning model;

[0059] In micron-scale channels, locations prone to turbulence usually include sharp turns, contraction or expansion locations, areas where the geometry changes suddenly (such as sudden narrowing or expansion), and any areas where obstacles or foreign matter exist. These locations are prone to significant changes in liquid flow rate due to sudden changes in the flow path or changes in the flow cross-sectional area, resulting in local pressure gradients and eddies, thereby triggering turbulence. In addition, areas with high surface roughness in the channel may also exacerbate flow instability due to friction, increasing the probability of turbulence formation.

[0060] There are many ways to collect flow information of trace amounts of liquid in micron-scale channels in real time, such as optical methods, pressure sensors, and micro-flow sensors. Optical methods include the use of microparticle imaging velocimetry (μPIV) technology to measure flow velocity and flow pattern by laser irradiation and microscopic observation of the movement of tracer particles in the liquid. Pressure sensors can be installed on the channel wall to detect pressure changes generated during the flow process, helping to identify turbulence and flow velocity fluctuations. Micro-flow sensors can directly sense the flow velocity and direction of the liquid, providing real-time and accurate data. In addition, the data acquisition system that combines these methods can efficiently integrate information for real-time flow state analysis.

[0061] From the real-time data collected at the monitoring points, key features that can reflect the occurrence of turbulence are extracted, wherein the extracted key features include the local pressure changes of trace liquid in the micron-sized channel and the velocity fluctuations in the flow of trace liquid. After extraction, under the monitoring window, the local pressure changes of trace liquid in the micron-sized channel and the velocity fluctuations in the flow of trace liquid are analyzed, and a pressure fluctuation quantification factor and a flow velocity disturbance quantification factor are generated respectively. The frequency of local pressure changes of trace liquid in the micron-sized channel is quantified by the pressure fluctuation quantification factor, reflecting the pressure instability caused by turbulence or other disturbances in the channel. The amplitude and frequency of velocity fluctuations in the flow of trace liquid are quantified by the flow velocity disturbance quantification factor, showing the severity and stability of the liquid flow velocity changes.

[0062] After obtaining the pressure fluctuation quantification factor and flow velocity disturbance quantification factor generated by analyzing the trace liquid flow information, the pressure fluctuation quantification factor and the flow velocity disturbance quantification factor are used to establish an input feature set, and the analyzed features in the input feature set are input into a pre-trained deep learning model to generate a turbulence generation coefficient. The potential turbulence phenomenon in the micron-scale channel is identified through the turbulence generation coefficient.

[0063] A pre-trained deep learning model is used to analyze and identify potential turbulent phenomena in micron-scale channels. This model can be a deep learning architecture based on a convolutional neural network (CNN) or a long short-term memory network (LSTM). CNN is a model structure that is particularly effective in processing images and spatiotemporal data. It can identify complex patterns by extracting local features of the input data, so it is very suitable for analyzing the flow characteristics of trace amounts of liquid in micron-scale channels. The CNN model can use multiple convolutional layers and pooling layers to gradually extract and compress the spatial features of flow velocity and pressure data. Through training, the CNN model is able to learn specific features related to turbulence, such as unstable patterns and local abnormal changes in flow, which is crucial for identifying turbulent phenomena.

[0064] On the other hand, the LSTM model performs well in processing time series data and can capture data dependencies and sequence information over long time spans. In the analysis of flow data in micron-scale channels, LSTM can help identify the dynamic changes and trends of liquid flow information. This is because LSTM can effectively retain important historical information through its unique memory unit structure and use this information to predict the current state. Combined with the LSTM model, the system can analyze the historical patterns of pressure fluctuations and velocity fluctuations in real-time data collected from monitoring points to identify potential periods of turbulence. The memory capacity of the LSTM network ensures that the system can adapt to and predict flow changes in real time, even the subtlest precursors to turbulence. In this way, the generated turbulence generation coefficient can accurately quantify the turbulence risk of the flow state.

[0065] The pre-trained deep learning model combines the advantages of CNN and LSTM, and can analyze both spatial features and time series features. With this hybrid model, the system can not only identify the current flow state, but also predict the turbulence that may occur in the next few seconds. This combination has efficient and robust recognition capabilities in complex flow environments, thereby significantly improving the stability and reliability of microfluidic detection. The model needs to use a large amount of experimental data with turbulence annotations during the training phase to ensure that it can quickly and accurately identify and quantify turbulence phenomena in micron-scale channels during actual detection.

[0066] The high frequency of local pressure changes of trace amounts of liquid in micron-sized channels does indicate that there may be potential turbulence in micron-sized channels. The reason is that turbulence is caused by the instability of liquid flow, which is characterized by random fluctuations of the fluid in space and time. Frequent changes in local pressure reflect the rapid changes in speed and direction of the fluid in a small range, which is one of the typical manifestations of turbulence. This pressure fluctuation will cause the liquid to form small eddies and irregular flows in the channel, hindering the normal laminar state, making the entire flow more unstable.

[0067] Under the monitoring window, the specific steps of analyzing the local pressure change of a trace amount of liquid in a micron-sized channel to generate a pressure fluctuation quantification factor are as follows:

[0068] Under the monitoring window, the local pressure change data of trace liquid in the micron-scale channel is collected in real time, and the time series t1, t2, ..., t n Indicates that at the moment of collection within the monitoring window, the collected instantaneous pressure values ​​are P(t1), P(t2), ..., P(t n ), on this basis, calculate the instantaneous pressure change rate ΔP at every two consecutive time points i , expressed as:

[0069] ΔP i =|P(t i+1 )-P(t i )| / (t i+1 -t i )

[0070] , where i = 1, 2, ..., n-1, i represents the index of the sampling time point, and n represents the total number of sampling time points in the monitoring window;

[0071] The instantaneous pressure change rate reflects the severity of the instantaneous pressure of the liquid in the local area.

[0072] In order to extract more details, a nonlinear change factor φ is introduced i , Where k is an adjustment parameter used to amplify small pressure changes in order to have a higher sensitivity to turbulence precursors, and finally obtain a change sequence Φ = {φ1, φ2, ..., φ n-1}, used for subsequent calculations;

[0073] The role of introducing nonlinear variation factors is to amplify small but potentially important pressure changes when extracting details of pressure changes, so as to improve the sensitivity of detection to turbulence precursors. By using nonlinear variation factors, such as exponential functions or other nonlinear transformations, the response to subtle fluctuations in the pressure change rate that are not obvious but have potential effects can be enhanced. In this way, when identifying turbulence, the system can more accurately capture those subtle fluctuations that may indicate the formation of turbulence, making the overall detection more sensitive and comprehensive, and avoiding missing key pressure change information.

[0074] Calculation of pressure fluctuation quantification factor PWI through the change sequence x , the pressure fluctuation quantization factor is defined as:

[0075]

[0076] , where α and β are adjustment parameters used to control the sensitivity of the fluctuation index to different pressure change rates;

[0077] This formula scales and accumulates each instantaneous rate of change of pressure in a nonlinear form, reflecting the frequency of overall pressure fluctuations. The nonlinear scaling term here is It can effectively amplify the contribution of abnormal pressure changes and weaken relatively stable changes. This nonlinear formula structure ensures that the pressure fluctuation quantification factor can produce a high response sensitivity to turbulence phenomena, which facilitates further identification of potential turbulence risks.

[0078] It can be seen from the pressure fluctuation quantification factor that, under the monitoring window, the larger the performance value of the pressure fluctuation quantification factor generated after analyzing the local pressure changes of trace liquids in the micron-level channel, the greater the hidden danger of potential turbulence in the micron-level channel. The pressure fluctuation quantification factor quantifies the frequency and intensity of local pressure changes of liquids in micron-level channels. When its performance value is high, it means that there are significant pressure instability and fluctuations in the liquid flow, which is a typical feature of turbulence or turbulence precursors. Conversely, when the value of the pressure fluctuation quantification factor is small, it means that the flow of the liquid is relatively stable, the local pressure changes are less, and the potential hidden dangers of turbulence are correspondingly reduced. Therefore, by monitoring the performance value of the pressure fluctuation quantification factor, the size of the turbulence risk in the channel can be effectively judged.

[0079] When a small amount of liquid flows in a micron-sized channel, large velocity fluctuations usually indicate the presence of potential turbulence. The reason is that turbulence is an unstable flow state, accompanied by random and violent fluctuations in velocity and pressure. Micron-sized channels are more likely to produce turbulence at high Reynolds numbers due to their small size and high flow rate. When the flow velocity fluctuates greatly, it means that the laminar state of the liquid in the channel is destroyed, and the flow begins to show irregular eddies and turbulence, which is one of the characteristics of turbulence. Therefore, the increase in the amplitude and frequency of velocity fluctuations is an important indicator for judging the potential turbulence in the channel.

[0080] Under the monitoring window, the specific steps of analyzing the velocity fluctuation in the flow of a trace amount of liquid and generating the flow velocity disturbance quantification factor are as follows:

[0081] The real-time velocity data collected from the monitoring point is recorded as v(t), where t represents time. The root mean square value (RMS) of the real-time velocity fluctuation is calculated within the monitoring window to quantify the velocity fluctuation amplitude. The calculation expression is:

[0082]

[0083] , where (i represents the index of the sampling time point, n represents the total number of sampling time points in the monitoring window), is the average flow rate in the monitoring window, T represents the total duration of the monitoring window, and RMS(v) represents the root mean square value of the real-time flow rate fluctuation;

[0084] The root mean square value of the real-time flow velocity fluctuation is used to capture the amplitude of the flow velocity change and reflect the fluctuation of the flow velocity. T represents the total duration of the monitoring window. The root mean square value of the real-time flow velocity fluctuation is used to further generate the flow velocity disturbance quantization factor.

[0085] The velocity disturbance quantification factor D is calculated using the RMS value of real-time velocity fluctuations. index , quantify the intensity and frequency of flow velocity changes, flow velocity disturbance quantification factor D index The calculation expression is:

[0086]

[0087] , where f v It is the fluctuation frequency obtained by counting the zero-crossing points of the flow rate changes within the monitoring window, reflecting the frequency of flow rate changes;

[0088] The frequency of fluctuation calculated by the number of zero crossings of the velocity change refers to the frequency of fluctuation calculated by counting the number of times the velocity signal crosses the zero line within a certain period of time. In the velocity data, the zero crossing refers to the intersection point where the velocity change curve changes from positive to negative or from negative to positive. Each crossing of the zero point represents half of a complete fluctuation cycle.

[0089] For example, suppose that in a monitoring window, the flow rate signal crosses the zero point multiple times, which indicates that the flow rate has experienced multiple acceleration and deceleration changes during this period of time. By counting the number of such zero crossings and calculating the frequency based on the length of the monitoring time window, the frequency of flow rate fluctuations can be quantified.

[0090] Through this combination, the velocity disturbance quantification factor not only reflects the amplitude of the velocity fluctuation (RMS normalization), but also takes into account the dynamics and severity of the fluctuation through the frequency part. The logarithmic term of the velocity disturbance quantification factor ensures a moderate response to high-frequency and low-amplitude fluctuations, thereby better capturing potential turbulence phenomena.

[0091] It can be seen from the flow velocity disturbance quantification factor that, under the monitoring window, the larger the performance value of the flow velocity disturbance quantification factor generated after analyzing the velocity fluctuation in the flow of trace liquid, the greater the hidden danger of potential turbulence in the micron-level channel. The flow velocity disturbance quantification factor reflects the instability and intensity of liquid flow by quantifying the amplitude and frequency of flow velocity fluctuations. When the value of the flow velocity disturbance quantification factor is high, it means that the flow velocity of the liquid in the channel changes frequently and violently, which is usually a signal that turbulence is about to occur or has occurred. On the contrary, when the flow velocity disturbance quantification factor is low, it indicates that the flow velocity fluctuation is small, the stability is high, and the risk of turbulence in the micron-level channel is low. Therefore, the flow velocity disturbance quantification factor is an important parameter for judging whether the flow state is close to turbulence.

[0092] The above deep learning model is not specifically limited here, and can realize the pressure fluctuation quantification factor PWI x and the velocity disturbance quantification factor D index Perform comprehensive analysis to generate turbulence generation coefficient Turbu Ge In order to realize the technical solution of the present invention, the present invention provides a specific implementation method; the turbulence generation coefficient Turbu Ge The resulting calculation formula is:

[0093]

[0094] , where d1 and d2 are pressure fluctuation quantification factors PWI respectively. x and the velocity disturbance quantification factor D index The preset proportional coefficient, and d1 and d2 are both greater than 0.

[0095] The "preset proportionality coefficient" refers to the constant parameter used in the calculation formula to adjust and weigh the impact of different indicators on the results. In the formula, d1 and d2 are the pressure fluctuation quantification factor PWI, respectively. x and the velocity disturbance quantification factor D index The preset proportional coefficients determine the effect of these two indicators in calculating the turbulence generation coefficient Cirl ac The relative importance and weight of the model in the equations. By adjusting these coefficients, the response of the model can be optimized to better suit specific application scenarios and flow characteristics. The preset proportional coefficients d1 and d2 are usually determined during system design or experiments to ensure that the model has optimal performance under different conditions.

[0096] It can be seen from the turbulence generation coefficient that, under the monitoring window, the larger the performance value of the pressure fluctuation quantification factor generated after analyzing the local pressure changes of trace liquid in the micron-level channel, the larger the performance value of the flow velocity disturbance quantification factor generated after analyzing the velocity fluctuation in the flow of trace liquid. That is, the larger the performance value of the turbulence generation coefficient generated when predicting the flow conditions of trace liquid in the micron-level channel by the pre-trained deep learning model, the greater the hidden danger of potential turbulence in the micron-level channel, and vice versa.

[0097] The flow period division unit divides the liquid flow process under the monitoring window into a turbulent period and a normal period based on the analysis results of the deep learning model;

[0098] The turbulence generation coefficient generated when predicting the flow of trace liquid in a micron-sized channel through a pre-trained deep learning model is compared and analyzed with the pre-set turbulence generation coefficient reference threshold, and the liquid flow process under the monitoring window is divided. The specific division steps are as follows:

[0099] If the turbulence generation coefficient is greater than the turbulence generation coefficient reference threshold, the liquid flow process under the monitoring window is divided into the turbulence occurrence period, which refers to the period of instability and turbulence in the liquid flow process in the micron-level channel. During this period, the fluid will show chaotic, irregular eddy and turbulent phenomena, and the flow rate and pressure will change frequently and irregularly. The occurrence of turbulence usually means that the liquid cannot maintain a stable laminar state in the channel, which may lead to deviations or inaccuracies in the test results, and even cause channel blockage and bubble generation, thereby affecting the reliability of the test.

[0100] If the turbulence generation coefficient is less than or equal to the turbulence generation coefficient reference threshold, the liquid flow process under the monitoring window is divided into a normal period. The normal period refers to the period when the liquid flows smoothly and stably in the micron-level channel. The flow state remains in a controllable laminar state during this period, and there is no obvious turbulence and turbulence. During the normal period, the velocity and pressure changes of the liquid are within the set safety range, reflecting the stability of the fluid flow. The flow state at this time helps to ensure that the sample and reagent are evenly mixed in the channel, and the accuracy and reliability of the test results can be guaranteed.

[0101] The efficient flow rate control unit controls the flow of trace liquid in the micron-scale channel at the initially set maximum flow rate according to the flow of trace liquid in normal time periods, so as to maintain high detection efficiency;

[0102] During the detection process, if it is determined through analysis that the liquid flow state in the current monitoring window belongs to the normal period, that is, there is no turbulence or instability, then the system will maintain the initial maximum flow rate to control the flow of trace liquid in the micron-level channel. This operation mode is designed to maximize the flow rate to improve the detection efficiency without sacrificing flow stability and detection accuracy, making the detection process faster, thereby speeding up sample processing and the output of test results. This helps to optimize the detection speed, which is especially important when rapid detection and high-throughput analysis are required.

[0103] The flow rate control unit during the turbulence period controls the initial maximum flow rate according to the real-time turbulence of the trace liquid flow predicted under the monitoring window, and controls the flow of the trace liquid in the micron-level channel with the actual flow rate after control, so as to avoid the generation of bubbles and improve the detection accuracy.

[0104] According to the flow of trace liquid during the turbulent period, the initially set maximum flow rate is adjusted, and the actual flow rate after adjustment is used to control the flow of trace liquid in the micron-sized channel. The specific steps are as follows:

[0105] During the detection process, the deep learning model is used to analyze the flow information in the monitoring window in real time to generate the current turbulence generation coefficient Turbu Ge , turbulence generation coefficient Turbu Ge Used to quantify the possibility of turbulence occurring in a trace amount of liquid in a micrometer-sized channel;

[0106] If the turbulence generation coefficient is greater than the turbulence generation coefficient reference threshold, it means that the current flow has the risk of turbulence, and the initial maximum flow velocity V needs to be adjusted. max This monitoring and identification step ensures that the system can respond quickly and is ready to adjust flow rate if a risk of turbulence arises.

[0107] Introduce the correction coefficient ψ to adjust the initial maximum flow rate V max , the calculation expression of the correction coefficient is:

[0108]

[0109] , where T ref is the reference threshold of the turbulence generation coefficient, δ is an adjustment factor used to control the sensitivity of flow velocity adjustment, which is usually set according to practical experience;

[0110] This formula ensures that when the turbulence generation coefficient Turbu Ge Close to the reference threshold value T of turbulence generation coefficient ref When the correction coefficient ψ is close to 1, the adjustment is small; when the turbulence generation coefficient Turbu Ge Much higher than the reference threshold value of turbulence generation coefficient T ref , the correction factor ψ will decrease significantly, indicating that the flow rate needs to be greatly reduced.

[0111] Based on the calculated correction coefficient ψ, the actual flow rate V after regulation adjusted , actual flow rate V adjusted The calculation expression is: V adjusted =V max ·ψ, the actual flow rate after this regulation V adjusted Used to control the flow of trace liquids in micron-scale channels, ensuring that the turbulence risk is reduced to a controllable range, avoiding the generation of bubbles and improving detection accuracy;

[0112] By using the actual flow rate V adjusted Instead of the initial maximum flow rate V max The system dynamically adjusts the flow rate during the turbulence period to maintain a stable flow state. This can minimize the instability caused by turbulence and ensure the reliability and accuracy of the test results.

[0113] The present invention collects the flow information of trace liquid in real time during the detection process and uses a deep learning model to identify and analyze potential turbulence. The system can intelligently divide the flow process into a turbulence occurrence period and a normal period. For the normal period, maintaining the initially set maximum flow rate ensures an efficient detection speed. During the turbulence occurrence period, the flow rate is dynamically controlled to avoid problems such as bubble generation and channel blockage caused by turbulence. This precise flow rate management and turbulence control mechanism not only improves the mixing uniformity and reaction integrity of samples and reagents in the micron-level channel, but also reduces the probability of false negative and false positive results, ensuring the reliability of the test results and the timeliness of patient diagnosis, thereby effectively reducing the risk of misdiagnosis and delayed treatment of sepsis, and ultimately helping to improve the survival rate and treatment effect of patients.

[0114] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0115] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0116] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid test, characterized in that: It includes an initial flow velocity control unit, a monitoring point setting and data collection unit, a feature extraction and input feature set construction unit, a deep learning model analysis unit, a flow period division unit, an efficient flow velocity control unit, and a turbulent period flow velocity control unit; An initial flow rate control unit, at the beginning of the detection, uses an initially set maximum flow rate to make a trace amount of liquid flow in the micrometer-sized channel; The monitoring point setting and data acquisition unit sets monitoring points at locations where turbulence is likely to occur in the micron-level channel to collect flow information of trace liquids in real time; The feature extraction and input feature set construction unit extracts key features that can reflect the occurrence of turbulence from the real-time data collected at the monitoring point, analyzes the extracted features under the monitoring window, and establishes an input feature set to provide high-quality data input for the subsequent deep learning model; A deep learning model analysis unit inputs the analyzed features in the input feature set into a pre-trained deep learning model, and identifies potential turbulence phenomena in the micron-scale channel through the deep learning model; The flow period division unit divides the liquid flow process under the monitoring window into a turbulent period and a normal period based on the analysis results of the deep learning model; The efficient flow rate control unit controls the flow of trace liquid in the micron-scale channel at the initially set maximum flow rate according to the flow of trace liquid in normal time periods, so as to maintain high detection efficiency; The flow rate control unit during the turbulence period adjusts the initially set maximum flow rate based on the trace liquid flow during the turbulence period through the real-time turbulence conditions of the trace liquid flow predicted under the monitoring window, and controls the flow of trace liquid in the micron-level channel with the actual flow rate after adjustment to avoid the generation of bubbles and improve the detection accuracy.

2. The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection according to claim 1, characterized in that: The maximum flow rate initially set is precisely set according to the structural design of the microfluidic chip, channel size, physical properties of the sample and reagent, and the required biochemical reaction time; its setting should meet the following requirements: Ensure that samples and reagents are fully mixed and reacted within the specified time to achieve an efficient testing process; Maintain uniform flow of samples and reagents to prevent low flow rates from causing prolonged reaction time and reduced detection efficiency; The flow rate setting needs to match the response speed and control capability of the detection equipment to ensure the stability and consistency of the detection, meet the clinical needs for fast and reliable detection, and optimize the detection process.

3. The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection according to claim 1, characterized in that: In micron-scale channels, locations where turbulence is likely to occur typically include sharp turns in the channel, locations where the channel contracts or expands, areas where the geometry changes abruptly, and any areas where there are obstacles or foreign objects.

4. The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection according to claim 1, characterized in that: Extract key features that can reflect the occurrence of turbulence from real-time data collected at monitoring points, where the extracted key features include local pressure changes of trace liquid in micron-sized channels and velocity fluctuations in the flow of trace liquid. After extraction, analyze the local pressure changes of trace liquid in micron-sized channels and velocity fluctuations in the flow of trace liquid under the monitoring window, and generate pressure fluctuation quantification factors and flow velocity disturbance quantification factors, respectively. The frequency of local pressure changes of trace liquid in micron-sized channels is quantified by the pressure fluctuation quantification factors, and the amplitude and frequency of velocity fluctuations in the flow of trace liquid are quantified by the flow velocity disturbance quantification factors. After obtaining the pressure fluctuation quantification factor and flow velocity disturbance quantification factor generated by analyzing the trace liquid flow information, the pressure fluctuation quantification factor and the flow velocity disturbance quantification factor are used to establish an input feature set, and the analyzed features in the input feature set are input into a pre-trained deep learning model to generate a turbulence generation coefficient. The potential turbulence phenomenon in the micron-scale channel is identified through the turbulence generation coefficient.

5. The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection according to claim 4, characterized in that: The turbulence generation coefficient generated when predicting the flow of trace liquid in a micron-sized channel through a pre-trained deep learning model is compared and analyzed with the pre-set turbulence generation coefficient reference threshold, and the liquid flow process under the monitoring window is divided. The specific division steps are as follows: If the turbulence generation coefficient is greater than the turbulence generation coefficient reference threshold, the liquid flow process under the monitoring window is divided into a turbulence generation period; If the turbulence generation coefficient is less than or equal to the turbulence generation coefficient reference threshold, the liquid flow process under the monitoring window is divided into a normal period.

6. The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection according to claim 4, characterized in that: Under the monitoring window, the specific steps of analyzing the local pressure change of a trace amount of liquid in a micron-sized channel to generate a pressure fluctuation quantification factor are as follows: Under the monitoring window, the local pressure change data of trace liquid in the micron-scale channel is collected in real time, and the time series t1, t2, ..., t n Indicates that at the moment of collection within the monitoring window, the collected instantaneous pressure values ​​are P(t1), P(t2), ..., P(t n ), on this basis, calculate the instantaneous pressure change rate ΔP at every two consecutive time points i , expressed as: ΔP i =|P(t i+1 )-P(t i )| / (t i+1 -t i ) Where i = 1, 2, ..., n-1, i represents the index of the sampling time point, and n represents the total number of sampling time points in the monitoring window; Introducing nonlinear variation factor φ i , Where k is an adjustment parameter used to amplify small pressure changes in order to have a higher sensitivity to turbulence precursors, and finally obtain a change sequence Φ = {φ1, φ2, ..., φ n-1 }, used for subsequent calculations; Calculation of pressure fluctuation quantification factor PWI through the change sequence x , the pressure fluctuation quantization factor is defined as: Among them, α and β are adjustment parameters used to control the sensitivity of the fluctuation index to different pressure change rates.

7. The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection according to claim 6, characterized in that: Under the monitoring window, the specific steps of analyzing the velocity fluctuation in the flow of a trace amount of liquid and generating the flow velocity disturbance quantification factor are as follows: The real-time velocity data collected from the monitoring point is recorded as v(t), where t represents time. The root mean square value of the real-time velocity fluctuation is calculated within the monitoring window to quantify the velocity fluctuation amplitude. The calculation expression is: in, is the average flow rate in the monitoring window, T represents the total duration of the monitoring window, and RMS(v) represents the root mean square value of the real-time flow rate fluctuation; The velocity disturbance quantification factor D is calculated using the RMS value of real-time velocity fluctuations. index , quantify the intensity and frequency of flow velocity changes, flow velocity disturbance quantification factor D index The calculation expression is: , where f v It is the fluctuation frequency obtained by counting the zero-crossing points of flow velocity changes within the monitoring window, reflecting the frequency of flow velocity changes.

8. The microfluidic chip for early diagnosis of sepsis and assessment of mitochondrial damage based on PCT-IL6 rapid detection according to claim 5, characterized in that: According to the flow of trace liquid during the turbulent period, the initially set maximum flow rate is adjusted, and the actual flow rate after adjustment is used to control the flow of trace liquid in the micron-sized channel. The specific steps are as follows: During the detection process, the deep learning model is used to analyze the flow information in the monitoring window in real time to generate the current turbulence generation coefficient Turbu Ge , turbulence generation coefficient Turbu Ge Used to quantify the possibility of turbulence occurring in a trace amount of liquid in a micrometer-sized channel; Introduce the correction coefficient ψ to adjust the initial maximum flow rate V max , the calculation expression of the correction coefficient is: Where, T ref is the reference threshold of turbulence generation coefficient, δ is an adjustment factor used to control the sensitivity of flow velocity adjustment; Based on the calculated correction coefficient ψ, the actual flow rate V after regulation adjusted , actual flow rate V adjusted The calculation expression is: V adjusted =V max ·ψ, the actual flow rate after this regulation V adjusted Used to control the flow of trace liquids in micron-sized channels.

Images