Online Detection System for Anesthetic Gas Concentration for Precise Dosing Control

Abstract

This invention discloses an online detection system for anesthetic gas concentration for precise dosage control, relating to the field of anesthetic gas concentration measurement technology. It addresses the technical problem of inaccurate measurement results during online detection of anesthetic gas concentration. The system includes: a data acquisition module for acquiring infrared absorption spectral data; an interference identification module for determining the start time of alcohol disinfection based on the temporal changes in the infrared absorption spectral data; a signal separation module for separating the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectral data and constructing an alcohol absorbance change model based on these values; the signal separation module also determines the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths from the infrared absorption spectral data; and a concentration calculation module for determining the concentration of the target anesthetic gas based on its absorbance values ​​at multiple preset wavelengths.

Description

Technical Field

[0001] This invention relates to the field of anesthetic gas concentration measurement technology, and specifically to an online anesthetic gas concentration detection system for precise dosage control. Background Technology

[0002] Online detection of anesthetic gas concentration is a crucial step in achieving precise drug administration during modern general anesthesia. By monitoring the patient's end-tidal gas concentration in real time, doctors can accurately control the depth of anesthesia, avoiding risks such as intraoperative awareness or excessive anesthesia. Currently, infrared spectroscopy-based gas concentration detection technology is widely used in operating rooms. This technology, based on Lambert-Beer's law, inverts the concentration by measuring the characteristic absorption of anesthetic gas molecules at specific wavelengths. However, in actual clinical environments, the vapor emitted by alcohol, widely used for surgical disinfection, can interfere with the detection process. Because alcohol and anesthetic molecules have overlapping absorption spectra in the infrared band, existing detection systems struggle to effectively distinguish the signal source, significantly reducing the accuracy and reliability of concentration detection results and directly affecting the precise control of anesthetic drug administration. Summary of the Invention

[0003] To address the technical problem of inaccurate measurement results in online anesthetic gas concentration detection caused by the mutual interference of the infrared absorption spectra of alcohol vapor and anesthetic gas, the present invention aims to provide an online anesthetic gas concentration detection system for precise dosage control. The specific technical solution adopted is as follows:

[0004] In a first aspect, the present invention provides an online detection system for anesthetic gas concentration for precise dosage control, comprising: a data acquisition module, an interference identification module, a signal separation module, and a concentration calculation module; the data acquisition module is used to acquire infrared absorption spectral data; wherein, the infrared absorption spectral data is used to characterize the absorbance values ​​of a patient's end-tidal gas sample at multiple preset wavelengths; the interference identification module is used to determine the start time of alcohol disinfection based on the temporal changes of the infrared absorption spectral data; wherein, the start time of alcohol disinfection is used to characterize the initial time point at which alcohol vapor begins to interfere with the detection of anesthetic gas concentration; the signal separation module is used to separate the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectral data based on the start time of alcohol disinfection, and to construct an alcohol absorbance change model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths; the signal separation module is also used to determine the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths from the infrared absorption spectral data based on the alcohol absorbance change model; the concentration calculation module is used to determine the concentration value of the target anesthetic gas based on the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths.

[0005] In one possible implementation, when the interference identification module determines the start time of alcohol disinfection based on the temporal changes of infrared absorption spectral data, it specifically performs the following steps: the interference identification module is also used to calculate the rate of change between the absorbance value at the current time and the absorbance values ​​at the previous two times for each preset wavelength; the interference identification module is also used to determine the current time as the start time of alcohol disinfection when the rate of change exceeds a preset threshold.

[0006] In one possible implementation, when the signal separation module separates the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectrum data based on the start time of alcohol disinfection, it specifically performs the following steps: The signal separation module is further configured to select absorbance values ​​from multiple historical times before the start time of alcohol disinfection for each target time after the start time of alcohol disinfection; wherein, the target time is any time after the start time of alcohol disinfection, and the historical times are multiple consecutive times before the start time of alcohol disinfection; The signal separation module is further configured to determine the initial absorbance value of alcohol vapor at the target time based on the difference between the absorbance value at the target time and the absorbance values ​​at multiple historical times; The signal separation module is further configured to perform weighted fusion of the initial absorbance values ​​corresponding to multiple historical times to obtain the absorbance values ​​of alcohol vapor at multiple preset wavelengths.

[0007] In one possible implementation, when the signal separation module determines the initial absorbance value of the alcohol vapor at the target time based on the difference between the absorbance value at the target time and the absorbance values ​​at multiple historical times, it specifically performs the following steps: The signal separation module is further configured to calculate the initial absorbance value for each preset wavelength based on the proportional relationship between the absorbance value at the target time and the absorbance values ​​at multiple historical times at each target wavelength; wherein, the target wavelength includes the preset wavelength and the wavelengths to the left and right of the preset wavelength.

[0008] In one possible implementation, when the signal separation module constructs an alcohol absorbance variation model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths, it specifically performs the following steps: The signal separation module is further configured to determine the optimal parameters of the biphasic decay model using a preset fitting algorithm based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths; wherein, the biphasic decay model is used to describe the decay characteristics of alcohol vapor absorbance changing with time; the signal separation module is further configured to predict the absorbance value of alcohol vapor at subsequent times using the biphasic decay model determined by the optimal parameters.

[0009] In one possible implementation, when the signal separation module determines the parameters of the biphasic attenuation model using a preset fitting algorithm based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths, it specifically performs the following steps: The signal separation module is further used to determine the prediction error of the biphasic attenuation model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths; wherein, the prediction error is used to characterize the degree of difference between the predicted value and the actual absorbance value of the biphasic attenuation model; the signal separation module is further used to determine the optimal parameters of the biphasic attenuation model by minimizing the prediction error.

[0010] In one possible implementation, when the signal separation module determines the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths from the infrared absorption spectral data based on the alcohol absorbance change model, it specifically performs the following steps: The signal separation module is also used to remove the absorbance values ​​of alcohol vapor at multiple preset wavelengths predicted by the alcohol absorbance change model from the infrared absorption spectral data to obtain the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths.

[0011] In one possible implementation, when the concentration calculation module determines the concentration value of the target anesthetic gas based on the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths, it specifically performs the following steps: the concentration calculation module is also used to establish a multivariate equation system of absorbance values ​​and concentration values ​​when multiple anesthetic gases are present; the concentration calculation module is also used to solve the multivariate equation system using a preset optimization algorithm to obtain the concentration value of each anesthetic gas.

[0012] In one possible implementation, when the data acquisition module acquires infrared absorption spectral data, it specifically performs the following steps: the data acquisition module is also used to acquire end-expiratory gas samples from patients; the data acquisition module is also used to perform water removal filtration on the end-expiratory gas samples; the data acquisition module is also used to measure the absorbance values ​​of the end-expiratory gas samples at multiple preset wavelengths using an infrared spectrometer to generate infrared absorption spectral data.

[0013] In one possible implementation, the system further includes a control output module; the control output module is used to generate an anesthetic infusion control command based on the target anesthetic gas concentration value determined by the concentration calculation module.

[0014] Secondly, the present invention provides an online detection method for anesthetic gas concentration for precise dosage control, comprising: acquiring infrared absorption spectral data; wherein the infrared absorption spectral data is used to characterize the absorbance values ​​of a patient's end-tidal gas sample at multiple preset wavelengths; determining the start time of alcohol disinfection based on the temporal changes of the infrared absorption spectral data; wherein the start time of alcohol disinfection is used to characterize the initial time point at which alcohol vapor begins to interfere with the detection of anesthetic gas concentration; separating the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectral data based on the start time of alcohol disinfection, and constructing an alcohol absorbance change model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths; determining the absorbance values ​​of a target anesthetic gas at multiple preset wavelengths from the infrared absorption spectral data based on the alcohol absorbance change model; and determining the concentration value of the target anesthetic gas based on the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths.

[0015] Thirdly, the present invention provides an electronic device, comprising: a processor and a memory; wherein the memory is used to store one or more programs, the one or more programs including computer-executable instructions, and when the electronic device is running, the processor executes the computer-executable instructions stored in the memory to cause the electronic device to perform the online detection method for precise dose control of anesthetic gas concentration as described in the first aspect and any possible implementation thereof.

[0016] Fourthly, the present invention provides a computer-readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by an electronic device of the present invention, cause the electronic device to perform the online detection method for precise dose control of anesthetic gas concentration as described in the first aspect and any possible implementation thereof.

[0017] Fifthly, the present invention provides a computer program product containing instructions that, when executed on a computer, cause the electronic device of the present invention to perform the online detection method for precise dose control of anesthetic gas concentration as described in the first aspect and any possible implementation thereof.

[0018] In a sixth aspect, the present invention provides a chip system applied to an online anesthetic gas concentration detection device for precise dosage control; the chip system includes one or more interface circuits and one or more processors. The interface circuits and the processors are interconnected via circuitry; the interface circuits are configured to receive signals from a memory of the online anesthetic gas concentration detection device for precise dosage control and to send the signals to the processors, the signals including computer instructions stored in the memory. When the processor executes the computer instructions, the online anesthetic gas concentration detection device for precise dosage control performs the online anesthetic gas concentration detection method for precise dosage control as described in the first aspect and any possible design thereof.

[0019] This invention offers the following advantages: By constructing a time-series spectral analysis framework, it effectively solves the problem of interference from alcohol vapor on the detection of anesthetic gas concentration, significantly improving the accuracy and clinical applicability of the detection system. The system can identify the starting point of alcohol disinfection interference in real time and accurately separate and superimpose spectral signals based on the dynamic characteristics of alcohol evaporation. This eliminates false increases in detection values ​​caused by alcohol vapor while maintaining a true reflection of the background concentration of anesthetic gas. Finally, through optimized algorithms, it outputs precise concentration values, providing reliable data support for the control of anesthetic drug dosage and effectively ensuring the precise management of anesthesia depth and clinical safety for surgical patients. Attached Figure Description

[0020] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the architecture of an online anesthetic gas concentration detection system for precise dosage control, provided in one embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of the architecture of a data acquisition module provided in one embodiment of the present invention;

[0023] Figure 3 This is a schematic diagram of the architecture of a signal separation module provided in one embodiment of the present invention;

[0024] Figure 4 This is a schematic flowchart of an online detection method for precise dosage control of anesthetic gas concentration provided in one embodiment of the present invention;

[0025] Figure 5 This is a schematic flowchart of another online detection method for precise dosage control of anesthetic gas concentration provided in an embodiment of the present invention. Detailed Implementation

[0026] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0028] The following description, in conjunction with the accompanying drawings, details a specific scheme for an online anesthetic gas concentration detection system for precise dosage control provided by the present invention.

[0029] For example, such as Figure 1 The diagram shown is an architectural schematic of an online anesthetic gas concentration detection system (hereinafter referred to as the online detection system) for precise dosage control according to an embodiment of the present invention. The online detection system 10 includes: a data acquisition module 11, an interference identification module 12, a signal separation module 13, a concentration calculation module 14, and a control output module 15. The modules are described below in sequence:

[0030] (1) Data acquisition module 11.

[0031] The data acquisition module 11 is responsible for collecting end-tidal gas samples from patients and generating infrared absorption spectrum data characterizing gas absorbance, providing raw data support for subsequent interference identification, signal separation and concentration calculation.

[0032] Optionally, the data acquisition module 11 is used to acquire infrared absorption spectral data. The infrared absorption spectral data is used to characterize the absorbance values ​​of the patient's end-tidal gas sample at multiple preset wavelengths.

[0033] For example, such as Figure 2 As shown, the data acquisition module 11 may include three sub-modules: an end-tidal gas sampling sub-module 111, a gas preprocessing sub-module 112, and an infrared spectroscopy detection sub-module 113. These will be described in detail below:

[0034] (1.1) End-tidal gas sampling submodule 111.

[0035] Optionally, the end-tidal gas sampling submodule 111 is used to collect end-tidal gas samples from the patient. Specifically, this submodule is connected to the patient's breathing circuit via a gas sampling three-way connector, and a sampling channel is constructed with a thin gas sampling tube. It continuously extracts end-tidal gas samples from the breathing circuit at a frequency of 10 times per second, with the sampling flow rate controlled at 150-200 mL / min—a flow rate that ensures sample representativeness (reflecting the true gas composition of the patient's alveoli) while avoiding interference with the breathing circuit pressure. The collected gas samples are directly delivered to the gas pretreatment submodule 112 to ensure the accuracy of subsequent testing.

[0036] (1.2) Gas pretreatment submodule 112.

[0037] Optionally, the gas pretreatment submodule 112 is used to perform water removal and filtration on the end-tidal gas sample. Specifically, the core function of the gas pretreatment submodule 112 is to eliminate the interference of moisture in the end-tidal gas on infrared spectroscopy detection. Since the patient's end-tidal gas contains a large amount of water vapor, if it directly enters the detection stage, the infrared absorption peak of water vapor may overlap with the absorption peak of anesthetic drugs or alcohol, leading to detection errors. Therefore, this submodule uses a dedicated water removal and filtration component (such as a polymer hydrophobic filter membrane) to remove water and filter the collected gas sample, removing liquid water and small particulate impurities from the sample. The processed dry gas sample is then delivered to the infrared spectroscopy detection submodule 113.

[0038] (1.3) Infrared spectroscopy detection submodule 113.

[0039] Optionally, the infrared spectroscopy detection submodule 113 is used to measure the absorbance value of the end-expiratory gas sample at multiple preset wavelengths using an infrared spectrometer to generate infrared absorption spectral data.

[0040] In one possible implementation, the infrared spectroscopy detection submodule 113 introduces a pre-treated dry gas sample into the analysis chamber of an infrared spectrometer. The instrument has multiple preset wavelengths matching the characteristic absorption peaks of the target anesthetic gas (such as sevoflurane or desflurane). Simultaneously, one auxiliary wavelength is added to each side of each preset wavelength (e.g., the interval between the preset wavelength and the auxiliary wavelengths is 0.1 μm, used for subsequent alcohol interference separation). Using an acousto-optic tunable filter or multiple sets of dedicated filters, the system rapidly cycles between the preset wavelength and the auxiliary wavelengths. The detector synchronously records the light intensity at each wavelength, and the absorbance value of the corresponding wavelength is calculated based on the light intensity attenuation. Finally, infrared absorption spectral data containing three dimensions—wavelength, absorbance, and time—is generated. This data is synchronously transmitted to the interference identification module 12 and the signal separation module 13 as the core input data for subsequent modules.

[0041] (2) Interference identification module 12.

[0042] The interference identification module 12 is responsible for identifying the key time point at which alcohol vapor begins to interfere with the detection based on the infrared absorption spectrum data output by the data acquisition module 11, so as to determine the start time T of alcohol disinfection and provide a time reference for the signal separation module 13 to accurately separate the absorbance of alcohol and anesthetic.

[0043] Specifically, for the temporal change of absorbance at each preset wavelength in the infrared absorption spectrum data, the interference identification module 12 calculates the "difference between the absorbance value at the current moment and the absorbance value at the previous moment," and performs a ratio calculation with the "absolute value of the difference between the absorbance value at the previous moment and the absorbance values ​​at the two moments before." When this ratio exceeds a preset threshold, the current moment is determined to be the start time T of alcohol disinfection. This is because during alcohol disinfection, the rapid evaporation of liquid alcohol causes a step increase in the alcohol concentration in the end-tidal gas, while the concentration of anesthetic drugs is precisely controlled by the vaporizer and changes gradually. Therefore, the rapid change in absorbance is only caused by alcohol interference.

[0044] The alcohol disinfection start time T determined by the interference identification module 12 will be directly transmitted to the signal separation module 13 as the core time reference for the separation of alcohol absorbance by the module 13.

[0045] (3) Signal separation module 13.

[0046] The signal separation module 13 is responsible for separating the absorbance of alcohol vapor and target anesthetic gas based on the infrared absorption spectrum data of the data acquisition module 11 and the alcohol disinfection start time T of the interference identification module 12, and finally extracting the absorbance data of pure anesthetic gas to provide accurate input for the concentration calculation module 14.

[0047] Optionally, the signal separation module 13 is used to separate the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectrum data according to the start time of alcohol disinfection, and to construct an alcohol absorbance change model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths.

[0048] Optionally, the signal separation module 13 is also used to determine the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths from infrared absorption spectral data based on the alcohol absorbance change model.

[0049] For example, such as Figure 3 As shown, the signal separation module 13 may include three sub-modules: a preliminary separation sub-module 131, a model construction sub-module 132, and an absorbance extraction sub-module 133. These sub-modules are described below:

[0050] (3.1) Preliminary separation of submodule 131.

[0051] Optionally, the preliminary separation submodule 131 is used to separate the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectrum data according to the start time of alcohol disinfection.

[0052] Specifically, the preliminary separation submodule 131, after the start time T of alcohol disinfection, initially separates the alcohol vapor absorbance at each preset wavelength and at each time point. The implementation involves two steps:

[0053] First, absorbance difference calculation: For any target time (T+l, l is the time interval after T) after the start time of alcohol disinfection, select multiple consecutive historical times before T (Tm, m is the time interval before T). Based on the symmetrical characteristics of the alcohol CH bond absorption peak (the absorbance of alcohol on the left and right sides is equal) and the fixed ratio of anesthetic absorbance (the ratio of anesthetic absorbance remains unchanged at the same wavelength before and after disinfection), construct an absorbance balance equation, calculate the absorbance difference between the target time and the historical time at each wavelength, and obtain the suspected alcohol absorbance.

[0054] Subsequently, weighted fusion optimization: considering that the closer the historical time T is to the disinfection time, the more stable the anesthetic concentration (small baseline drift), an exponential weighted algorithm is used to fuse the suspected alcohol absorbance corresponding to multiple historical times, and finally obtain the preliminary alcohol absorbance value at each time and each preset wavelength after the start of disinfection.

[0055] The output data of the initial separation submodule 131 will be transmitted to the model building submodule 132.

[0056] (3.2) Model building submodule 132.

[0057] Optionally, the model building submodule 132 is used to build an alcohol absorbance change model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths.

[0058] Specifically, the model building submodule 132 implements the construction of the alcohol absorbance change model in three steps:

[0059] Model Selection: Based on the two-phase characteristics of alcohol evaporation (fast phase: rapid evaporation of liquid alcohol on the surface, resulting in a decrease in absorbance index; slow phase: slow diffusion and evaporation of alcohol in porous materials, resulting in a slow decrease in absorbance), a two-phase decay function was selected as the model basis; wherein, the decay rate constant of the two-phase decay function is... and The decay rates corresponding to the fast and slow phases were obtained through extensive experimental data calibration in the early stages. During online fitting, only the amplitude parameter was optimized. a and b .

[0060] Usability assessment: For the initial alcohol absorbance value at each time point, the maximum-minimum normalization function (f function) is used to calculate its usability for model construction. Priority is given to selecting data with high usability in the early stage of evaporation (fast phase dominated) for model fitting, to avoid model bias caused by slow phase data in the later stage.

[0061] Parameter optimization: With the goal of minimizing the sum of squared errors between the model prediction value and the initial alcohol absorbance value, the unknown parameters of the biphasic decay function are solved by the least squares method to obtain the optimal biphasic decay model. This model can accurately predict the alcohol absorbance value at any time after the start of disinfection, and the prediction results are transmitted to the absorbance extraction submodule 133.

[0062] (3.3) Absorbance extraction submodule 133.

[0063] Optionally, the absorbance extraction submodule 133 is used to determine the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths from infrared absorption spectral data based on the alcohol absorbance change model.

[0064] Specifically, the absorbance extraction submodule 133 eliminates the superposition interference of alcohol absorbance and extracts the pure absorbance of the target anesthetic gas from the infrared absorption spectrum data. This is achieved based on the linear superposition characteristic of absorbance according to Lambert-Beer's law, i.e., total absorbance of the mixed gas = alcohol absorbance + anesthetic absorbance. The anesthetic absorbance is calculated in two scenarios:

[0065] When l≤4 (the first 4 moments after disinfection begins): Since the biphasic decay model has not yet fully converged, the "preliminary alcohol absorbance value" output by the preliminary separation submodule 131 is directly used to calculate the absorbance of the anesthetic drug using a preset formula; for details of the preset formula, please refer to S404 below, which will not be repeated here.

[0066] When l>4 (4 moments after the start of disinfection): The absorbance of alcohol predicted by the biphasic decay model constructed by model construction submodule 132 is used to calculate the absorbance of anesthetic drug using the same formula.

[0067] The "target anesthetic gas absorbance data" output by this submodule will be directly transmitted to the concentration calculation module 14 as the core basis for concentration calculation.

[0068] (4) Concentration calculation module 14.

[0069] The concentration calculation module 14 is responsible for calculating the real-time concentration of the anesthetic gas based on the absorbance data of the target anesthetic gas output by the signal separation module 13 and the Lambert-Beer law, providing a basis for the control output module 15.

[0070] Optionally, the concentration calculation module 14 is used to determine the concentration value of the target anesthetic gas based on the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths.

[0071] Specifically, when determining the concentration value of the target anesthetic gas, the concentration calculation module 14 divides the process into the following two scenarios:

[0072] For single anesthetic gas scenarios: Based directly on Lambert-Beer's law, combined with the inherent absorption coefficient of the anesthetic at the corresponding wavelength and the fixed optical path of the detection cavity, the concentration value is directly calculated from the absorbance of the target anesthetic gas.

[0073] Multiple anesthetic gas scenarios: Due to overlapping absorption peaks of various anesthetics at certain wavelengths, a multivariate relationship between absorbance and the concentration of each anesthetic needs to be established. Since the number of measurement wavelengths exceeds the number of anesthetic types, the least squares method is used to solve the relevant equations to find the concentration combination that minimizes the sum of squared errors between theoretical and actual absorbance, ultimately obtaining the concentration value of each anesthetic. The "real-time concentration value of anesthetic gas" output by the concentration calculation module 14 is transmitted to the control output module 15 for subsequent dosing regulation.

[0074] (5) Control output module 15.

[0075] The control output module 15 is responsible for generating anesthetic drug infusion control commands based on the real-time concentration value of the anesthetic drug gas output by the concentration calculation module 14, thereby achieving precise control of the anesthetic drug dosage.

[0076] Optionally, the control output module 15 is used to generate an anesthetic infusion control command based on the target anesthetic gas concentration value determined by the concentration calculation module.

[0077] Specifically, the control output module 15 compares the real-time concentration value output by the concentration calculation module 14 with a preset "anesthesia treatment window" (a concentration range that prevents intraoperative awareness while avoiding excessive anesthesia). If the real-time concentration is lower than the lower limit of the treatment window, a control command to "increase the anesthetic infusion rate" is generated; if the real-time concentration is higher than the upper limit of the treatment window, a control command to "decrease the anesthetic infusion rate" is generated; if the real-time concentration is within the treatment window, a control command to "maintain the current infusion rate" is generated. This control command is directly sent to the anesthetic vaporizer and other drug delivery devices to dynamically adjust the anesthetic output dose, ensuring that the patient's intraoperative anesthesia depth is always within a safe and effective range.

[0078] The above describes the online anesthetic gas concentration detection system 10 for precise dosage control and its included modules.

[0079] For example, such as Figure 4The diagram shown is a flowchart illustrating an online detection method for precise dosage control of anesthetic gas concentration according to an embodiment of the present invention, comprising the following steps:

[0080] S401. Acquire infrared absorption spectral data. The infrared absorption spectral data is used to characterize the absorbance values ​​of the patient's end-tidal gas sample at multiple preset wavelengths.

[0081] For example, this step can be performed by the data acquisition module 11 described above, and specifically includes the following steps:

[0082] (1) Collect end-tidal gas samples from patients.

[0083] The end-tidal gas sample mentioned in this invention refers to the gas sample expelled from the alveoli during the last stage of expiration. This gas has the closest composition to alveolar gas, and the concentration of anesthetic gas in it can directly reflect the drug concentration in the blood and even brain tissue, making it a key physiological indicator for clinical monitoring of the depth of anesthesia. Continuous collection and analysis of end-tidal gas using a sidestream sampling method enables real-time, non-invasive monitoring of the patient's anesthetic state, providing a physiological signal source for the precise concentration detection and dosage control of this invention.

[0084] Optionally, the end-tidal gas sampling submodule 111 in the data acquisition module 11 collects end-tidal gas samples from the patient through a bypass sampling system. The bypass sampling system includes a three-way valve, a thin gas collection tube, and a gas delivery line. The three-way valve is connected to the patient's breathing circuit, and one end of the thin gas collection tube is connected to the three-way valve, while the other end is connected to the gas pretreatment submodule 112 of the data acquisition module 11. The end-tidal gas sampling submodule 111 continuously extracts end-tidal gas samples from the patient's breathing circuit at a sampling frequency of 10 times per second, with the sampling flow rate controlled between 150 mL / min and 200 mL / min. This sampling frequency and flow rate ensure that the samples accurately reflect the gas composition within the patient's alveoli while avoiding interference with the pressure balance of the patient's breathing circuit.

[0085] (2) The end-expiratory gas sample was subjected to water removal filtration treatment.

[0086] Optionally, the gas pretreatment submodule 112 in the data acquisition module 11 performs water removal filtration on the end-expiratory gas sample; wherein, a polymer hydrophobic filter membrane is used as the water removal filtration component, and the filter membrane pore size is set to 0.2μm to 0.5μm, which can effectively remove liquid water, respiratory secretion microparticles and other impurities in the end-expiratory gas sample, avoid the interference of overlapping infrared absorption peaks caused by moisture, and ensure the accuracy of subsequent spectral detection. The processed dry gas sample is transported to the infrared spectral detection submodule 113 through a dedicated pipeline.

[0087] (3) Use an infrared spectrometer to measure the absorbance of the end-expiratory gas sample at multiple preset wavelengths to generate infrared absorption spectrum data.

[0088] Optionally, the infrared spectroscopy detection submodule 113 in the data acquisition module 11 is used to measure the absorbance value of the end-tidal gas sample at multiple preset wavelengths using an infrared spectrometer to generate infrared absorption spectral data. The analysis chamber of the infrared spectrometer is a sealed structure, and is equipped with an acousto-optic tunable filter or multiple sets of dedicated filters. The preset wavelengths include multiple core wavelength points that match the characteristic absorption peaks of the target anesthetic gas (such as sevoflurane or desflurane). An auxiliary wavelength point is added to the left and right of each core wavelength point. The wavelength difference between the auxiliary wavelength point and the corresponding core wavelength point is controlled between 0.1 μm and 0.3 μm for subsequent separation of alcohol interference signals.

[0089] Subsequently, the infrared spectroscopy detection submodule 113 uses an acousto-optic tunable filter or multiple sets of dedicated filters to rapidly cycle between a preset wavelength and auxiliary wavelengths on the left and right sides. The detector synchronously records the light intensity at each wavelength point, and then calculates the absorbance value of the corresponding wavelength based on the light intensity attenuation, finally generating infrared absorption spectrum data containing three dimensions of information: wavelength, absorbance, and time.

[0090] S402. Determine the start time of alcohol disinfection based on the temporal changes in infrared absorption spectral data. The start time of alcohol disinfection is used to characterize the initial time point at which alcohol vapor begins to interfere with the detection of anesthetic gas concentration.

[0091] For example, this step can be performed by the interference identification module 12 described above, and specifically includes the following steps:

[0092] (1) For each preset wavelength, calculate the rate of change between the absorbance value at the current time and the absorbance values ​​at the previous two times.

[0093] Specifically, the interference identification module 12 first extracts the absorbance values ​​of all preset wavelengths at each sampling time from the infrared absorption spectrum data output by the data acquisition module 11. The preset wavelengths include core wavelengths matching the characteristic absorption peaks of the target anesthetic gas and auxiliary wavelengths to the left and right of each core wavelength, maintaining consistency with the preset wavelengths of the data acquisition module 11. Then, for each preset wavelength, the degree of change in absorbance at the current time is calculated sequentially according to the sampling time. This degree of change is quantified by the rate of change, and the specific calculation formula is as follows:

[0094]

[0095] In the above formula, This represents the rate of change of absorbance at the t-th sampling time, which is the start time of alcohol disinfection, under the i-th preset wavelength. It is dimensionless, and the larger the value, the more significant the change in absorbance at that time.

[0096] It represents the absorbance value corresponding to the i-th preset wavelength at the t-th sampling time. It has no unit and is directly taken from the three-dimensional infrared absorption spectrum data of "sampling time-wavelength-absorbance" generated by the data acquisition module 11.

[0097] This represents the absorbance value corresponding to the i-th preset wavelength at the (t-1)-th sampling time. It is dimensionless and represents the data from the previous sampling time at the current time.

[0098] This represents the absorbance value corresponding to the i-th preset wavelength at the (t-2)-th sampling time. It has no unit and is the data from the two sampling times before the current time, ensuring that abrupt changes in absorbance can be captured by comparing data from previous and subsequent times.

[0099] This represents the minimum correction term, with a value of 10 to the power of negative 4 (unitless). It is used to avoid computational instability when the denominator is zero or close to zero, and to prevent excessive amplification of sensor noise during signal stability. When close to or equal to zero, This ensures that the formula is mathematically valid and computationally stable. In scenarios where alcohol disinfection causes significant abrupt changes in absorbance, Usually much larger ,at this time The impact on the calculated rate of change is negligible.

[0100] It should be noted that the numerator in the formula... The absolute value is not used, and its possible positive or negative values ​​are retained. This is because the core purpose of this step is to specifically distinguish and capture the positive abrupt change in absorbance signal caused by the step increase in alcohol vapor concentration at the beginning of alcohol disinfection, rather than responding to all types of absorbance changes. In this way, the start of alcohol disinfection can be distinguished from other types of signal changes (such as concentration decrease, negative pulse, etc.).

[0101] and Absolute value form is used to eliminate the interference of the absorbance change direction in the first two time moments on the calculation results; when the denominator is affected... and Approaching and converging And molecules When it is a clearly positive value, It will approach a maximum value, which perfectly matches the characteristic of the instantaneous change in absorbance during alcohol disinfection.

[0102] (2) When the rate of change exceeds the preset threshold, the current time is determined as the start time of alcohol disinfection.

[0103] Specifically, the interference identification module 12 calculates the rate of change for each preset wavelength. It compares the value with a preset threshold in real time.

[0104] For example, the preset threshold is set to 3. This value is calibrated based on a large amount of clinical trial data: When alcohol disinfection occurs, liquid alcohol evaporates rapidly, causing a step increase in the alcohol concentration in the end-tidal breath. However, the concentration of anesthetic drugs is precisely controlled by the vaporizer and changes gradually. Therefore, the rapid change in absorbance is only caused by alcohol interference, and the rate of change is low. It will significantly exceed 3; if alcohol disinfection is not performed, the absorbance is dominated only by the concentration of anesthetic, and the change is gradual with a low rate of change. It will remain below 3.

[0105] Furthermore, when the rate of change corresponding to any preset wavelength When the preset threshold 3 is exceeded, the interference identification module 12 determines that the t-th sampling time is the alcohol disinfection start time T. If the rate of change of multiple preset wavelengths exceeds the threshold at different sampling times, the earliest sampling time that meets the condition is taken as the final alcohol disinfection start time T to ensure the uniqueness and accuracy of the time reference.

[0106] S403. Based on the start time of alcohol disinfection, separate the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectrum data, and construct an alcohol absorbance change model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths.

[0107] For example, this step can be performed by the signal separation module 13 described above, and specifically includes the following steps:

[0108] (1) Based on the start time of alcohol disinfection, the absorbance values ​​of alcohol vapor at multiple preset wavelengths are separated from the infrared absorption spectrum data.

[0109] Optionally, the preliminary separation submodule 131 in the signal separation module 13 first performs absorbance difference calculation: For any target time (T+l, l is the time interval after T) after the start time of alcohol disinfection, select multiple consecutive historical times (Tm, m is the time interval before T) before T. Based on the symmetrical characteristics of the alcohol CH bond absorption peak (the absorbance of alcohol on the left and right wavelengths is equal) and the fixed ratio of anesthetic absorbance (the ratio of anesthetic absorbance remains unchanged at the same wavelength before and after disinfection), construct an absorbance balance equation, calculate the absorbance difference between the target time and the historical time at each wavelength, and obtain the suspected alcohol absorbance (that is, the initial absorbance value in the detailed process description of S501-S503 below).

[0110] Next, the preliminary separation submodule 131 performs weighted fusion optimization: considering that the closer the historical time T is to the disinfection time, the more stable the anesthetic concentration (small baseline drift), an exponential weighting algorithm is used to fuse the absorbance values ​​of suspected alcohol at multiple historical times, finally obtaining the absorbance values ​​of alcohol vapor at each time and at each preset wavelength after the start of disinfection. It should be noted that the specific process of separating the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectral data according to the above steps is described in S501-S503 below, and will not be repeated here.

[0111] In another possible implementation, when the preliminary separation submodule 131 separates the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectral data, it can also pre-establish a standard absorption spectral library of alcohol vapor at different concentrations. During the disinfection interference stage, the real-time acquired spectral data is matched with the spectral library. Combining the change law of alcohol concentration first rising rapidly and then slowly decreasing after the start of disinfection, the absorbance value corresponding to alcohol vapor is directly separated by calculating the spectral similarity. This method is suitable for detection environments where the baseline data is unstable.

[0112] Alternatively, the preliminary separation submodule 131 can construct a baseline trend model of the absorbance of anesthetic gas by performing smoothing filtering on the absorbance data of the baseline period before disinfection. During the disinfection interference period, the absorbance value of alcohol vapor is obtained by subtracting the real-time absorbance value from the predicted value of the baseline trend model. This method can effectively suppress the interference of small fluctuations in the baseline period on the separation results.

[0113] Thus, the preliminary separation submodule 131 accurately separates the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectrum data output by the data acquisition module 11, laying the foundation for subsequent construction of an alcohol absorbance change model and elimination of interference.

[0114] (2) Construct an alcohol absorbance variation model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths.

[0115] Optionally, the model building submodule 132 in the signal separation module 13 selects the biphase decay function as the model basis. Then, for the initial alcohol absorbance value at each time point, the usability for model building is calculated using a maximum-minimum normalization function. Data with high usability in the early stages of evaporation is prioritized for model fitting to avoid model bias caused by data from the later, slower phases. Finally, with the goal of minimizing the sum of squared errors between the model's predicted value and the initial alcohol absorbance value, the unknown parameters of the biphase decay function are solved using the least squares method to obtain the optimal biphase decay model. It should be noted that the specific process for constructing the alcohol absorbance change model according to the above steps is described in S504-S505 below, and will not be repeated here.

[0116] In another possible implementation, when the model building submodule 132 builds the alcohol absorbance change model, it can also divide the entire alcohol evaporation process into two stages, a fast phase and a slow phase, through piecewise fitting and weighted fusion modeling. Corresponding exponential decay models are established for each stage. By evaluating the reliability of the data in the two stages and assigning different weights, the output results of the two models are finally weighted and fused. This method reduces the computational complexity through piecewise modeling and is suitable for application scenarios with clear boundaries between the evaporation stages.

[0117] Alternatively, the model building submodule 132 can also adopt a machine learning prediction modeling approach, using the separated alcohol absorbance data as training samples, combining time series features and environmental parameters to construct feature vectors, using time series prediction algorithms to train the model, and through continuous iterative optimization, enabling the model to autonomously grasp the changing patterns of alcohol absorbance. This method does not rely on the assumption of the two-phase characteristics of alcohol evaporation and can adapt to the modeling needs in complex environments.

[0118] Therefore, the model construction submodule 132 can construct an alcohol absorbance change model that accurately describes the dynamic change of alcohol absorbance over time based on the absorbance values ​​of the separated alcohol vapor at multiple preset wavelengths, providing a reliable predictive basis for subsequent continuous elimination of alcohol interference and extraction of anesthetic absorbance.

[0119] S404. Based on the alcohol absorbance change model, determine the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths from infrared absorption spectral data.

[0120] For example, this step can be performed by the absorbance extraction submodule 133 in the signal separation module 13, specifically by removing the absorbance values ​​of alcohol vapor predicted by the alcohol absorbance change model at multiple preset wavelengths from the infrared absorption spectrum data, and obtaining the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths.

[0121] Specifically, in the initial stage after the start of alcohol disinfection (T), the accumulated data points are too few, leading to unstable parameter fitting of the biphasic decay model and potentially significant deviations in model predictions. The pre-separated alcohol absorbance value, however, offers more real-time data. Therefore, the absorbance extraction submodule 133 employs a segmented processing strategy: after the start of alcohol disinfection (T), when the accumulated data points are less than N0 (e.g., N0=20, corresponding to 2 seconds of data), the pre-separated alcohol vapor absorbance value is used as the standard alcohol absorbance value for determining the absorbance of the target anesthetic gas. When the accumulated data points reach or exceed N0, the alcohol vapor absorbance value predicted by the biphasic decay model is used as the standard alcohol absorbance value. Furthermore, based on the linear superposition property of the Lambert-Beer law—that the total absorbance of a gas mixture at a given wavelength is equal to the arithmetic sum of the absorbances of its components—the absorbance of the target anesthetic gas can be obtained by subtracting the standard alcohol absorbance value from the total absorbance. The specific calculation formula is as follows:

[0122]

[0123] in, Indicates the time after alcohol disinfection begins (T), the th l At the [time]th moment i Each wavelength corresponds to the absorbance value of the target anesthetic gas in the absorption peak; Indicates the time after alcohol disinfection begins (T), the th l At the [time]th moment i The total absorbance value at each wavelength is directly derived from infrared absorption spectroscopy data; This indicates the time T after the start of alcohol disinfection. The absorbance of the alcohol vapor initially separated at time i at the absorption peak corresponding to the i-th wavelength. Indicates the first The predicted absorbance of alcohol vapor is obtained by substituting the time into the two-phase decay model.

[0124] Therefore, through the above segmented calculation, the absorbance extraction submodule 133 can continuously and accurately separate the pure absorbance signal of the target anesthetic gas throughout the entire period of alcohol interference, laying a solid foundation for subsequent concentration calculation.

[0125] S405. Determine the concentration of the target anesthetic gas based on the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths.

[0126] For example, this step can be performed by the concentration calculation module 14 described above, and specifically includes the following steps:

[0127] (1) When multiple anesthetic gases are present, establish a multivariate equation system of absorbance and concentration values.

[0128] Specifically, according to the Lambert-Beer law and the theory of molecular vibrational spectroscopy, the total absorbance of the anesthetic gas measured at any wavelength is equal to the sum of the absorbances of each component anesthetic gas at that wavelength. It should be noted that the wavelengths referred to below, from the 1st to the Pth, are numbered according to a fixed wavelength order preset in the system for concentration calculation. This order is explicit and unique within the system algorithm, ensuring the standardization and repeatability of the equation set. For example, the wavelength order is arranged in ascending order of wavelength values. Thus, the following multivariate equation set is established:

[0129]

[0130] in, Indicates the time after alcohol disinfection begins (T), the th l The first wavelength at each time point corresponds to the total absorbance value of all anesthetic gases in the absorption peak. Indicates the time after alcohol disinfection begins (T), the th l The second wavelength at each time point corresponds to the total absorbance value of all anesthetic gases in the absorption peak. Indicates the time after alcohol disinfection begins (T), the th l At time P, the wavelength corresponds to the total absorbance value of all anesthetic gases in the absorption peak, where N represents the number of different types of anesthetic gases. , , These represent the molar absorptivity of the nth anesthetic gas at wavelengths 1, 2, and M, respectively, with units such as L·mol⁻¹. -1 ·cm -1 ; This represents the concentration of the nth anesthetic gas, in mol / L. d The optical path length of the infrared spectrometer's analysis cavity is expressed in cm; P represents the total number of preset wavelengths used for concentration calculation.

[0131] (2) The multivariate equation system is solved by using a preset optimization algorithm to obtain the concentration value of each anesthetic gas.

[0132] Since the number of wavelengths P used in actual measurements is usually greater than the number of anesthetic gas types N, the established equation system is an overdetermined system. The concentration calculation module 14 uses the least squares method as the preset optimization algorithm to solve the problem. By finding a set of concentration values ​​that minimizes the sum of squared errors between the theoretical absorbance calculated from this set and the measured absorbance, the optimal solution for the concentration value of each anesthetic gas is obtained. This algorithm effectively utilizes redundant wavelength measurement information, suppresses the influence of random measurement noise, and significantly improves the accuracy and stability of the concentration calculation results.

[0133] If only a single anesthetic gas exists, there is no need to establish a set of equations. The concentration calculation module 14 can directly calculate the real-time concentration of the anesthetic gas by measuring the absorbance value at the characteristic absorption peak according to the Lambert-Beer law and combining it with the known absorption coefficient and optical path length d.

[0134] Based on the above technical solution, this invention effectively solves the problem of interference from alcohol vapor on the detection of anesthetic gas concentration by constructing a time-series spectral analysis framework, significantly improving the accuracy and clinical applicability of the detection system. This system can identify the starting point of alcohol disinfection interference in real time and accurately separate and superimpose spectral signals based on the dynamic characteristics of alcohol evaporation. This eliminates false increases in detection values ​​caused by alcohol vapor while maintaining a true reflection of the background concentration of anesthetic gas. Finally, through optimized algorithms, it outputs precise concentration values, providing reliable data support for the control of anesthetic drug dosage and effectively ensuring the precise management of anesthesia depth and clinical safety for surgical patients.

[0135] For example, in combination Figure 4 ,like Figure 5 The diagram shown is a flowchart illustrating another online detection method for precise dosage control of anesthetic gas concentration provided by an embodiment of the present invention. In this method, based on the start time of alcohol disinfection, the absorbance values ​​of alcohol vapor at multiple preset wavelengths are separated from infrared absorption spectral data, and an alcohol absorbance change model is constructed based on these absorbance values. Specifically, the method includes the following steps:

[0136] S501. For each target time after the start time of alcohol disinfection, select the absorbance values ​​of multiple historical times before the start time of alcohol disinfection.

[0137] The target time is any time after the start of alcohol disinfection, and is synchronized with the sampling frequency (10 times per second) of the data acquisition module 11.

[0138] Furthermore, the historical timeframes are the 10-20 consecutive timeframes preceding the start of alcohol disinfection (corresponding to T-1 to T-20, where m is the interval number between the historical timeframe and the start of disinfection, m=1,2,...,20). The core basis for selecting consecutive historical timeframes is that within a short period (1-2 seconds) before disinfection, the drug concentration set by the anesthesiologist is stable, and the concentration of anesthetic in the patient's breathing circuit and alveoli is in dynamic equilibrium with a gradual change in concentration. This can serve as a stable baseline, avoiding errors in alcohol absorbance separation due to baseline drift.

[0139] Therefore, after determining the target time and historical time, the preliminary separation submodule 131 selects the absorbance values ​​of multiple historical times before the start time of alcohol disinfection.

[0140] S502. Determine the initial absorbance value of alcohol vapor at the target time based on the difference between the absorbance value at the target time and the absorbance values ​​at multiple historical times.

[0141] Optionally, when the preliminary separation submodule 131 performs this step, for each preset wavelength, it calculates the initial absorbance value based on the ratio of the absorbance value at the target time to the absorbance value at historical times at each target wavelength; wherein, the target wavelength includes the preset wavelength and the wavelengths to the left and right of the preset wavelength.

[0142] For example, the initial separation submodule 131 can calculate the initial absorbance value of the alcohol vapor at the target time using the following formula:

[0143]

[0144] In the above formula, Indicates the target time Preset wavelength Below, based on historical moments The initial absorbance value of the alcohol vapor was calculated; it has no unit.

[0145] Indicates the target time Preset wavelength The total absorbance value is unitless and taken from infrared absorption spectroscopy data.

[0146] Representing historical moments Preset wavelength The absorbance values ​​at this time are unitless and are taken from infrared absorption spectroscopy data. Since this is before the start of alcohol disinfection, the absorbance mainly reflects the signal of the target anesthetic gas (and other background gases that may not interfere with the measurement).

[0147] Representing historical moments Left wavelength The absorbance values ​​(absorbance values ​​of anesthetic drugs) are unitless and are taken from infrared absorption spectroscopy data.

[0148] Representing historical moments right-side wavelength The absorbance values ​​(absorbance values ​​of anesthetic drugs) are unitless and are taken from infrared absorption spectroscopy data.

[0149] Indicates the target time Left wavelength The total absorbance value is unitless and taken from infrared absorption spectroscopy data.

[0150] Indicates the target time right-side wavelength The total absorbance value is unitless and taken from infrared absorption spectroscopy data.

[0151] This represents the minimum correction term, which can take the value of 10 to the power of negative 4 (unitless), and is used to avoid calculation instability when the denominator is zero or close to zero.

[0152] It should be noted that the above formula aims to decouple the mixed signal at the current moment by utilizing the spectral characteristics of pure anesthetic gases from historical times (before disinfection). Molecular part This indicates the background signal at the core wavelength of a historical moment. Based on this, the signal is scaled proportionally to the sum of the total absorbance at the symmetrical wavelengths on both sides at the current moment. This operation aims to estimate: if the signal enhancement at the wavelengths on both sides at the current moment is entirely caused by the broad-spectrum absorption of alcohol, then at the core wavelength... Above, the absorbance scale that alcohol vapor may contribute; the denominator in the formula. This represents the coefficient used to normalize historical baseline data. The denominator represents the sum of the pure absorbance at the left and right wavelengths before disinfection, signifying the historical baseline morphology without alcohol interference. Its function is to provide a stable normalized benchmark based on the pure historical signal for scaling the proportions in the numerator. Therefore, the right side of the formula represents the reconstructed signal at the core wavelength based on the total signal change between the pure spectral morphology before disinfection and the symmetrical wavelengths after disinfection. The above should be attributed to the absorbance estimate of alcohol vapor. This is determined by the total absorbance from the target time. Subtracting this estimated value from the mixed signal can effectively remove the estimated alcohol interference component from the mixed signal, thereby obtaining the initial absorbance value of the alcohol vapor at the target time.

[0153] Therefore, the above formula achieves signal separation through three key designs: First, the total absorbance at the target time is used directly. As a calculation benchmark; secondly, utilizing historical moments before disinfection. absorbance data The background signal characteristics of the anesthetic gas are characterized. Finally, a correction term is constructed by combining the arithmetic mean of the absorbance at the current wavelength and the historical proportional relationship. This design establishes a benchmark reference through historical data and combines it with real-time monitoring values ​​at symmetrical wavelengths, effectively distinguishing the superposition effect of alcohol signals and anesthetic gas signals, thus maintaining theoretical rationality and significantly improving computational stability.

[0154] S503. Weighted fusion of the initial absorbance values ​​corresponding to multiple historical moments is performed to obtain the absorbance values ​​of alcohol vapor at multiple preset wavelengths.

[0155] For example, the preliminary separation submodule 131 can calculate the absorbance values ​​of alcohol vapor at multiple preset wavelengths using the following formula:

[0156]

[0157] In the above formula, Indicates the target time Preset wavelength The absorbance value of the alcohol vapor (unitless) is used as input data for subsequent model construction.

[0158] Representing historical moments The corresponding weighting coefficients are calculated based on an exponential function with the natural constant as the base. They are dimensionless, range from 0 to 1, and decrease exponentially as m increases.

[0159] M represents the total number of historical moments, i.e., the upper limit of the summation; m is the summation index, representing the offset of the historical moment, with values ​​ranging from 1 to M.

[0160] The design logic of the above formula is based on the following considerations: First, through the aforementioned steps, multiple initial estimates based on comparisons at different historical moments have been obtained. These estimates vary depending on the selected historical time points. To obtain a stable and reliable final result, these estimates need to be fused. The formula employs an exponentially weighted average algorithm to achieve this fusion process, where the weighting coefficients are designed so that historical time points closer to the start of disinfection (i.e., smaller m-values) are assigned higher weights. This weighting mechanism aligns with clinical practice—within a short timeframe, anesthetic gas concentrations and system baselines are more stable, thus initial estimates based on more recent historical time points have higher reliability. This weighted fusion method fully utilizes the data redundancy advantage of multiple historical time points while highlighting the contribution of high-quality data sources through weight allocation, thereby significantly improving the estimation accuracy and anti-interference capability of the final alcohol absorbance value.

[0161] S504. Based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths, the optimal parameters of the two-phase decay model are determined using a preset fitting algorithm. The two-phase decay model describes the decay characteristics of alcohol vapor absorbance over time.

[0162] For example, when the model building submodule 132 performs this step, it includes the following sub-steps:

[0163] (1) The prediction error of the two-phase decay model is determined based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths. The prediction error is used to characterize the degree of difference between the predicted value and the actual absorbance value of the two-phase decay model.

[0164] Specifically, the basic form of the two-phase decay model must first be defined:

[0165]

[0166] in, Indicates the first The predicted value of alcohol vapor absorbance obtained by substituting the time into the two-phase decay model; a , b These are the amplitude parameters to be fitted for the biphasic decay model; , The decay rate constants, determined in advance through experimental calibration, correspond to the decay characteristics of the fast and slow phases of alcohol evaporation, respectively. Indicates the time relative to the start of alcohol disinfection T The time offset.

[0167] Based on the above model form, the prediction error is calculated using the following formula:

[0168]

[0169] Where E represents the prediction error of the biphasic decay model, and L represents the total number of time intervals considered after the start of alcohol disinfection (to ensure the stability of the fit, L should not be less than N0, for example, N0=20, that is, at least 2 seconds of data should be accumulated). Indicates the time relative to the start of alcohol disinfection T The time offset, with values ​​ranging from 1 to L;

[0170] Indicates the time T after the start of alcohol disinfection. The usability of the alcohol vapor absorbance initially separated at each time point for constructing a model of alcohol absorbance changes is demonstrated by... Calculate, where e represents the natural constant;

[0171] This indicates the time T after the start of alcohol disinfection. The alcohol vapor initially separated at the [time] [time] [before] [time] i Each wavelength corresponds to the absorbance of the absorption peak. Indicates the first The absorbance prediction value is obtained by substituting it into the above two-phase attenuation model function.

[0172] (2) Determine the optimal parameters of the biphasic decay model by minimizing the prediction error.

[0173] In this step, the model building submodule 132 uses the least squares method to analyze the prediction error. E Minimization is performed by using an iterative optimization algorithm to find the solution that minimizes the prediction error. E Amplitude parameter at minimum value a , b These two parameters are determined as the optimal parameters for the two-phase decay model. Decay rate constant , The parameters are not involved in online optimization; their values ​​are obtained through extensive prior experimental data calibration to ensure the stability of the model structure. This optimization process fully considers the differences in data point availability at different times, and weighted least squares estimation ensures the accuracy and reliability of the model parameter estimates.

[0174] S505. Predict the absorbance values ​​of alcohol vapor at subsequent times using a two-phase decay model determined by optimal parameters.

[0175] In this step, the model building submodule 132 will use the time offset of subsequent time steps. l Substitute into the two-phase decay model with the already determined optimal parameters: In this process, the predicted absorbance of alcohol vapor at the corresponding time is directly calculated, providing an accurate quantitative benchmark for alcohol interference in subsequent signal separation.

[0176] Based on the above technical solutions, this invention effectively improves the accuracy and reliability of anesthetic gas concentration detection under alcohol interference by combining time-series analysis and dynamic modeling. First, an alcohol absorbance separation method based on multi-time-point comparison and weighted fusion fully utilizes the reference value of a stable baseline before disinfection, significantly enhancing the anti-interference capability of initial signal separation through data redundancy and weight allocation mechanisms. Then, an alcohol absorbance change model based on biphasic decay characteristics is introduced, accurately describing the dynamic process of alcohol evaporation through weighted least squares fitting, achieving precise prediction of alcohol interference at subsequent times. The entire method is interconnected, overcoming the limitations of single-time-point data being susceptible to noise and solving the problem of fixed models being unable to adapt to the complex dynamics of alcohol evaporation, ultimately providing a solid guarantee for the accurate calculation of anesthetic gas concentration.

[0177] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0178] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An online detection system for anesthetic gas concentration for precise dosage control, characterized in that, The system includes: a data acquisition module, an interference identification module, a signal separation module, and a concentration calculation module; The data acquisition module is used to acquire infrared absorption spectral data; wherein, the infrared absorption spectral data is used to characterize the absorbance values ​​of the patient's end-tidal gas sample at multiple preset wavelengths; The interference identification module is used to determine the start time of alcohol disinfection based on the temporal changes of the infrared absorption spectrum data; wherein, the start time of alcohol disinfection is used to characterize the initial time point at which alcohol vapor begins to interfere with the detection of anesthetic gas concentration; The signal separation module is used to separate the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectrum data according to the start time of alcohol disinfection, and to construct an alcohol absorbance change model based on the absorbance values ​​of alcohol vapor at multiple preset wavelengths. The signal separation module is also used to determine the absorbance value of the target anesthetic gas at the multiple preset wavelengths from the infrared absorption spectrum data according to the alcohol absorbance change model. The concentration calculation module is used to determine the concentration value of the target anesthetic gas based on the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths. When the interference identification module determines the start time of alcohol disinfection based on the temporal changes in the infrared absorption spectrum data, it specifically performs the following steps: The interference identification module is also used to calculate the rate of change between the absorbance value at the current moment and the absorbance values ​​at the previous two moments for each preset wavelength; The interference identification module is also used to determine the current time as the start time of the alcohol disinfection when the rate of change exceeds a preset threshold. When the signal separation module constructs an alcohol absorbance change model based on the absorbance values ​​of the alcohol vapor at multiple preset wavelengths, it specifically performs the following steps: The signal separation module is further configured to determine the prediction error of the two-phase decay model based on the absorbance values ​​of the alcohol vapor at the multiple preset wavelengths; wherein the two-phase decay model is used to describe the decay characteristics of the absorbance of alcohol vapor over time, and the prediction error is used to characterize the degree of difference between the predicted value of the two-phase decay model and the actual absorbance value. The signal separation module is further configured to determine the optimal parameters of the biphasic attenuation model by minimizing the prediction error; The signal separation module is also used to predict the absorbance value of alcohol vapor at subsequent times using a two-phase decay model determined by optimal parameters.

2. The online anesthetic gas concentration detection system for precise dosage control according to claim 1, characterized in that, When the signal separation module separates the absorbance values ​​of alcohol vapor at multiple preset wavelengths from the infrared absorption spectrum data based on the start time of alcohol disinfection, it specifically performs the following steps: The signal separation module is further configured to select absorbance values ​​from multiple historical moments prior to the start time of alcohol disinfection for each target moment after the start time of alcohol disinfection; wherein, the target moment is any moment after the start time of alcohol disinfection, and the historical moments are multiple consecutive moments prior to the start time of alcohol disinfection. The signal separation module is further configured to determine the initial absorbance value of the alcohol vapor at the target time based on the difference between the absorbance value at the target time and the absorbance values ​​at the plurality of historical times. The signal separation module is also used to perform weighted fusion of the initial absorbance values ​​corresponding to the multiple historical times to obtain the absorbance values ​​of the alcohol vapor at the multiple preset wavelengths.

3. The online anesthetic gas concentration detection system for precise dosage control according to claim 2, characterized in that, When the signal separation module determines the initial absorbance value of the alcohol vapor at the target time based on the difference between the absorbance value at the target time and the absorbance values ​​at multiple historical times, it specifically performs the following steps: The signal separation module is further configured to calculate the initial absorbance value for each preset wavelength based on the ratio of the absorbance value at the target time to the absorbance value at the historical time at each target wavelength; wherein the target wavelength includes the preset wavelength and the wavelengths to the left and right of the preset wavelength.

4. The online anesthetic gas concentration detection system for precise dosage control according to claim 1, characterized in that, When the signal separation module determines the absorbance value of the target anesthetic gas at multiple preset wavelengths from the infrared absorption spectral data based on the alcohol absorbance change model, it specifically performs the following steps: The signal separation module is further configured to remove the absorbance values ​​of alcohol vapor predicted by the alcohol absorbance change model at multiple preset wavelengths from the infrared absorption spectral data, thereby obtaining the absorbance values ​​of the target anesthetic gas at multiple preset wavelengths.

5. The online anesthetic gas concentration detection system for precise dosage control according to claim 1, characterized in that, When determining the concentration of the target anesthetic gas based on its absorbance values ​​at multiple preset wavelengths, the concentration calculation module specifically performs the following steps: The concentration calculation module is also used to establish a multivariate equation system of absorbance and concentration values ​​when multiple anesthetic gases are present. The concentration calculation module is also used to solve the multivariate equation system using a preset optimization algorithm to obtain the concentration value of each anesthetic gas.

6. The online anesthetic gas concentration detection system for precise dosage control according to claim 1, characterized in that, When the data acquisition module acquires infrared absorption spectrum data, it specifically performs the following steps: The data acquisition module is also used to collect end-tidal gas samples from patients; The data acquisition module is also used to perform water removal filtration on the end-tidal gas sample; The data acquisition module is also used to measure the absorbance value of the end-expiratory gas sample at multiple preset wavelengths using an infrared spectrometer, and generate the infrared absorption spectrum data.

7. The online detection system for anesthetic gas concentration for precise dosage control according to any one of claims 1-6, characterized in that, The system also includes a control output module; The control output module is used to generate an anesthetic infusion control command based on the target anesthetic gas concentration value determined by the concentration calculation module.

Images