Ventilator failure prediction method and system accompanying intensive care unit nursing stage

Abstract

The present invention relates to the technical field of ventilator fault data processing, and specifically to a ventilator fault prediction method and system accompanying the nursing stage of an intensive care unit. The present invention obtains a turning point by analyzing the difference between the tangent at each data point on the temperature time series data fitting curve and the fitting curve; analyzes the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor; analyzes the fluctuation of the temperature time series data points in the window area according to the number of updated turning points and their corresponding retention factors to obtain the risk level; divides the temperature time series data into multiple fluctuation segments, analyzes the similarity between the current fluctuation segment and the historical fluctuation segment, and obtains the matching degree between the current fluctuation segment and the historical fluctuation segment; and performs weighted processing on the temperature time series data and predicts the temperature data according to the risk level and the matching degree. The accuracy of temperature data prediction is improved, and the accuracy of fault prediction is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of ventilator fault data processing, and in particular to a ventilator fault prediction method and system accompanying a nursing stage in an intensive care unit. Background Art

[0002] Patients in intensive care units are often seriously ill and have impaired body functions, which results in their spontaneous breathing being unable to meet the body's oxygen needs. A ventilator can ensure that patients inhale enough oxygen to maintain normal gas exchange in the body. A respiratory humidifier is an important component of a ventilator. It uses physical heating to provide appropriate humidity and temperature for the dry air-oxygen mixture according to the preset temperature range of the patient's condition. If the heating system of the respiratory humidifier fails, it will directly affect the patient's treatment effect and cause harm to the patient. Therefore, it is necessary to predict the failure of the respiratory humidifier to reduce the failure rate.

[0003] When predicting faults of respiratory humidifiers, the first thing to do is to obtain predicted temperature data. The existing technology generally obtains the predicted temperature data time series through the autoregressive integrated moving average model (ARIMA). However, temperature data often has correlation in time series, that is, the temperature prediction value at a future moment will be affected by the temperature data at the current moment. Therefore, the weight corresponding to each historical temperature data is also different. Directly predicting based on historical temperature data will reduce the accuracy of the predicted temperature data, thereby reducing the accuracy of fault prediction. Summary of the invention

[0004] In order to solve the above technical problems, the object of the present invention is to provide a ventilator failure prediction method and system accompanying the intensive care unit nursing stage.

[0005] According to the first aspect of the embodiment of the present application, a method for predicting ventilator failure during the intensive care unit nursing phase is provided, and the technical solution adopted is as follows:

[0006] Obtain the preset temperature range and temperature time series data of the respiratory humidifier;

[0007] In a rectangular coordinate system, curve fitting is performed on the temperature time series data points, a window area of ​​the temperature time series data points is set, and based on the window area, the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve is analyzed to obtain a turning point;

[0008] Analyze the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor;

[0009] Based on the window area, according to the number of the update turning points and the corresponding retention factors, analyzing the fluctuation of the temperature time series data points in the window area to obtain the risk level of the temperature time series data points;

[0010] Dividing the temperature time series data into multiple fluctuation segments, analyzing the similarity between the current fluctuation segment and the historical fluctuation segment, and obtaining the matching degree between the current fluctuation segment and the historical fluctuation segment;

[0011] Obtaining the importance of the temperature time series data point according to the risk degree and the matching degree;

[0012] According to the importance, the temperature time series data is weighted and processed, and temperature data prediction is performed to obtain predicted temperature data.

[0013] In some embodiments of the present invention, curve fitting is performed on the temperature time series data points in a rectangular coordinate system, including:

[0014] Based on the temperature time series data, a rectangular coordinate system is constructed with time as the horizontal axis and temperature data as the vertical axis;

[0015] In a rectangular coordinate system, the temperature time series data is curve-fitted using the least square method to obtain a fitting curve and a fitting function of the fitting curve.

[0016] In some embodiments of the present invention, setting a window area of ​​a temperature time series data point includes:

[0017] In the rectangular coordinate system, the temperature time series data point is the center and the horizontal length is The area is used as the window area of ​​the temperature time series data points.

[0018] In some embodiments of the present invention, based on the window area, analyzing the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve to obtain the turning point includes:

[0019] In the window area of ​​the temperature time series data points, draw a straight line perpendicular to the horizontal axis through the first and last temperature time series data points to obtain two target straight lines;

[0020] Analyze the tangent line at the temperature time series data point, the two target straight lines and the area enclosed by the horizontal axis to obtain the tangent line feature;

[0021] Analyze the fitting curve, the two target straight lines and the area enclosed by the horizontal axis to obtain the characteristics of the fitting curve;

[0022] Analyze the difference between the tangent line characteristics and the fitting curve characteristics to obtain the turning factor of the temperature time series data point;

[0023] A turning factor threshold is set, and a turning point is obtained according to the turning factor.

[0024] In some embodiments of the present invention, analyzing the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor includes:

[0025] Analyze the minimum value of the difference between the temperature value corresponding to the turning point and the upper and lower temperature values ​​in the preset temperature range to obtain the retention factor of the turning point;

[0026] A retention factor threshold is set, and based on the retention factor, an update turning point and its corresponding retention factor are obtained.

[0027] In some embodiments of the present invention, based on the window area, according to the number of the update turning points and the corresponding retention factors, analyzing the fluctuation of the temperature time series data points in the window area to obtain the risk level of the temperature time series data points includes:

[0028] In the window area of ​​the temperature time series data point, the risk level of the temperature time series data point is obtained according to the number of the update turning points and the average value of the retention factors corresponding to all the update turning points.

[0029] In some embodiments of the present invention, the temperature time series data is divided into a plurality of fluctuation segments, including:

[0030] In a rectangular coordinate system, a reference straight line is drawn with a vertical coordinate of 37°C and parallel to the horizontal axis, the intersection of the fitting curve and the reference straight line is recorded as a reference point, and the fitting curve between two adjacent reference points is recorded as a fluctuation segment, thereby obtaining multiple fluctuation segments.

[0031] In some embodiments of the present invention, analyzing the similarity between the current fluctuation segment and the historical fluctuation segment to obtain the matching degree between the current fluctuation segment and the historical fluctuation segment includes:

[0032] Analyze the distance relationship between each temperature time series data point in the current fluctuation segment and all temperature time series data points in the historical fluctuation segment to obtain the distance coefficient;

[0033] Analyze the time interval between the current fluctuation segment and the historical fluctuation segment to obtain the time coefficient;

[0034] According to the distance coefficient and the time coefficient, the matching degree between the current fluctuation segment and the historical fluctuation segment is obtained.

[0035] According to a second aspect of an embodiment of the present application, a ventilator failure prediction system accompanying an intensive care unit nursing stage is provided, comprising: a memory and a processor, wherein:

[0036] The memory is used to store program code;

[0037] The processor is used to read the program code stored in the memory and execute the method described in the first aspect of the embodiment of the present application.

[0038] In some embodiments of the present invention, the processor comprises:

[0039] A temperature data acquisition module is used to obtain the preset temperature range and temperature time series data of the respiratory humidifier;

[0040] A turning point acquisition module is used to perform curve fitting on the temperature time series data points in a rectangular coordinate system, set a window area for the temperature time series data points, and analyze the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve based on the window area to obtain the turning point; and analyze the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor;

[0041] A risk level analysis module, configured to analyze the fluctuation of the temperature time series data points in the window area based on the number of the update turning points and the corresponding retention factors, and obtain the risk level of the temperature time series data points;

[0042] A matching degree analysis module is used to divide the temperature time series data into multiple fluctuation segments, analyze the similarity between the current fluctuation segment and the historical fluctuation segment, and obtain the matching degree between the current fluctuation segment and the historical fluctuation segment;

[0043] An importance acquisition module, used to obtain the importance of the temperature time series data point according to the risk level and the matching degree;

[0044] The temperature data prediction module is used to perform weighted processing on the temperature time series data according to the importance, and perform temperature data prediction to obtain predicted temperature data.

[0045] Compared with the prior art, the ventilator failure prediction method and system provided by the present invention during the intensive care unit nursing stage have the following beneficial effects:

[0046] The present invention performs curve fitting on the temperature time series data points in a rectangular coordinate system, sets a window area of ​​the temperature time series data points, analyzes the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve based on the window area, and obtains the turning point; and analyzes the relationship between the temperature value corresponding to the turning point and the preset temperature range, obtains the updated turning point and its corresponding retention factor, and the updated turning point obtained above is the temperature time series data point that may affect the operation of the ventilator in the local area before and after the temperature time series data point. Based on the window area, according to the number of updated turning points and their corresponding retention factors, the temperature fluctuation in the local area centered on the temperature time series data point is analyzed to obtain the risk level of the temperature time series data point; the temperature time series data is divided into multiple fluctuation segments, and the similarity between the current fluctuation segment and the historical fluctuation segment is analyzed to obtain the matching degree of the current fluctuation segment and the historical fluctuation segment. According to the similarity between the historical fluctuation segment and the current fluctuation segment, the time difference and the risk level of each temperature time series data point, each temperature time series data point is weighted, and the correlation of the temperature data in the time series is comprehensively considered, which improves the accuracy of the prediction of future temperature data and ensures the accuracy of fault prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A basic flow chart of a method for predicting ventilator failures accompanying the intensive care unit nursing stage provided by one embodiment of the present invention;

[0049] Figure 2 A schematic diagram of the basic components of a ventilator failure prediction system accompanying the intensive care unit nursing stage provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the ventilator failure prediction method and system accompanying the intensive care unit nursing stage proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0051] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. Terms such as "comprises", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a circuit structure, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such article or device. In the absence of further restrictions, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the article or device including the element.

[0052] The conditions of patients in the intensive care unit are usually critical, and the respiratory humidifier is an important component in the ventilator to maintain the oxygen supply temperature. Once a failure occurs, it will pose a serious threat to the patient's life safety. Therefore, it is necessary to analyze the acquired temperature data before the actual failure occurs to obtain predicted temperature data, and judge the possibility of failure based on the predicted temperature data. In order to improve the accuracy of fault prediction, it is necessary to improve the accuracy of predicted temperature data, and temperature data is correlated in time series, so the importance of each temperature time series data point in the temperature data time series sequence is analyzed.

[0053] The specific scheme of a ventilator failure prediction method accompanying the intensive care unit nursing stage provided by the present invention is described in detail below with reference to the accompanying drawings.

[0054] See also Figure 1 , which shows the basic process of a ventilator failure prediction method accompanying the intensive care unit nursing stage provided by an embodiment of the present invention.

[0055] like Figure 1 As shown, a ventilator failure prediction method accompanying the intensive care unit nursing stage provided by an embodiment of the present invention specifically includes:

[0056] S100: Obtaining a preset temperature range and temperature time series data of a respiratory humidifier.

[0057] A temperature sensor is used to collect a time series sequence of temperature data to obtain the temperature time series data of the respiratory humidifier. The collection frequency is once every 5 seconds, and the collected temperature time series data is subjected to Gaussian filtering.

[0058] When the ventilator is in use, the doctor will set the appropriate temperature range according to the specific condition and treatment needs of the patient in the intensive care unit, and then the respiratory humidifier will adjust the temperature according to the preset temperature range. Therefore, it is also necessary to obtain the preset temperature range of the respiratory humidifier.

[0059] At this point, the preset temperature range and temperature time series data of the respiratory humidifier are obtained.

[0060] The acquired temperature time series data in the respiratory humidifier itself will fluctuate, but as long as the fluctuation is within a safe range, it will not affect the safe operation of the ventilator, and the temperature time series data often has correlation in the time series, that is, the temperature prediction value at the future moment will be affected by the temperature data at the current moment. Therefore, according to the fluctuation in the acquired temperature time series data and the correlation between the historical temperature data and the temperature data at the current moment, all temperature data are analyzed for importance, and the importance of each temperature data is obtained, thereby improving the accuracy of temperature data prediction. The specific method includes steps S200 to S600.

[0061] S200: In a rectangular coordinate system, curve fitting is performed on the temperature time series data points, a window area of ​​the temperature time series data points is set, and based on the window area, the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve is analyzed to obtain a turning point.

[0062] The human body has certain requirements for the temperature of inhaled oxygen, especially for patients in intensive care units. Due to the decline in body function, overheated or overcooled oxygen will irritate the patient's respiratory tract. In the process of the respiratory humidifier regulating the temperature of oxygen, the heating element gradually ages over time, and its heating efficiency and stability may decrease, causing fluctuations in temperature data. Fluctuations within the preset temperature range are normal, but when the number and degree of fluctuations exceed the safe range, it is necessary to analyze the reasons for the changes in a timely manner and take corresponding measures to ensure the normal operation of the ventilator.

[0063] Based on the above analysis, this embodiment performs curve fitting on the temperature time series data points in a rectangular coordinate system, sets a window area for the temperature time series data points, and analyzes the degree of difference between the tangent at each temperature time series data point on the fitting curve and the fitting curve based on the window area to obtain the degree of trend change of each temperature time series data point. For temperature time series data points with a large degree of trend change, they may correspond to potential risk points of ventilator failure, and thus the turning points in the temperature time series data points can be obtained, providing a basis for subsequent fault prediction.

[0064] Specifically, in a rectangular coordinate system, curve fitting is performed on the temperature time series data points, including: constructing a rectangular coordinate system based on the temperature time series data, with time as the horizontal axis and temperature data as the vertical axis; in the rectangular coordinate system, curve fitting is performed on the temperature time series data using the least square method to obtain a fitting curve and a fitting function of the fitting curve. Setting a window area for the temperature time series data points includes: in the rectangular coordinate system, taking the temperature time series data point as the center and the horizontal length as The area is used as the window area of ​​the temperature time series data points, where the horizontal length The value of can be 30 seconds. Based on the window area, the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve is analyzed to obtain the turning point, including:

[0065] In the window area of ​​the temperature time series data points, draw a straight line perpendicular to the horizontal axis through the first and last temperature time series data points to obtain two target straight lines; analyze the tangent line at the temperature time series data point, the two target straight lines and the area enclosed by the horizontal axis to obtain the tangent line characteristics; analyze the fitting curve, the two target straight lines and the area enclosed by the horizontal axis to obtain the fitting curve characteristics; analyze the degree of difference between the tangent line characteristics and the fitting curve characteristics to obtain the turning factor of the temperature time series data point; set the turning factor threshold, and obtain the turning point according to the turning factor.

[0066] The calculation formula for the turning factor of each temperature time series data point in the rectangular coordinate system is:

[0067]

[0068]

[0069] In the formula, For the The area enclosed by the tangent line at each temperature time series data point, the two target straight lines, and the horizontal axis; Indicates The turning factor of each temperature time series data point; Indicates The horizontal coordinate of the first temperature time series data point in the window area of ​​100 temperature time series data points; Indicates The horizontal coordinate of the last temperature time series data point in the window area of ​​100 temperature time series data points; Indicates The ordinate of the intersection point of the first target straight line and the tangent line in the window area of ​​the temperature time series data point; Indicates The ordinate of the intersection point between the last target straight line and the tangent line in the window area of ​​the temperature time series data point; Represents the fitting function corresponding to the fitting curve in the rectangular coordinate system.

[0070] Indicates The difference between the area enclosed by the fitting curve and the tangent line in the window area of ​​the temperature time series data point. The larger the absolute value of the difference, the The greater the difference between the tangent line of the temperature time series data point and the fitting curve, the The greater the turning factor of each temperature time series data point.

[0071] The larger the turning factor of each temperature time series data point, the greater the difference between the fluctuation at the current temperature time series data point and the overall trend, that is, the more likely it is a turning point. Therefore, the turning factor threshold can be set to 0.6, and the temperature time series data points with a turning factor greater than 0.6 are recorded as turning points. A turning point may mean that the temperature data has changed significantly, which may indicate that the working state of the respiratory humidifier has changed, and may even be a precursor to a failure.

[0072] S300: Analyze the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor.

[0073] The turning point means that the temperature time series data point may fluctuate, but when the fluctuation has a smaller impact on the operation of the ventilator, the fluctuation of the turning point will have a smaller impact on the ventilator, that is, when the temperature at the turning point is within the safe range, it can be judged that the fluctuation of the turning point will not affect the operation of the ventilator. Therefore, it is necessary to judge whether the temperature at each turning point exceeds the safe range, further screen all turning points, and retain the turning points that may affect the operation of the ventilator.

[0074] Based on the above analysis, this embodiment analyzes the relationship between the temperature value corresponding to the turning point and the preset temperature range, and obtains the updated turning point and its corresponding retention factor. It further includes: analyzing the minimum value of the difference between the temperature value corresponding to the turning point and the upper and lower temperature values ​​in the preset temperature range to obtain the retention factor of the turning point; setting the retention factor threshold, and obtaining the updated turning point and its corresponding retention factor according to the retention factor.

[0075] The calculation formula for the retention factor of the construction turning point is:

[0076]

[0077] In the formula, Indicates The retention factor of the turning point; Indicates The temperature value of the turning point; Indicates the upper limit of the preset temperature range; Indicates the lower limit of the preset temperature range; represents the minimum function; Expressed as a natural constant An exponential function with base .

[0078] The smaller the value, the closer the temperature value of the turning point is to the boundary of the preset temperature range, which means that the adjustment ability of the temperature control system is close to the limit. At this time, a slight change in the parameters will cause the temperature to exceed the safe range. Therefore, the retention factor of the turning point should be larger.

[0079] The retention factor threshold can be set to 0.7, and the turning points with a retention factor greater than or equal to 0.7 are recorded as update turning points to obtain update turning points and their corresponding retention factors; the retention factors less than 0.7 are non-turning points.

[0080] S400: Based on the window area, according to the number of updated turning points and their corresponding retention factors, the fluctuation of the temperature time series data points in the window area is analyzed to obtain the risk level of the temperature time series data points.

[0081] When patients breathe rapidly, they are usually accompanied by a large amount of gas exchange, which will cause the respiratory humidifier to experience faster temperature fluctuations, and the degree of fluctuation may be large. This is important monitoring data and should be given a larger weight to ensure the accuracy of subsequent temperature predictions, and then ensure the accuracy of fault predictions. Therefore, this embodiment is based on the window area, and according to the number of updated turning points and their corresponding retention factors, the fluctuation of the temperature time series data points in the window area is analyzed to obtain the risk level of the temperature time series data points. Furthermore, in the window area of ​​the temperature time series data points, according to the number of updated turning points and the average value of the retention factors corresponding to all updated turning points, the fluctuation of the temperature time series data points in the window area is analyzed to obtain the risk level of the temperature time series data points.

[0082] The calculation formula for constructing the risk level of temperature time series data points is:

[0083]

[0084] In the formula, Indicates The risk level of each temperature time series data point; Indicates Update the number of turning points in the window area of ​​100 temperature time series data points; Indicates The mean of the retention factors of all updated turning points in the window area of ​​100 temperature time series data points; represents the linear normalization function.

[0085] No. Update the number of turning points in the window area of ​​​​temperature time series data points The more, the more frequent the fluctuation of the temperature in the local area centered on the temperature time series data point, and the greater the risk level of the corresponding temperature time series data point; The mean of the retention factors of all updated turning points in the window area of ​​the temperature time series data points The larger it is, the more likely the temperature fluctuation in the local area centered on the temperature time series data point is to exceed the safety range. That is, the greater the fluctuation degree of the temperature time series data point, the greater the risk degree of the corresponding temperature time series data point.

[0086] S500: Divide the temperature time series data into multiple fluctuation segments, analyze the similarity between the current fluctuation segment and the historical fluctuation segment, and obtain the matching degree between the current fluctuation segment and the historical fluctuation segment.

[0087] Normally, the temperature of the respiratory humidifier will fluctuate around the normal human body temperature of 37°C, so the temperature time series data can be divided into several fluctuation segments according to the temperature fluctuation. The temperature time series data is divided into multiple fluctuation segments, specifically including: in a rectangular coordinate system, a reference straight line with a ordinate of 37°C and parallel to the horizontal axis is drawn, the intersection of the fitting curve and the reference straight line is recorded as a reference point, and the fitting curve between two adjacent reference points is recorded as a fluctuation segment, and multiple fluctuation segments are obtained.

[0088] It is known that different respiratory states of patients correspond to different fluctuations. In the temperature prediction process, the historical fluctuation segments similar to the patient's current respiratory state should be given a larger weight to ensure the accuracy of the prediction. Therefore, this embodiment obtains the matching degree between the current fluctuation segment and the historical fluctuation segment by analyzing the similarity between the current fluctuation segment and the historical fluctuation segment.

[0089] Since the current fluctuation segment is not a complete fluctuation segment, while the historical fluctuation segments are all complete fluctuation segments, the current fluctuation segment will be distorted when matching with the historical fluctuation segments. The result of directly calculating the DTW value of the current fluctuation segment and each historical fluctuation segment is inaccurate, and the length of each historical fluctuation segment is different. The longer the length, the larger the DTW value of the current fluctuation segment. Therefore, the similarity between the current fluctuation segment and the historical fluctuation segment is analyzed to obtain the matching degree between the current fluctuation segment and the historical fluctuation segment. Further, the distance relationship between each temperature time series data point in the current fluctuation segment and all temperature time series data points in the historical fluctuation segment is analyzed to obtain the distance coefficient. Here, by averaging the distances corresponding to each temperature time series data point in the current fluctuation segment and all temperature time series data points in the historical fluctuation segment, it will not be affected by the change in the length of the historical fluctuation segment; at the same time, the time interval between the current fluctuation segment and the historical fluctuation segment is analyzed to obtain the time coefficient; finally, according to the distance coefficient and the time coefficient, the matching degree between the current fluctuation segment and the historical fluctuation segment is obtained.

[0090] The calculation formula for constructing the matching degree between the current fluctuation segment and the historical fluctuation segment is:

[0091]

[0092] in, Indicates the current fluctuation segment and the The matching degree of historical fluctuation segments; Indicates the current fluctuation segment The temperature time series data point is The mean of the Euclidean distance values ​​between all temperature time series data points in a historical fluctuation segment; Indicates the number of temperature time series data points in the current fluctuation segment; Indicates The first temperature time series data collection moment in a historical fluctuation segment, Indicates the collection time of the first temperature time series data in the current fluctuation segment.

[0093] Indicates that all temperature data in the current fluctuation segment are consistent with the The similarity of the temperature time series data in the historical fluctuation segments, the similarity of the temperature time series data in the current fluctuation segment The temperature time series data point is The larger the mean of the Euclidean distances of all temperature time series data points in a historical fluctuation segment, the smaller the similarity. Indicates the current fluctuation segment and the The longer the time interval between the historical fluctuation segments and the current fluctuation segment, the The lower the credibility of a historical fluctuation segment, the smaller the corresponding matching degree.

[0094] S600: Obtain the importance of the temperature time series data points according to the risk level and the matching degree.

[0095] Combine the current fluctuation segment with the The matching degree of the historical fluctuation segment is taken as the The initial importance of each temperature time series data point in a historical fluctuation segment, the higher the matching degree of the historical fluctuation segment with the current fluctuation segment, the greater the initial importance of each temperature time series data point in the historical fluctuation segment, and the greater the impact on the temperature data at future moments. Increasing the weight of the temperature time series data point can improve the accuracy of the predicted temperature data, thereby further ensuring the accuracy of fault prediction.

[0096] Therefore, according to the matching degree between the current fluctuation segment and the historical fluctuation segment, combined with the risk level of the temperature time series data points in the historical fluctuation segment, the importance of each temperature time series data point in each historical fluctuation segment is obtained. The calculation formula for constructing the importance of the temperature time series data point is:

[0097]

[0098] In the formula, Indicates The first of the historical fluctuations The importance of each temperature time series data point; Indicates the current fluctuation segment and the The matching degree of historical fluctuation segments; Indicates The first of the historical fluctuations The risk level of each temperature time series data point; represents the linear normalization function.

[0099] The current fluctuation range is The matching degree of historical fluctuation segments The larger the value, the more the current fluctuation range is different from the first The greater the similarity between the historical fluctuation segments, the The greater the importance of the temperature time series data point in the historical fluctuation segment, the greater the importance of the temperature time series data point in the historical fluctuation segment. The first of the historical fluctuations The risk level of each temperature time series data point The larger it is, the more important the temperature time series data point is for monitoring and should be given a larger weight. The greater the importance of the corresponding temperature time series data point is.

[0100] S700: weighting the temperature time series data according to the importance, and predicting the temperature data to obtain predicted temperature data.

[0101] The greater the importance of a temperature time series data point in a historical fluctuation segment, the greater the impact of the temperature time series data point in the historical fluctuation segment on future temperature data. Therefore, this embodiment weights the temperature time series data according to the importance, and predicts the temperature data to obtain predicted temperature data.

[0102] Specifically, the The first of the historical fluctuations The importance of each temperature time series data point As the weight of the temperature time series data, in addition, 1 is used as the weight of each temperature time series data point in the current segment.

[0103] According to the above method, the weight of each temperature time series data is obtained. According to the weight of each temperature time series data, the weighted ARIMA model is used to predict the temperature time series data to obtain predicted temperature data.

[0104] According to step S200 to step S400, the risk level of the predicted temperature data is obtained. It should be noted that since the predicted temperature data is the last data in the time series, the corresponding window area takes the area in front of the predicted temperature data as the window area for analysis, and the remaining operations are the same as step S200 to step S400. When the risk level of the predicted temperature data is greater than 0.6, the corresponding position of the ventilator is highlighted on the central monitoring screen of the intensive care unit, and the temperature abnormality warning information is marked with a striking color (such as red or yellow flashing), completing the fault risk warning.

[0105] Based on the same inventive concept as the above method, this embodiment also provides a ventilator failure prediction system accompanying the intensive care unit nursing stage.

[0106] See also Figure 2 , which shows the basic composition of a ventilator failure prediction system accompanying the intensive care unit nursing stage provided by an embodiment of the present invention.

[0107] like Figure 2 As shown, a ventilator failure prediction system accompanying the intensive care unit nursing stage includes: a memory 10 and a processor 20, wherein:

[0108] A memory 10, used for storing program codes;

[0109] The processor 20 is used to read the program code stored in the memory 10, and execute to obtain the preset temperature range and temperature time series data of the respiratory humidifier; in a rectangular coordinate system, curve fitting is performed on the temperature time series data points, a window area of ​​the temperature time series data points is set, and based on the window area, the difference between the tangent at each temperature time series data point on the fitting curve and the fitting curve is analyzed to obtain the turning point; the relationship between the temperature value corresponding to the turning point and the preset temperature range is analyzed to obtain the updated turning point and its corresponding retention factor; based on the window area, according to the number of updated turning points and their corresponding retention factors, the fluctuation of the temperature time series data points in the window area is analyzed to obtain the risk level of the temperature time series data points; the temperature time series data is divided into multiple fluctuation segments, the similarity between the current fluctuation segment and the historical fluctuation segment is analyzed to obtain the matching degree between the current fluctuation segment and the historical fluctuation segment; according to the risk level and the matching degree, the importance of the temperature time series data points is obtained; according to the importance, the temperature time series data is weighted and the temperature data is predicted to obtain the predicted temperature data.

[0110] The processor 20 includes a temperature data acquisition module 21, a turning point acquisition module 22, a risk level analysis module 23, a matching degree analysis module 24, an importance acquisition module 25 and a temperature data prediction module 26.

[0111] The temperature data acquisition module 21 is used to obtain the preset temperature range and temperature time series data of the respiratory humidifier;

[0112] The turning point acquisition module 22 is used to perform curve fitting on the temperature time series data points in a rectangular coordinate system, set a window area for the temperature time series data points, and analyze the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve based on the window area to obtain the turning point; and analyze the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain the updated turning point and its corresponding retention factor;

[0113] The risk level analysis module 23 is used to analyze the fluctuation of the temperature time series data points in the window area based on the number of updated turning points and their corresponding retention factors, and obtain the risk level of the temperature time series data points;

[0114] The matching degree analysis module 24 is used to divide the temperature time series data into multiple fluctuation segments, analyze the similarity between the current fluctuation segment and the historical fluctuation segment, and obtain the matching degree between the current fluctuation segment and the historical fluctuation segment;

[0115] The importance acquisition module 25 is used to obtain the importance of the temperature time series data point according to the risk level and the matching degree;

[0116] The temperature data prediction module 26 is used to perform weighted processing on the temperature time series data according to the importance, and perform temperature data prediction to obtain predicted temperature data.

[0117] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0118] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for predicting ventilator failure during the intensive care unit nursing phase, characterized in that: The method comprises: Obtain the preset temperature range and temperature time series data of the respiratory humidifier; In a rectangular coordinate system, curve fitting is performed on the temperature time series data points, a window area of ​​the temperature time series data points is set, and based on the window area, the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve is analyzed to obtain a turning point; Analyze the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor; Based on the window area, according to the number of the update turning points and the corresponding retention factors, analyzing the fluctuation of the temperature time series data points in the window area to obtain the risk level of the temperature time series data points; Dividing the temperature time series data into multiple fluctuation segments, analyzing the similarity between the current fluctuation segment and the historical fluctuation segment, and obtaining the matching degree between the current fluctuation segment and the historical fluctuation segment; Obtaining the importance of the temperature time series data point according to the risk degree and the matching degree; According to the importance, weighted processing is performed on the temperature time series data, and temperature data is predicted to obtain predicted temperature data; In the rectangular coordinate system, curve fitting is performed on the temperature time series data points, including: Based on the temperature time series data, a rectangular coordinate system is constructed with time as the horizontal axis and temperature data as the vertical axis; In a rectangular coordinate system, using the least squares method to perform curve fitting on the temperature time series data to obtain a fitting curve and a fitting function of the fitting curve; Set the window area of ​​the temperature time series data points, including: In the rectangular coordinate system, the area with the temperature time series data point as the center and the horizontal length of l is used as the window area of ​​the temperature time series data point; Based on the window area, the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve is analyzed to obtain the turning point, including: In the window area of ​​the temperature time series data points, draw a straight line perpendicular to the horizontal axis through the first and last temperature time series data points to obtain two target straight lines; Analyze the tangent line at the temperature time series data point, the two target straight lines and the area enclosed by the horizontal axis to obtain the tangent line feature; Analyze the fitting curve, the two target straight lines and the area enclosed by the horizontal axis to obtain the characteristics of the fitting curve; Analyze the difference between the tangent line characteristics and the fitting curve characteristics to obtain the turning factor of the temperature time series data point; Setting a turning factor threshold, and obtaining a turning point according to the turning factor; Analyzing the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor, including: Analyze the minimum value of the difference between the temperature value corresponding to the turning point and the upper and lower temperature values ​​in the preset temperature range to obtain the retention factor of the turning point; Setting a retention factor threshold, and obtaining an update turning point and its corresponding retention factor according to the retention factor; Based on the window area, according to the number of the update turning points and the corresponding retention factors, analyzing the fluctuation of the temperature time series data points in the window area to obtain the risk level of the temperature time series data points includes: In the window area of ​​the temperature time series data points, according to the number of the update turning points and the average value of the retention factors corresponding to all the update turning points, the fluctuation of the temperature time series data points in the window area is analyzed to obtain the risk level of the temperature time series data points; The temperature time series data is divided into a plurality of fluctuation segments, including: In a rectangular coordinate system, a reference straight line is drawn with a ordinate of 37°C and parallel to the horizontal axis, the intersection of the fitting curve and the reference straight line is recorded as a reference point, and the fitting curve between two adjacent reference points is recorded as a fluctuation segment, thereby obtaining multiple fluctuation segments; Analyze the similarity between the current fluctuation segment and the historical fluctuation segment to obtain the matching degree between the current fluctuation segment and the historical fluctuation segment, including: Analyze the distance relationship between each temperature time series data point in the current fluctuation segment and all temperature time series data points in the historical fluctuation segment to obtain the distance coefficient; Analyze the time interval between the current fluctuation segment and the historical fluctuation segment to obtain the time coefficient; According to the distance coefficient and the time coefficient, the matching degree between the current fluctuation segment and the historical fluctuation segment is obtained.

2. A ventilator failure prediction system accompanying the intensive care unit nursing stage, characterized in that: The system comprises: a memory and a processor, wherein: The memory is used to store program codes; The processor is used to read the program code stored in the memory and execute the method as claimed in claim 1.

3. The ventilator failure prediction system accompanying the intensive care unit nursing stage according to claim 2 is characterized in that: The processor comprises: A temperature data acquisition module is used to obtain the preset temperature range and temperature time series data of the respiratory humidifier; A turning point acquisition module is used to perform curve fitting on the temperature time series data points in a rectangular coordinate system, set a window area for the temperature time series data points, and analyze the difference between the tangent line at each temperature time series data point on the fitting curve and the fitting curve based on the window area to obtain the turning point; and analyze the relationship between the temperature value corresponding to the turning point and the preset temperature range to obtain an updated turning point and its corresponding retention factor; A risk level analysis module, configured to analyze the fluctuation of the temperature time series data points in the window area based on the number of the update turning points and the corresponding retention factors, and obtain the risk level of the temperature time series data points; A matching degree analysis module is used to divide the temperature time series data into multiple fluctuation segments, analyze the similarity between the current fluctuation segment and the historical fluctuation segment, and obtain the matching degree between the current fluctuation segment and the historical fluctuation segment; An importance acquisition module, used to obtain the importance of the temperature time series data point according to the risk level and the matching degree; The temperature data prediction module is used to perform weighted processing on the temperature time series data according to the importance, and perform temperature data prediction to obtain predicted temperature data.

Images