Method and system for measuring dehydration amount of hemodialysis instrument

Abstract

The invention belongs to the technical field of medical information, and discloses a hemodialysis instrument dehydration amount measuring method and system.The method comprises the steps that the mass variation, measured by an electronic balance, of purified water in a container is obtained in real time, and environmental influence parameters around the electronic balance are obtained in real time; pre-processing the mass variation and the environmental influence parameters; according to the pre-processed environmental influence parameters, a pre-trained measurement error identification model is used to acquire a measurement error of the electronic balance so as to correct the pre-processed mass variation; periodic interference components in the corrected mass variation are eliminated based on a fast Fourier transform algorithm and an LMS filter; the dehydration amount is calculated according to the mass variation after the periodic interference components are eliminated; therefore, the dehydration amount can be automatically measured, the measurement precision is high, and the practicability is good.

Description

Technical Field

[0001] The present application relates to the field of medical information technology, and in particular to a method and system for measuring dehydration of a hemodialysis machine. Background Art

[0002] Hemodialysis is an essential medical intervention for the treatment of end-stage renal disease and chronic renal failure. By removing toxins and waste from the blood, hemodialysis effectively prolongs the life of patients and significantly improves their quality of life, allowing them to maintain their daily activities and stay healthy while waiting for a kidney transplant.

[0003] The principle of hemodialysis is to draw blood out of the body and let it flow in the opposite direction to the dialysate in the dialyzer. By utilizing the concentration gradient and hydrostatic pressure gradient on both sides of the membrane, through diffusion, convection and ultrafiltration, small and medium-molecule metabolic waste and excess water in the blood are selectively removed, while retaining large-molecule blood cells and proteins, and supplementing the required physiological ions, thereby achieving blood purification.

[0004] During hemodialysis, accurate measurement of dehydration is particularly important. Dehydration refers to the amount of excess water removed from the patient's body during dialysis. Since the renal function of patients with end-stage renal disease is severely impaired, the balance of water and electrolytes in the body depends on effective dehydration during dialysis. Excessive dehydration may lead to adverse reactions such as hypotension and muscle cramps, while too little dehydration may cause complications such as water and sodium retention and heart failure. Therefore, accurate measurement and control of dehydration through hemodialysis can effectively avoid these potential risks and ensure patient safety and treatment effects.

[0005] The weighing method is a common method for measuring dehydration. The specific operation is: connect the hemodialysis machine and the blood circuit according to the treatment mode, and immerse the arteriovenous end of the blood circuit in a container of pure water placed on an electronic balance. After the blood circuit is filled with pure water, adjust the temperature and dialysate flow rate of the hemodialysis machine. After the flow rate is stabilized, operate according to the preset ultrafiltration rate, and use a stopwatch to determine the ultrafiltration time. At different ultrafiltration rates, weigh the mass reduction of pure water to ensure the accuracy and reliability of the measurement. Finally, the dehydration amount is calculated based on the measured data using the following formula: ; in, To calculate the dehydration amount, is the mass change of pure water in the container, is the density of pure water, is the ultrafiltration time.

[0006] The above weighing method has the following problems: 1. Complicated operation: The balance and stopwatch need to be operated manually, requiring the operator to constantly weigh and record the time. This repetitive operation not only increases the workload, but also easily leads to operational errors. In order to ensure the measurement accuracy, it is necessary to repeat the measurement at multiple time points and calculate the average value, which means that the entire measurement process may require long-term continuous attention and operation, increasing the workload.

[0007] 2. Low precision: Electronic balances are very sensitive to the surrounding environment, such as air flow, vibration, temperature changes, etc., which will affect the measurement accuracy. In a hospital environment, these interference factors are difficult to avoid; manual operation is prone to human errors, such as inaccurate recording time, reading errors, etc. These errors will accumulate and affect the accuracy of the final result; during the weighing process, the flow and fluctuation of the liquid will also affect the stability of the reading, and it is necessary to wait until the liquid is completely still before reading, which further increases the difficulty of operation.

[0008] 3. The practical application has great limitations: the weighing method is time-consuming and inefficient, and cannot meet the needs of dialysis processes that require rapid adjustment and monitoring; the weighing method cannot achieve real-time monitoring. Real-time monitoring of dehydration is very important for timely adjustment of dialysis parameters, while the weighing method can only provide intermittent and discrete measurement data.

[0009] Therefore, the prior art needs to be improved and enhanced. Summary of the invention

[0010] The purpose of the present application is to provide a method and system for measuring the dehydration amount of a hemodialysis machine, which can automatically measure the dehydration amount with high measurement accuracy and good practicality.

[0011] In a first aspect, the present application provides a method for measuring the dehydration amount of a hemodialysis machine, which measures the dehydration amount based on an electronic balance, on which a container containing pure water is placed, and the arteriovenous end of a blood circuit connected to the hemodialysis machine according to a preset treatment mode is immersed in the container; the method for measuring the dehydration amount of a hemodialysis machine comprises the steps of: A1. Real-time acquisition of the mass change of the pure water in the container measured by the electronic balance, and real-time acquisition of environmental impact parameters around the electronic balance; A2. Preprocessing the mass change and the environmental impact parameters; A3. According to the pre-processed environmental impact parameters, the measurement error of the electronic balance is obtained by using the pre-trained measurement error recognition model to correct the mass change after pre-processing; A4. Eliminate the periodic interference components in the corrected mass variation based on the fast Fourier transform algorithm and LMS filter; A5. Calculate the dehydration amount based on the mass change after eliminating the periodic interfering components.

[0012] The hemodialysis instrument dehydration amount measurement method realizes the automatic measurement of the dehydration amount in the hemodialysis process, does not need manual weighing and timing, has high efficiency, avoids measurement errors caused by manual operation, can realize real-time continuous monitoring, and improves practicality; the mass change measured by the electronic balance is corrected in combination with the environmental influence parameter, and the periodic interference components of the corrected mass change are eliminated, which can improve the accuracy of the measurement data, thereby improving the measurement accuracy of the dehydration amount.

[0013] Preferably, the environmental influencing parameter includes at least one of temperature, humidity, pressure and air flow rate.

[0014] These environmental parameters have a great impact on the measurement accuracy of the electronic balance. By collecting these environmental parameters in real time and using them to correct the mass change, the accuracy and effectiveness of the measurement data can be effectively improved.

[0015] Preferably, step A3 comprises: Inputting the preprocessed environmental impact parameter into a pretrained measurement error recognition model to obtain a measurement error output by the pretrained measurement error recognition model; The measurement error is subtracted from the pre-processed mass change to obtain the corrected mass change.

[0016] The measurement error can be accurately estimated using the pre-trained measurement error identification model, thereby ensuring that the corrected mass change has a high accuracy.

[0017] Preferably, step A4 comprises: Using the first sliding window to intercept the corrected mass variation to obtain first intercepted data; Performing a fast Fourier transform on the first intercepted data to obtain frequency domain data corresponding to the first intercepted data; Extracting periodic component frequency information according to the frequency domain data; According to the periodic component frequency information, the periodic interference component in the corrected mass variation is filtered out by using an LMS filter to obtain the mass variation after eliminating the periodic interference component.

[0018] Preferably, after step A3, the following steps are further included: Obtaining the mass change rate according to the corrected mass change; The sampling frequency of the electronic balance is adjusted according to the mass change rate.

[0019] Preferably, after step A5, the method further comprises the following steps: identifying a treatment pattern using a deep learning model based on the amount of dehydration; Comparing the recognition result of the treatment mode with the preset treatment mode to determine whether the system is abnormal; An alarm signal is issued when the system is abnormal.

[0020] Preferably, the deep learning model is an LSTM neural network model.

[0021] In a second aspect, the present application provides a hemodialysis machine dehydration measurement system, comprising a hemodialysis machine, an electronic balance, an environmental sensor, an intelligent terminal and a cloud server, wherein a container containing pure water is placed on the electronic balance, the hemodialysis machine and the blood circuit are connected according to a preset treatment mode, the arteriovenous end of the blood circuit is immersed in the container, and the electronic balance, the environmental sensor and the cloud server are all connected to the intelligent terminal for communication; The electronic balance is used to measure the mass change of the pure water in the container in real time; The environmental sensor is used to measure the environmental impact parameters around the electronic balance in real time; The intelligent terminal is used to preprocess the mass change and the environmental impact parameter, and obtain the measurement error of the electronic balance according to the preprocessed environmental impact parameter using a pretrained measurement error recognition model to correct the preprocessed mass change, eliminate the periodic interference component in the corrected mass change based on a fast Fourier transform algorithm and an LMS filter, calculate the dehydration amount according to the mass change after eliminating the periodic interference component, and send the dehydration amount to the cloud server; The cloud server is used to store the dehydration amount.

[0022] Preferably, the environmental influencing parameter includes at least one of temperature, humidity, pressure and air flow rate, and the environmental sensor correspondingly includes at least one of a temperature sensor, a humidity sensor, a pressure sensor and an air flow rate sensor.

[0023] Preferably, the cloud server is also used to identify the treatment mode according to the dehydration amount using a deep learning model, compare the identification result of the treatment mode with the preset treatment mode, determine whether the system is abnormal, and issue an alarm signal when the system is abnormal.

[0024] Beneficial effects: The hemodialysis machine dehydration amount measurement method and system provided in the present application realize the automatic measurement of the dehydration amount in the hemodialysis process, without the need for manual weighing and timing, with high efficiency, and avoids measurement errors caused by manual operation, and can realize real-time continuous monitoring, thereby improving practicality; the mass change measured by the electronic balance is corrected in combination with the environmental impact parameters, and the periodic interference components of the corrected mass change are eliminated, which can improve the accuracy of the measurement data, thereby improving the measurement accuracy of the dehydration amount. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A flow chart of a method for measuring dehydration of a hemodialysis machine provided in an embodiment of the present application.

[0026] Figure 2 A schematic diagram of the structure of a hemodialysis machine dehydration measurement system provided in an embodiment of the present application.

[0027] Figure 3 A schematic diagram of the structure of an intelligent terminal.

[0028] Figure 4 Schematic diagram of the connection structure of the electronic balance, container and blood circuit.

[0029] Explanation of reference numerals: 1. Hemodialysis machine; 2. Electronic balance; 3. Environmental sensor; 4. Intelligent terminal; 401. Main controller; 402. Warning light; 403. Display; 404. Memory; 405. Wireless transceiver module; 5. Cloud server; 90. Container. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0031] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0032] Please refer to Figure 1 , Figure 1 A method for measuring dehydration of a hemodialyzer in some embodiments of the present application is based on measuring dehydration using an electronic balance, a container containing pure water is placed on the electronic balance, and the arteriovenous end of the blood circuit connected to the hemodialyzer according to a preset treatment mode is immersed in the container (such as Figure 4 The method for measuring the dehydration amount of a hemodialyzer comprises the following steps: A1. Obtain the mass change of pure water in the container measured by the electronic balance in real time, and obtain the environmental impact parameters around the electronic balance in real time; A2. Preprocess the quality change and environmental impact parameters; A3. According to the pre-processed environmental impact parameters, the measurement error of the electronic balance is obtained using the pre-trained measurement error recognition model to correct the mass change after pre-processing; A4. Eliminate the periodic interference components in the corrected mass variation based on the fast Fourier transform algorithm and LMS filter; A5. Calculate the dehydration amount based on the mass change after eliminating the periodic interfering components.

[0033] The hemodialysis instrument dehydration amount measurement method realizes the automatic measurement of the dehydration amount in the hemodialysis process, does not need manual weighing and timing, has high efficiency, avoids measurement errors caused by manual operation, can realize real-time continuous monitoring, and improves practicality; the mass change measured by the electronic balance is corrected in combination with the environmental influence parameter, and the periodic interference components of the corrected mass change are eliminated, which can improve the accuracy of the measurement data, thereby improving the measurement accuracy of the dehydration amount.

[0034] Among them, the hemodialysis machine dehydration measurement method can be applied to Figure 2 The hemodialysis machine dehydration measurement system is shown.

[0035] The electronic balance collects the mass change at a sampling frequency, thereby obtaining a series of mass change values. In step A1, the mass change collected by the electronic balance at a sampling frequency is obtained. The sampling frequency can be fixed or automatically adjustable.

[0036] The environmental impact parameters include at least one of temperature, humidity, pressure, and air velocity. Preferably, they include temperature, humidity, pressure, and air velocity. These environmental impact parameters have a great influence on the measurement accuracy of the electronic balance. By collecting these environmental impact parameters in real time to correct the mass change, the accuracy and effectiveness of the measurement data can be effectively improved. Each environmental impact parameter can be measured by a corresponding environmental sensor.

[0037] Among them, in step A2, the preprocessing of the mass change includes abnormal value replacement processing. Specifically, the average value and standard deviation of the mass change are calculated, the absolute value of the difference between each mass change value and the average value is calculated, recorded as the mass change deviation, and the mass change value whose ratio of the mass change deviation to the standard deviation is higher than a preset ratio threshold (which can be set according to actual needs, such as 3, but not limited to this) is determined as an abnormal value, and the replacement value of the abnormal value is calculated based on the non-abnormal mass change value adjacent to the abnormal value using an interpolation algorithm (such as linear interpolation or polynomial interpolation), and the corresponding abnormal value is replaced by the replacement value.

[0038] In step A2, the preprocessing of the mass change amount may also include filtering processing, for example, filtering the mass change amount by a Kalman filter. Since the mass change amount measured by the electronic balance has a certain noise interference component, filtering processing by a Kalman filter can filter out most of the noise interference components. Since the noise interference component in the mass change amount is not fixed, the Kalman filter is an adaptive filter that can observe the change of the interference component and adaptively adjust the filter parameters to better filter out the interference component.

[0039] In step A2, the preprocessing of the environmental impact parameters includes filtering the environmental impact parameters through a Kalman filter. Since the environmental impact parameters measured by the environmental sensor also have certain noise interference components, if they are directly used to correct the mass change, additional errors will be introduced, and the accuracy and reliability of the correction results cannot be guaranteed. Therefore, filtering the environmental impact parameters through a Kalman filter first can improve the accuracy of the environmental impact parameters, thereby ensuring the accuracy and reliability of the correction results. Since the noise interference components in the environmental impact parameters are not fixed, and the Kalman filter is an adaptive filter that can observe the changes in the interference components and adaptively adjust the filter parameters to better filter out the interference components.

[0040] Specifically, step A3 includes: A301. Input the preprocessed environmental impact parameters into the pre-trained measurement error recognition model to obtain the measurement error output by the pre-trained measurement error recognition model; A302. Subtract the measurement error from the preprocessed mass change to obtain the corrected mass change.

[0041] The measurement error can be accurately estimated using the pre-trained measurement error identification model, thereby ensuring that the corrected mass change has a high accuracy.

[0042] Among them, the pre-trained measurement error identification model is obtained in the following manner: obtaining multiple historical measurement samples to form a historical measurement sample set, the historical measurement samples including the historical mass changes measured using the same electronic balance (the electronic balance used in the hemodialysis machine dehydration measurement method of the present application), the corresponding historical environmental impact parameters and the corresponding mass change reference value (i.e., the true value of the mass change); constructing an initial measurement error identification model (the existing deep learning model structure can be used to construct the initial measurement error identification model according to actual needs); using the historical measurement sample set to train the initial measurement error identification model to iteratively optimize its model parameters (for example, using gradient descent method, adam optimizer, etc. to optimize the model parameters) until the loss function converges or the number of iterations reaches a preset threshold, thereby obtaining a pre-trained measurement error identification model.

[0043] Since the performance of an electronic balance changes with its service life, it is necessary to collect actual data to expand the historical measurement sample set during actual use, and to regularly perform reinforcement training on the pre-trained measurement error recognition model to ensure the accuracy of the error recognition results of the pre-trained measurement error recognition model.

[0044] Since the environmental influencing parameters will change, in order to ensure the accuracy of the calculated measurement error, the measurement error can be calculated periodically according to a preset period (which can be set according to actual needs), and before the next measurement error calculation, the pre-processed mass change collected subsequently can be corrected according to the latest calculated measurement error.

[0045] However, when calculating the measurement error according to the preset period, if the period is set too long, the measurement error will not change in time according to the changes in the environmental influencing parameters, and the sudden and drastic changes in the environmental influencing parameters may be ignored, resulting in inaccurate correction results; if the period is set too short, it is necessary to frequently call the pre-trained measurement error identification model to calculate the measurement error, which results in a large computational load and low data processing efficiency.

[0046] To this end, in some preferred implementations, step A301 may include: At the initial moment, the current preprocessed environmental impact parameter is input into the pre-trained measurement error recognition model to obtain the measurement error output by the pre-trained measurement error recognition model; After the initial moment, compare the current preprocessed environmental impact parameters with the preprocessed environmental impact parameters when the measurement error was calculated last time to determine whether the measurement error needs to be updated. If necessary, input the current preprocessed environmental impact parameters into the pretrained measurement error recognition model to obtain the measurement error output by the pretrained measurement error recognition model.

[0047] Therefore, in step A302, before the next measurement error identification is performed, the pre-processed mass change currently acquired is subtracted from the most recently calculated measurement error to obtain a corrected mass change.

[0048] Through the above method, the measurement error is updated only when a certain deviation occurs between the current preprocessed environmental impact parameters and the preprocessed environmental impact parameters when the measurement error was calculated last time. This can avoid the measurement error being calculated too frequently, resulting in a large calculation load and low data processing efficiency, and avoid the measurement error being updated too long, resulting in untimely measurement error updates and inaccurate correction results.

[0049] Comparing the current pre-processed environmental impact parameter with the pre-processed environmental impact parameter when the measurement error was calculated last time to determine whether the measurement error needs to be updated may include: The comprehensive change rate of environmental impact parameters is calculated according to the following formula: ; in, is the comprehensive change rate of environmental impact parameters, is the parameter value of the i-th environmental impact parameter after preprocessing when the measurement error was calculated last time, is the parameter value of the i-th environmental impact parameter after the current preprocessing, n is the total number of environmental impact parameters (for example, if the environmental impact parameters include temperature, humidity, pressure and air flow rate, then n=4), is the weight factor of the i-th environmental impact parameter (the weight factor of each environmental impact parameter can be set according to actual needs); If the comprehensive change rate of the environmental impact parameters exceeds a preset change rate threshold (which can be set according to actual needs), it is determined that the measurement error needs to be updated; otherwise, it is determined that the measurement error does not need to be updated.

[0050] The comprehensive change rate of environmental impact parameters reflects the comprehensive deviation of all environmental impact parameters. When the comprehensive change rate of environmental impact parameters does not exceed the preset change rate threshold, the actual measurement error changes little, and there is no need to update the measurement error. Otherwise, the measurement error needs to be updated to avoid the measurement error used in the correction process from deviating too much from the actual measurement error, resulting in inaccurate correction results.

[0051] Specifically, step A4 includes: A401. Using the first sliding window to intercept the corrected mass change to obtain the first intercepted data; A402. Perform a fast Fourier transform on the first intercepted data to obtain frequency domain data corresponding to the first intercepted data; A403. Extract periodic component frequency information based on frequency domain data; A404. According to the frequency information of the periodic component, the LMS (Least Mean Squares) filter is used to filter out the periodic interference component in the corrected mass change to obtain the mass change after eliminating the periodic interference component.

[0052] The length of the first sliding window can be set according to actual needs, and the sliding step size of the first sliding window can be set according to actual needs. In step A401, the mass change value is intercepted from the corrected mass change through the first sliding window to obtain a plurality of mass change values ​​(whose number is equal to the length of the first sliding window) to form the first intercepted data.

[0053] In step A402, fast Fourier transform is a prior art and will not be described in detail here.

[0054] In the frequency domain data, a spike value will appear at the frequency of the periodic component. In step A403, by identifying the spike value, the frequency of the periodic component can be extracted to obtain the periodic component frequency information. The method of identifying the spike value is a prior art and will not be described in detail here.

[0055] After the frequency information of the periodic component is identified, the corresponding periodic interference component can be filtered out by a filter, for example, by a bandpass filter. In step A404, the periodic interference component is filtered out by an LMS filter, which is an adaptive filter that can observe the periodic variation of the periodic interference component according to the first intercepted data intercepted by each first sliding window, and adaptively adjust the filter parameters, so as to better suppress the periodic interference component and make the output signal (i.e., the mass variation after eliminating the periodic interference component) smoother.

[0056] Specifically, in step A5, the dehydration amount is calculated according to the following formula: ; Where, t is the duration of hemodialysis (i.e. ultrafiltration time), is the dehydration amount at time t, is the mass change after eliminating the periodic interference components at time t, is the density of pure water.

[0057] In some preferred embodiments, step A3 further includes: A6. Obtain the mass change rate according to the corrected mass change; A7. Adjust the sampling frequency of the electronic balance according to the mass change rate.

[0058] Wherein, step A6 comprises: Extract the most recently collected N corrected mass change values ​​as reference mass change values, where N is a preset positive integer value not less than 2 (which can be set according to actual needs, for example, 4, but not limited thereto); The mass change rate was calculated according to the following formula: ; in, is the mass change rate, is the value of the ith reference mass change, is the i+1th reference mass change value; the reference mass change values ​​are sorted in the order of acquisition time.

[0059] Among them, in step A7, the target sampling frequency can be determined according to the mass change rate range segment into which the mass change rate falls, and the sampling frequency of the electronic balance is adjusted to the target sampling frequency. Multiple mass change rate range segments can be pre-set, and a reference sampling frequency can be assigned to each mass change rate range segment, wherein the higher the mass change rate, the greater the reference sampling frequency corresponding to the mass change rate range segment. In step A7, according to the mass change rate range segment into which the mass change rate falls, the reference sampling frequency of the corresponding mass change rate range segment is extracted as the target sampling frequency. After adjusting the sampling frequency of the electronic balance, the electronic balance will subsequently collect the mass change amount according to the adjusted sampling frequency. Thus, when the mass change rate is high, a larger sampling frequency is used to collect the mass change amount to ensure that the details of the rapid changes are captured. When the mass change rate is low, a smaller sampling frequency is used to collect the mass change amount to reduce the amount of data processing and save computing resources and storage space.

[0060] In some embodiments, after step A5, the method further comprises the following steps: A8. Use deep learning models to identify treatment patterns based on the amount of dehydration; A9. Compare the recognition result of the treatment mode with the preset treatment mode to determine whether the system is abnormal; A10. Send out alarm signal when the system is abnormal.

[0061] Wherein, step A8 comprises: A801. Using the second sliding window to intercept the dehydration amount to obtain the second intercepted data; wherein the length and sliding step of the second sliding window can be set according to actual needs; A802. normalize the second intercepted data and record it as the third intercepted data; A803. Input the third intercepted data into the deep learning model to obtain the treatment pattern recognition result output by the deep learning model.

[0062] Among them, the deep learning model can use an LSTM (long short-term memory) neural network model, but is not limited to this.

[0063] The deep learning model is obtained in the following manner: multiple groups of historical third-party interception data under various treatment modes are obtained and the corresponding treatment modes are labeled to obtain multiple training samples to form a training sample set; after constructing the initial deep learning model, the initial deep learning model is trained with the training samples to iteratively optimize its model parameters until the loss function converges or the number of iterations reaches a preset threshold, thereby obtaining the final deep learning model.

[0064] In some embodiments, step A9 includes: if the currently identified treatment mode is different from the preset treatment mode, determining that the system is abnormal.

[0065] However, the recognition results of the deep learning model are not absolutely correct. Judging whether the system is abnormal based on only one recognition result can easily lead to misjudgment. Therefore, in another embodiment, step A9 includes: if the number of times the treatment mode identified within a preset time period (which can be set according to actual needs) before the current moment is different from the preset treatment mode exceeds a preset upper limit (which can be set according to actual needs), the system is judged to be abnormal.

[0066] In step A8, treatment mode recognition can be performed for each set of second intercepted data. Therefore, the period of treatment mode recognition depends on the sliding step of the second sliding window, but the sliding step should not be set too large to avoid missing key information. Therefore, treatment mode recognition is more frequent.

[0067] Sometimes, in order to avoid excessive computational load caused by calling the deep learning model too frequently for treatment mode recognition, treatment mode recognition is performed at a preset time interval, which is larger than the time interval corresponding to the sliding step of the second sliding window (the time interval corresponding to the sliding step of the second sliding window is equal to the product of the sliding step of the second sliding window and the sampling period of the dehydration amount). Each time treatment mode recognition is performed, a second intercepted data intercepted within the last preset time interval is randomly selected for recognition (using steps A802 and A803 for recognition); if the recognition result of one treatment mode recognition is different from the preset treatment mode, all second intercepted data intercepted within the last preset time interval are further recognized (using steps A802 and A803 for recognition). If the number of times that the further recognized recognition result is different from the preset treatment mode exceeds the preset number limit, the system is determined to be abnormal. On the one hand, the accuracy of the judgment result can be improved, and on the other hand, it can avoid excessive computational load caused by calling the deep learning model too frequently for treatment mode recognition.

[0068] By sending out an alarm signal, medical staff can be reminded to take corresponding measures in time to avoid adverse effects on patients.

[0069] In some embodiments, after step A8, the method may further include the following steps: A11. Send the identification results of the dehydration amount and the treatment mode to the user terminal (the user terminal may be a mobile terminal).

[0070] The identification results of dehydration amount and treatment mode are sent to the user end so that the user can remotely monitor the hemodialysis process.

[0071] refer to Figure 2 The present application provides a hemodialysis machine dehydration measurement system (the system can be used to implement the hemodialysis machine dehydration measurement method described above), comprising a hemodialysis machine 1, an electronic balance 2, an environmental sensor 3, an intelligent terminal 4 and a cloud server 5, a container 90 containing pure water is placed on the electronic balance 2, the hemodialysis machine 1 and the blood circuit are connected according to a preset treatment mode, and the arteriovenous end of the blood circuit is immersed in the container 90 (such as Figure 4 As shown, a in the figure is the arterial end of the blood circuit, and b is the venous end of the blood circuit), the electronic balance 2, the environmental sensor 3 and the cloud server 5 are all connected to the intelligent terminal 4 for communication; The electronic balance 2 is used to measure the mass change of the pure water in the container 90 in real time (refer to the above step A1 for details); The environmental sensor 3 is used to measure the environmental impact parameters around the electronic balance 2 in real time (refer to step A1 above for details); The intelligent terminal 4 is used to pre-process the mass change and the environmental impact parameters, and obtain the measurement error of the electronic balance according to the pre-processed environmental impact parameters using the pre-trained measurement error recognition model to correct the pre-processed mass change, eliminate the periodic interference components in the corrected mass change based on the fast Fourier transform algorithm and the LMS filter, calculate the dehydration amount according to the mass change after eliminating the periodic interference components (refer to the above steps A2 to A5 for details), and send the dehydration amount to the cloud server 5; The cloud server 5 is used to store the dehydration amount.

[0072] Preferably, the environmental influencing parameter includes at least one of temperature, humidity, pressure and air flow rate, and the environmental sensor 3 correspondingly includes at least one of a temperature sensor, a humidity sensor, a pressure sensor and an air flow rate sensor.

[0073] Preferably, the intelligent terminal 4 is further used to obtain the mass change rate according to the corrected mass change amount after correction of the pre-processed mass change amount, and adjust the sampling frequency of the electronic balance according to the mass change rate (refer to steps A6 and A7 above for details).

[0074] Preferably, the cloud server 5 is also used to identify the treatment mode according to the dehydration amount using a deep learning model, compare the identification result of the treatment mode with the preset treatment mode, determine whether the system is abnormal, and send an alarm signal when the system is abnormal (refer to steps A8-A10 above for details). The cloud server 5 implements the function of identifying the treatment mode, and can use the powerful computing power of the cloud server 5 to provide data analysis support, reduce the performance requirements for the local smart terminal 4, and help improve the intelligence and self-adaptation capabilities of the system, making it more adaptable to complex medical environments and application requirements.

[0075] Among them, when the cloud server 5 sends an alarm signal, it can send the alarm signal to the smart terminal 4, and the smart terminal 4 can display and / or broadcast the alarm signal. It can also send the alarm signal to the user end (that is, the terminal of the service platform of the cloud server 5 through an account, and the user end can be a mobile terminal) to facilitate the user to remotely monitor the hemodialysis process.

[0076] In addition, the cloud server 5 can also send the identification results of the dehydration amount and treatment mode to the user end, so that the user can remotely monitor the hemodialysis process.

[0077] In some embodiments, see Figure 3 The intelligent terminal 4 includes a main controller 401, a warning light 402, a display 403, a memory 404 and a wireless transceiver module 405, wherein the electronic balance 2, the environmental sensor 3, the warning light 402, the display 403, the memory 404 and the wireless transceiver module 405 are all electrically connected to the main controller 401. The warning light 402 is used to emit a light signal according to the alarm signal under the control of the main controller 401, and the warning light 402 can be an LED light, but is not limited thereto. The display is used to display the alarm signal. The wireless transceiver module 405 is used to communicate with the cloud server 5.

[0078] In summary, the hemodialysis machine dehydration measurement method and system provided in this application have the following advantages: 1. It can avoid the influence of environmental factors such as temperature, humidity, pressure and air velocity on the measurement results, and improve the accuracy and stability of the measurement results; 2. Use Kalman filter to filter the mass change and environmental impact parameters, and dynamically adjust the sampling frequency of mass change according to the mass change rate, so as to correct the data in real time, respond quickly to mass changes, and ensure the real-time and accuracy of the data; 3. Through the fast Fourier transform algorithm and LMS filter, the periodic interference components can be effectively identified and eliminated, significantly improving the signal quality and the reliability of the measurement results; 4. The automatic measurement of dehydration volume is realized, which greatly reduces the complexity of manual intervention and operation; 5. By utilizing the computing power of the cloud server 5, real-time data analysis can be performed and remote monitoring can be supported, thereby improving the manageability and intelligence level of the system.

[0079] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0080] In addition, the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0081] Furthermore, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0082] In this document, relational terms such as first and second, etc. are used merely to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0083] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for measuring the dehydration amount of a hemodialysis machine, based on measuring the dehydration amount using an electronic balance, on which a container containing pure water is placed, and the arteriovenous end of a blood circuit connected to the hemodialysis machine according to a preset treatment mode is immersed in the container; characterized in that: The method for measuring the dehydration amount of a hemodialyzer comprises the steps of: A1. Real-time acquisition of the mass change of the pure water in the container measured by the electronic balance, and real-time acquisition of environmental impact parameters around the electronic balance; A2. Preprocessing the mass change and the environmental impact parameters; A3. According to the pre-processed environmental impact parameters, the measurement error of the electronic balance is obtained by using the pre-trained measurement error recognition model to correct the mass change after pre-processing; A4. Eliminate the periodic interference components in the corrected mass variation based on the fast Fourier transform algorithm and LMS filter; A5. Calculate the dehydration amount based on the mass change after eliminating the periodic interfering components.

2. The method for measuring dehydration of a hemodialyzer according to claim 1, characterized in that: The environmental influencing parameter includes at least one of temperature, humidity, pressure, and air flow rate.

3. The method for measuring dehydration of a hemodialyzer according to claim 1, characterized in that: Step A3 includes: Inputting the preprocessed environmental impact parameter into a pretrained measurement error recognition model to obtain a measurement error output by the pretrained measurement error recognition model; The measurement error is subtracted from the pre-processed mass change to obtain the corrected mass change.

4. The method for measuring dehydration of a hemodialyzer according to claim 1, characterized in that: Step A4 includes: Using the first sliding window to intercept the corrected mass change amount to obtain first intercepted data; Performing a fast Fourier transform on the first intercepted data to obtain frequency domain data corresponding to the first intercepted data; Extracting periodic component frequency information according to the frequency domain data; According to the periodic component frequency information, the periodic interference component in the corrected mass variation is filtered out by using an LMS filter to obtain the mass variation after eliminating the periodic interference component.

5. The method for measuring dehydration of a hemodialyzer according to claim 1, characterized in that: Step A3 and later also include: Obtaining the mass change rate according to the corrected mass change; The sampling frequency of the electronic balance is adjusted according to the mass change rate.

6. The method for measuring dehydration of a hemodialyzer according to claim 1, characterized in that: After step A5, the method further includes the following steps: identifying a treatment pattern using a deep learning model based on the amount of dehydration; Comparing the recognition result of the treatment mode with the preset treatment mode to determine whether the system is abnormal; An alarm signal is issued when the system is abnormal.

7. The method for measuring dehydration of a hemodialyzer according to claim 6, characterized in that: The deep learning model is an LSTM neural network model.

8. A hemodialysis machine dehydration measurement system, characterized in that: It includes a hemodialysis machine, an electronic balance, an environmental sensor, an intelligent terminal and a cloud server, wherein a container containing pure water is placed on the electronic balance, the hemodialysis machine and the blood circuit are connected according to a preset treatment mode, the arteriovenous end of the blood circuit is immersed in the container, and the electronic balance, the environmental sensor and the cloud server are all connected to the intelligent terminal for communication; The electronic balance is used to measure the mass change of the pure water in the container in real time; The environmental sensor is used to measure the environmental impact parameters around the electronic balance in real time; The intelligent terminal is used to preprocess the mass change and the environmental impact parameter, and obtain the measurement error of the electronic balance according to the preprocessed environmental impact parameter using a pretrained measurement error recognition model to correct the preprocessed mass change, eliminate the periodic interference component in the corrected mass change based on a fast Fourier transform algorithm and an LMS filter, calculate the dehydration amount according to the mass change after eliminating the periodic interference component, and send the dehydration amount to the cloud server; The cloud server is used to store the dehydration amount.

9. The hemodialysis machine dehydration measurement system according to claim 8, characterized in that: The environmental influencing parameter includes at least one of temperature, humidity, pressure, and air flow rate, and the environmental sensor correspondingly includes at least one of a temperature sensor, a humidity sensor, a pressure sensor, and an air flow rate sensor.

10. The hemodialysis machine dehydration measurement system according to claim 8, characterized in that: The cloud server is also used to identify the treatment mode according to the dehydration amount using a deep learning model, compare the identification result of the treatment mode with the preset treatment mode, determine whether the system is abnormal, and send an alarm signal when the system is abnormal.

Images