Methods and systems for ventilators

By establishing a modeling-based parameter difference tracking method in the ventilator, and utilizing a predictor-corrector and edge computing framework, the reliability of the ventilator can be monitored and predicted in real time. This solves the problem of sensor and actuator deviation in environments with changing humidity, and improves the reliability of the ventilator and patient comfort.

CN116139375BActive Publication Date: 2026-11-17GE PRECISION HEALTHCARE LLC
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202211421946.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2021-11-23
Filing Date
2022-11-14
Publication Date
2026-11-17
Estimated Expiration
2042-11-14

AI Technical Summary

Technical Problem

In intensive care units, the sensors and actuators of ventilators are prone to deviations in environments with changing humidity, causing patient discomfort. Existing technologies struggle to effectively monitor and predict these deviations to ensure the reliability of ventilators.

Method used

By establishing a method for tracking the difference between modeled ventilator parameters and sensed parameters, using a predictor-corrector and edge computing framework, combined with cloud networks and swarm models, the reliability of the ventilator can be monitored and predicted in real time, and operating parameters can be adjusted to reduce deviations.

Benefits of technology

It improves the reliability of ventilators and patient comfort, reduces discomfort caused by sensor and actuator deviations, and improves the maintenance efficiency of mechanical ventilators.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116139375B_ABST
    Figure CN116139375B_ABST
Patent Text Reader

Abstract

This paper provides systems and methods for analyzing the performance of mechanical ventilators. In one example, one method includes tracking the differences over time between modeled ventilator parameters and sensed ventilator parameters, and further predicting ventilator reliability during operation based on ventilator utilization.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the subject matter disclosed herein relate to the analysis of ventilator performance, and more specifically to the reliability of ventilator components, including sensors and actuators. Background Technology

[0002] Some types of medical equipment (such as incubators and anesthesia machines) may include advanced respiratory systems for directing breathable gases into and out of a patient's lungs. Advanced respiratory systems may include devices for monitoring the airflow rate of a patient's breathing. In some examples, the device may include a flexible reed perpendicular to the airflow path, sensitive to changes in the airflow rate through the device. For example, the flexible reed may bend bidirectionally in proportion to the mass flow rate of each breath, such as bending in a first direction during inhalation and in the opposite second direction during exhalation. The device may include two pressure ports, one on each side of the flexible reed, to determine the pressure difference across the reed. This pressure difference can be used to determine the mass flow rate of breathing.

[0003] Ventilators providing air to patients without spontaneous breathing are a common necessity in intensive care units (ICUs). In critical care settings, patient comfort can depend on the performance and / or consistency of components within the ventilator system architecture, such as sensors and actuators. In such cases, even a decrease in performance and / or consistency, even for a few seconds, can lead to patient discomfort. In one example, during bed transfers between the ICU and surgical wards, the oxygen sensor in the ventilator may be susceptible to humidity variations, which could increase sensor skew due to frequent changes in humidity levels. The ventilator may operate in less controlled environments where humidity can change frequently. Some more advanced ventilators may incorporate redundant sensors, improving the accuracy of ventilator components, but at a higher manufacturing cost. Summary of the Invention

[0004] In one embodiment, a method for a ventilator includes: predicting the reliability of the ventilator during operation based on time-tracked differences between modeled ventilator parameters and sensed ventilator parameters, and ventilator utilization.

[0005] It should be understood that the above brief description is provided to introduce selected concepts further described in the detailed embodiments in a simplified form. This is not intended to identify key or essential features of the claimed subject matter, the scope of which is uniquely defined by the claims following the detailed embodiments. Furthermore, the claimed subject matter is not limited to embodiments that address any shortcomings mentioned above or in any part of this disclosure. Attached Figure Description

[0006] The invention will be better understood by referring to the following description of non-limiting embodiments, in which:

[0007] Figure 1 A system for a mechanical ventilator is shown.

[0008] Figure 2 A state machine based on a ventilator model and one or more sensor signals is shown, wherein the transition from a first state to a second state can be based on the reliability of the ventilator's predictions.

[0009] Figure 3 A schematic diagram of an exemplary predictor-corrector operation based on a ventilator model and sensor signals is shown.

[0010] Figure 4 A schematic diagram of the aerodynamic architecture of an existing ventilator system is shown.

[0011] Figure 5 An exemplary edge cloud architecture for predicting sensor bias based on environmental conditions tracked over time, utilization levels, and a group model is shown.

[0012] Figure 6 It shows the basis Figure 5 The example described in the document is an edge cloud architecture for condition-based maintenance of ventilator groups. Detailed Implementation

[0013] The following description relates to various implementation schemes of methods and systems for predicting the reliability of mechanical ventilators in ventilator systems. Figure 1 An example of such a ventilator system employing a reliability prediction approach is shown. This ventilator system includes an adaptive ventilator model and multiple on-premises and cloud-based functionalities supporting the management of the mechanical ventilator. The method may include state machine control of ventilator operation and tracking ventilator reliability through a predictor-corrector architecture, an example of which is shown in [example missing]. Figure 2 As shown in the image. Figure 3 The figure provides an exemplary illustration of reliability prediction based on predicted sensor bias and sensor failure. Figure 4 The diagram illustrates an exemplary mechanical ventilator aerodynamic architecture for which adaptive ventilator models and state machine control operations can be developed. Figure 5 The image shows a mechanical ventilator communicating with a supervisory system that uses an edge cloud architecture to track sensor bias and utilization over time. Figure 6 This paper illustrates an exemplary condition-based maintenance for a ventilator group based on the reliability of predictions made for each ventilator in the group over time. As a result of the systems and methods described herein, the ease of use of ventilator systems may be increased, mechanical ventilator maintenance may become more efficient, and the likelihood of patient discomfort may be reduced.

[0014] Embodiments of this disclosure will now be described by way of example with reference to the accompanying drawings, wherein Figure 1 A block diagram of system 100 is shown. In the illustrated embodiment, system 100 is a mechanical ventilator system for mechanical ventilator 115. Mechanical ventilator 115 is connected to components for detecting and displaying reliability predictions in system 100.

[0015] The mechanical ventilator 115 can receive input from an operator (e.g., a doctor, nurse), such as parameters for error value detection, thresholds for sensor and / or actuator bias, etc. The mechanical ventilator 115 can be connected to a ventilator model 102, wherein input from the operator to the mechanical ventilator 115 can also be input to the ventilator model 102 as parameters for calculations performed by the ventilator model 102. The mechanical ventilator 115 can be connected to a predictor-corrector 116, wherein input from the operator to the mechanical ventilator 115 and readings from the mechanical ventilator 115 (such as oxygen level, positive end-expiratory pressure (PEEP), etc.) can also be input to the predictor-corrector 116. The mechanical ventilator 115 can also be connected to a ventilator monitoring and control system 118, wherein any readings from the mechanical ventilator 115 can be input to the ventilator monitoring and control system 118 for monitoring and correcting any potential errors detected in the components of the mechanical ventilator 115.

[0016] In some embodiments, an operator may additionally or alternatively control one or more operating parameters of the mechanical ventilator 115 via an electronic controller 114. The controller 114 includes a processor operatively connected to a memory. The memory may be a non-transitory computer-readable medium and may be configured to store computer-executable code (e.g., instructions) that will be processed by the processor to execute one or more routines (such as those described herein). The memory may also be configured to store data received by the processor. The controller 114 may be communicatively coupled (e.g., via a wired or wireless connection) to one or more external or remote computing devices, such as a hospital computing system, and may be configured to send and receive various types of information, such as electronic medical record information, procedural information, etc. The controller 114 may also be electrically coupled to various other components of the mechanical ventilator 115.

[0017] The controller 114 receives signals from various sensors of the mechanical ventilator 115 and uses various actuators of the mechanical ventilator 115 to adjust the operation of the mechanical ventilator 115 based on the received signals and instructions stored in the memory of the controller 114. For example, the flow of gas to the intake port can be controlled via an input device (e.g., keyboard, touch screen, etc.) connected to the electronic controller of the mechanical ventilator 115. The controller 114 can display the operating parameters of the mechanical ventilator 115 via a GUI display of the ventilator interface 120. The controller 114 can receive signals (e.g., electrical signals) via input devices and can adjust the operating parameters of the mechanical ventilator 115 in response to (e.g., acknowledgment) the received signals.

[0018] For example, the operator can input the desired concentration of the anesthetic to be delivered to the patient. The corresponding valve positions of one or more valves of the mechanical ventilator (e.g., the positions of one or more bypass valves) can be determined empirically and stored in a predetermined lookup table or function in the controller's memory. For example, the controller can receive the desired anesthetic concentration via an input device and determine the opening amount of one or more valves corresponding to the desired anesthetic concentration based on a lookup table, where the input is the anesthetic concentration and the output is the valve position of one or more valves. The controller can transmit electrical signals to the actuators of one or more valves to adjust each of the one or more valves to its corresponding output valve position. In some examples, for instance, the controller can compare the desired gas flow rate with a measured gas flow rate, such as via an inspiratory flow sensor.

[0019] Although controller 114 is Figure 1 The controller 114 is shown for illustrative purposes, but it should be understood that it may be located at various locations within, around, and / or away from the mechanical ventilator 115. For example, the controller 114 may include multiple devices / modules that may be distributed throughout the mechanical ventilator 115. Thus, the controller 114 may include multiple controllers at various locations within the mechanical ventilator 115. As another example, additionally or alternatively, the controller 114 may include one or more devices / modules located outside the mechanical ventilator 115, near the mechanical ventilator 115 (e.g., in the same room), or away from the mechanical ventilator (e.g., a remote server). In each example, the multiple devices / modules may be communicatively coupled via wired and / or wireless connections.

[0020] Ventilator model 102 may be a semi-empirical, physics-based model running in real time on a processor (e.g., the processor of controller 114) to estimate the current state of sensors and / or actuators of mechanical ventilator 115. In one example, ventilator model 102 may use multidimensional linearized equations derived from mass, energy, and momentum balance equations, where parameters and linearization are performed in real time via machine learning methods. Ventilator model 102 may use information obtained from components of mechanical ventilator 115 (e.g., high-pressure passage, blower speed, blower afterpressure value) to generate estimates of sensor and / or actuator functionality. Ventilator model 102 may be embedded in or near ventilator system 100 via a computing module. In some examples, input to the model may be obtained from mechanical ventilator 115 or via a bidirectional cloud connection to cloud network 108. Ventilator model 102 can be connected to predictor-corrector 116, where one or more estimates calculated by ventilator model 102 (such as oxygen level estimate, PEEP estimate, error threshold estimate, etc.) can be input to predictor-corrector 116. Ventilator model 102 can also be connected to edge computing box 101, where one or more calculations performed by ventilator model 102 can be input to edge computing box 101, thus allowing evaluation of the performance of ventilator model 102. In one example, edge computing box 101 can be an edge processor.

[0021] Edge computing frame 101 may have a bidirectional cloud connection to cloud network 108, wherein any diagnostics from mechanical ventilator 115, ventilator monitoring and control system 118, and ventilator model 102 can be input into cloud network 108 for data aggregation with other ventilator systems via cloud network 108. Cloud network 108 may input error thresholds and other parameters for ventilator component readings into edge computing frame 101, enabling edge computing frame 101 to communicate with other components in ventilator system 100 given data shared via the bidirectional cloud connection. Edge computing frame 101 may be connected to ventilator monitoring and control system 118, wherein any error thresholds from cloud network 108 can be input into ventilator monitoring and control system 118, and any diagnostics from ventilator monitoring and control system 118 can be input into edge computing frame 101 communicating via the bidirectional cloud connection to cloud network 108.

[0022] Cloud network 108 may include mechanical ventilator group data from multiple mechanical ventilators, which may include readings under various conditions (e.g., ambient environment, utilization technology, etc.). Cloud network 108 may also include multiple group models to facilitate sensor bias calculations performed by the system to detect local sensor bias. Multiple group models can be considered as sensor bias models based on group data, where each group model can distinguish bias accumulation caused by environmental conditions of the ventilators (e.g., humidity variations) or utilization (e.g., misuse due to lack of experience). Tracking ventilator utilization over time (such as how bias can accumulate) may help classify the reliability of local sensors for the ventilator group. Multiple group models can decouple bias correction from environmental conditions and / or utilization, and can transmit fault thresholds and fault threshold-based operating modes to edge computing box 101. In this way, multiple group models can separately consider and categorize changes in data from mechanical ventilators' sensors and / or actuators caused by environmental conditions and / or utilization, as well as changes in data from mechanical ventilators' sensors and / or actuators caused by corrections performed by the system or operator. The accumulated group deviation data categorized by environmental conditions and utilization can be deployed to edge computing box 101 to adapt to ventilator model 102, thereby improving the reliability of sensors and other components of the ventilators in the group. Multiple group models can indicate any ventilator connected to the group that is approaching failure (e.g., conditions causing patient discomfort) based on environmental conditions and / or utilization, meaning that a ventilator approaching failure can enter a fault-tolerant operating mode to accommodate any potential failures. A ventilator is considered to be approaching failure based on conditions such as the absolute value of the difference between the data from the ventilator's sensors and / or actuators and a failure threshold increasing less than a predetermined number of times. Multiple group models can differentiate ventilators connected to a group based on their proximity to the triggering fault, where ventilators closer to the triggering fault can be classified differently from those that may be new ventilators or may not be close to the triggering fault. In this way, cloud network 108 can use multiple group models to help determine the health status of ventilators based on environmental conditions, and can increase the efficiency of any services that may occur on ventilators due to the use of multiple group models to determine the health status of ventilators.

[0023] The predictor-calibrator 116 compares real-time sensed outputs (such as oxygen levels and / or PEEP levels) from the mechanical ventilator 115 with the outputs (e.g., predictive estimates) from the ventilator model 102 to determine potential prediction biases in the readings from the mechanical ventilator 115 due to component errors. The predictor-calibrator 116 can be connected to the ventilator interface 120, where readings from the mechanical ventilator 115 and estimates from the ventilator model 102 can be input to the ventilator interface 120 for display. The real-time determinations made by the predictor-calibrator 116 can be displayed on the GUI of the ventilator interface 120. The predictor-calibrator 116 can be connected to the ventilator monitoring and control system 118. In one example, the output of the predictor-calibrator 116 can be monitored by the ventilator monitoring and control system 118, for example, based on a comparison or filtered reading of the difference between the readings from the mechanical ventilator 115 and the estimates from the ventilator model 102. The ventilator monitoring and control system 118 can communicate with the mechanical ventilator 115 and the edge computing box 101. In some examples, the ventilator monitoring and control system 118 can provide input to the edge computing box 101 to potentially adapt to the ventilator model 102 and communicate with the cloud network 108.

[0024] The ventilator interface 120 can display sensor readings from the mechanical ventilator 115 and estimates from the ventilator model 102 on a GUI. In one example, a warning indication (e.g., an alarm) can be displayed to the operator if the difference in sensor values ​​exceeds an error threshold, prompting any potential operator-based corrections and / or calibrations to components of the mechanical ventilator. In one example, the ventilator interface 120 can display oxygen levels, PEEP levels, or other parameters indicating patient comfort (e.g., a patient comfort index).

[0025] Figure 2 The diagram shows a device used for tracking mechanical ventilation (such as...) Figure 1 An exemplary state machine 200 for predicting reliability in a mechanical ventilator 115. This state machine can be used with a ventilator monitoring and control system (such as...) Figure 1 Ventilator monitoring and control system 118), ventilator model (such as Figure 1 The ventilator model 102) and signals from sensors and / or actuators of the mechanical ventilator (such as those described below) Figure 4 The signals from those sensors and / or actuators described in [the document / document]). The ventilator monitoring and control system can be local or from a cloud network (such as [other systems]). Figure 1The cloud network 108) receives a fault threshold. In one example, the reliability prediction of the mechanical ventilator during operation can be compared with a fault threshold, such that if any reliability prediction of the mechanical ventilator exceeds the associated fault threshold, the mechanical ventilator can transition from one state to another. The reliability prediction may include differences indicating sensor bias, actuator bias, gas leakage, etc., which are caused by a variety of factors, including but not limited to component aging and / or degradation, improper use, and changes in environmental conditions. In one example, upon detecting a difference, the mechanical ventilator can transition to a biased or degraded state, where the ventilator monitoring and control system can operate under a warning indicating degradation. In another example, the ventilator monitoring and control system can adjust ventilator operation based on a comparison of observed data with expected data supplied from the ventilator model and one or more thresholds. In yet another example, the ventilator monitoring and control system can generate actions for the operator, such as adjusting the operation of the mechanical ventilator. In some examples, even if a fault is detected, restricted ventilator operation can continue in a degraded operating state by including corrections to enable fault tolerance. When a more severe discrepancy is detected, the ventilator may be shut down based on exceeding one or more high-severity fault thresholds. In some examples, shutdown may include deactivation of ventilator operation. In one example, when a fault is detected, the operator can be notified by presenting a notification on the GUI display, allowing the operator to manually adjust operating parameters to correct the fault if the mechanical ventilator enters a deviation and / or degradation state.

[0026] When the mechanical ventilator operates with standard operating parameters (referred to herein as Normal Mode 201), it can be considered to be operating in normal operating condition. Normal Mode 201 may include all standard and typical behaviors of the mechanical ventilator, such as supplying a predetermined amount of oxygen to the patient, measuring the oxygen level in the mechanical ventilator via an oxygen sensor, adjusting the pressure of the oxygen supplied to the patient, filtering the oxygen supplied to the patient, measuring the pressure level in the mechanical ventilator via a pressure sensor, filtering bacteria during the patient's inhalation and exhalation, etc. The normal operating condition parameters that may be included in Normal Mode 201 can be predetermined by the ventilator monitoring and control system, and the operator can verify the normal operating condition parameters included in Normal Mode 201. During operation, data supplied by sensors and actuators in the mechanical ventilator can be displayed on an associated GUI display (such as...). Figure 1 On the ventilator interface 120).

[0027] Based on the pressure deviation detected by the system, the predictive reliability of the mechanical ventilator can transition from normal mode 201 to a pressure deviation state (referred to herein as pressure deviation mode 204), indicated by arrow 202. The detected pressure deviation may be caused by frequent changes in environmental conditions (e.g., humidity), and the pressure sensor included in the mechanical ventilator may show deviations in sensor readings. In one example, an airflow sensor configurable within the mechanical ventilator can be used to measure the pressure. The airflow sensor may include a thin, flexible metal reed coupled perpendicular to the airflow path of the airflow sensor. This reed deflects proportionally to the airflow through the airflow path, and the pressure at both ends of the reed (e.g., determined based on pressure measurements upstream and downstream of the reed) can be used to determine the airflow rate. For example, the airflow sensor can be calibrated during manufacturing to determine the relationship between the pressure at both ends of the reed and the flow rate, which can be referenced to determine the airflow rate during the use of the mechanical ventilator. However, this relationship may change over time due to metal fatigue of the reed and other sources of calibration drift. Therefore, the accuracy of the flow rate determined based on the pressure at both ends of the reed may change over time. As a predictor-corrector (such as...) Figure 1 The predictor-calibrator (116) and its architecture compare pressure measurements from sensors with estimated pressure measurements from a ventilator model, detecting pressure deviations via pressure sensors or airflow sensors. Figure 1 The edge computing box 101) allows a comparison of pressure measurements from sensors with predicted pressure measurements to be input into the group model for pressure deviations due to humidity, in order to determine whether the mechanical ventilator is nearing or in a fault. If the group model determines that the mechanical ventilator is not nearing a fault, the mechanical ventilator can switch to normal mode 201 without taking any additional generated actions to prevent patient discomfort, as indicated by arrow 203.

[0028] If a fault threshold set by the group model in the cloud network is exceeded, the mechanical ventilator may transition from pressure deviation mode 204 to a degraded operating state (referred to herein as first fault mode 206), indicated by arrow 205. Exceeding the fault threshold may include a value that is normally greater than the fault threshold but is now less than the fault threshold, or a value that is normally less than the fault threshold but is now greater than the fault threshold, both of which may lead to patient discomfort. If the group model determines that the mechanical ventilator is nearing a fault while in pressure deviation mode 204, the mechanical ventilator may also transition to first fault mode 206 as a precaution against potential future patient discomfort. First fault mode 206 may indicate to the operator on the GUI display that first fault mode 206 is active, and may also include displaying to the operator fault-related model outputs associated with first fault mode 206 and readings from any fault-related sensors, allowing the operator to make corresponding corrections to offset reliability and prevent patient discomfort. The operator may monitor sensor measurements and component behavior while the mechanical ventilator is in first fault mode 206, and the operator may determine when the mechanical ventilator can transition to normal mode 201. In one example, prompts can be presented to the operator on the GUI display to allow the operator to indicate when the mechanical ventilator can be switched to normal mode 201.

[0029] Based on actuator deviation detected by the system, the mechanical ventilator can transition from normal mode 201 to actuator deviation state (referred to herein as actuator deviation mode 209), indicated by arrow 207. Detected actuator deviation can be caused by a variety of conditions, including but not limited to changes in the overall flow rate of the mechanical ventilator, component degradation, or misuse. In one example, the controller (such as...) Figure 1The controller 114) can transmit electrical signals to the actuators of one or more valves in the mechanical ventilator to adjust each of the one or more valves to a corresponding output valve position, and component degradation can result in a difference between the indicated output valve position and the actual output valve position, which can lead to the detection of actuator deviation. In some examples, the controller can compare a desired gas flow rate with a measured gas flow rate (such as that measured by an inhalation flow sensor), and actuator deviation can be detected if the absolute value of the difference between the desired gas flow rate and the measured gas flow rate exceeds a threshold amount over a threshold duration. In one example, an air flow sensor configured within the mechanical ventilator can be used to determine actuator deviation, as described above. The air flow sensor calibration at manufacturing may change over the ventilator's lifespan due to metal fatigue of the reed and other sources of calibration drift. Therefore, the accuracy of the flow rate determined based on the pressure across the reed may change over time. Actuator deviation can be detected as a result of the controller using a lookup table to compare known values ​​of expected pressure and flow rate for a given expected input voltage with actual values ​​of pressure and flow rate for a given actual input voltage. As a result of comparing airflow measurements from sensors with estimated airflow measurements from the ventilator model in the predictor-corrector architecture, flow rate deviations via pressure or airflow sensors can be detected. The predictor-corrector outputs a comparison of the two measurements to an edge computation box, where the comparison between the airflow measurement from the sensor and the predicted airflow measurement can be input into a group model for actuator deviations due to airflow, to determine whether the mechanical ventilator is nearing or in a fault state. If the group model determines that the mechanical ventilator is not nearing a fault, the mechanical ventilator can switch to normal mode 201, and no generated action to prevent patient discomfort may be taken, as indicated by arrow 208. In one example, the lookup table used for detecting actuator deviations can be recalibrated with new values ​​and / or value ranges based on the current instance of actuator deviation detection to reduce the number of instances where the mechanical ventilator switches from actuator deviation mode 209 to normal mode 201 due to inaccurate actuator deviation detection.

[0030] If a fault threshold set by the group model in the cloud network is exceeded, the mechanical ventilator may transition from actuator deviation mode 209 to a second degraded state (referred to herein as second fault mode 211), indicated by arrow 210. Exceeding the fault threshold may include a value that is normally greater than the fault threshold but is now less than the fault threshold, or a value that is normally less than the fault threshold but is now greater than the fault threshold, both of which may lead to patient discomfort. If the group model and / or lookup table determine that the mechanical ventilator is close to failure while in actuator deviation mode 209, the mechanical ventilator may also transition to second fault mode 211 as a precaution against potential future patient discomfort. Second fault mode 211 may indicate to the operator on the GUI display that second fault mode 211 is active, and may also include displaying to the operator fault-related model outputs associated with second fault mode 211 and readings from any fault-related sensors, allowing the operator to make corresponding corrections to counteract reliability issues and prevent patient discomfort. The operator can monitor sensor measurements and component behavior when the mechanical ventilator is in the second fault mode 211, and can determine when the mechanical ventilator can switch to normal mode 201. In one example, a prompt can be presented to the operator on the GUI display to allow the operator to indicate when the mechanical ventilator can switch to normal mode 201.

[0031] Based on the oxygen level detected by the mechanical ventilator, the mechanical ventilator can transition from normal mode 201 to an oxygen estimation state (referred to herein as oxygen estimation mode 214), indicated by arrow 212. The oxygen level detected by the mechanical ventilator may be outside the range of values ​​representing the expected oxygen level of the mechanical ventilator, thus transitioning the mechanical ventilator to oxygen estimation mode 214. Because the mechanical ventilator may operate in less controlled environments, such as where humidity variations are high and interfere with oxygen probe functionality, the oxygen level detected by the mechanical ventilator may be outside the range of expected oxygen levels. Inaccurate oxygen measurements may be detected as a result of comparing oxygen measurements from the sensor with estimated oxygen measurements from the ventilator model, as a result of the predictor-corrector architecture. The predictor-corrector outputs a comparison of the two measurements to an edge computation box, where the comparison of the oxygen measurement from the sensor with the predicted oxygen measurement can be input into a group model for oxygen bias due to utilization, to determine whether the mechanical ventilator is nearing or already faulty. If the group model determines that the mechanical ventilator is not approaching a fault, the mechanical ventilator can switch to normal mode 201 and no generated action is required to prevent patient discomfort, as indicated by arrow 213.

[0032] If oxygen levels are detected to be outside the first oxygen threshold, the mechanical ventilator can transition from oxygen estimation mode 214 to oxygen standby mode (referred to herein as oxygen standby mode 217), indicated by arrow 215. In oxygen standby mode, an oxygen sensor malfunction may be detected as a result of comparing oxygen measurements from the sensor with estimated oxygen measurements from the ventilator model, as part of a predictor-corrector architecture. The predictor-corrector can output a comparison of the two measurements to an edge computation box, where the comparison between the oxygen measurement from the sensor and the predicted oxygen measurement can be input into the group model for oxygen bias due to utilization, to determine whether the mechanical ventilator is close to or already in a second malfunction. In one example, oxygen standby mode could be considered a difference of higher severity due to sensor reliability.

[0033] If oxygen supply is detected to be outside a higher severity threshold, the mechanical ventilator may switch from oxygen standby mode 217 to shutdown (referred to herein as third fault mode 219), indicated by arrow 218. Exceeding the higher severity oxygen threshold may include a value that is normally greater than the threshold but is now less than the threshold, or a value that is normally less than the threshold but is now greater than the threshold, both of which may lead to patient discomfort. If the group model and / or lookup table determine that the mechanical ventilator is nearing failure while in oxygen standby mode 217, the mechanical ventilator may also switch to third fault mode 219 as a precaution against potential future patient discomfort. The mechanical ventilator may also switch from normal mode 201 to third fault mode, indicated by arrow 216. Third fault mode 219 may indicate to the operator on the GUI display that third fault mode 219 is active, and may also include fault-related diagnostics associated with third fault mode 219 and instructions for using a different mechanical ventilator. In one example, instructions for a mechanical ventilator in third fault mode 219 may include discontinuing ventilator operation until maintenance is performed. In another example, the operator can monitor sensor measurements and component behavior when the mechanical ventilator is in a third fault mode, and the operator can determine when the mechanical ventilator can switch to normal mode 201. In one example, a prompt can be presented to the operator on a GUI display to allow the operator to indicate when the mechanical ventilator can switch to normal mode 201.

[0034] Based on the amount of gas leakage detected in the ventilator tubing, the mechanical ventilator can transition from normal mode 201 to a gas leakage state (referred to herein as gas leakage mode 221), indicated by arrow 220. In one example, the gas leakage may include a leak of oxygen gas. The amount of gas leakage detected in the ventilator tubing may be outside the range of values ​​representing the expected gas leakage amount, thus transitioning the mechanical ventilator to gas leakage mode 221. The detected gas leakage amount may be outside the range of values ​​representing the expected gas leakage amount due to conditions such as degradation or misuse of components in the flow path (e.g., valves, pressure regulators, flow limiters). In one example, the controller may transmit electrical signals to the actuators of one or more valves in the mechanical ventilator to adjust each of the one or more valves to a corresponding output valve position, and component degradation may result in a difference between the indicated output valve position and the actual output valve position, which can lead to a gas leakage. In some examples, the controller can compare a desired gas pressure with a measured gas pressure (such as that measured by an inspiratory gas pressure sensor), and detect a gas leak if the absolute value of the difference between the desired and measured gas pressures exceeds a threshold amount over a threshold duration. Gas leaks can also be detected as a result of the controller using a lookup table to compare a known value of the expected pressure for a given expected adjustment with the actual values ​​of pressure and flow rate for a given actual adjustment. Gas leaks via pressure sensors or other sensors, such as airflow, can be detected as a result of the predictor-corrector architecture comparing gas pressure measurements from sensors with estimated pressure measurements from a ventilator model. The predictor-corrector outputs a comparison of two measurements to an edge computation box, where the comparison of the gas pressure measurement from the sensor with the predicted gas pressure measurement can be input into a group model for gas leaks due to gas pressure to determine if the mechanical ventilator is nearing or already faulty. In one example, gas leak mode 221 could be considered a degraded operating state and a higher severity fault type, causing the mechanical ventilator to briefly remain in gas leak mode. In gas leak mode, the mechanical ventilator can increase the gas intake force to compensate for the leak and prevent patient discomfort. In one example, the lookup table used to detect gas leaks can be recalibrated with new values ​​and / or value ranges based on the current instance of gas leak detection to estimate the amount of gas pressure to be increased to maintain an appropriate gas flow rate.

[0035] The mechanical ventilator can transition from a gas leak mode to a shutdown state (referred to herein as fourth fault mode 223), indicated by arrow 222. Fourth fault mode 223 can be indicated to the operator on the GUI display as active, and may also include fault-related diagnostics associated with fourth fault mode 223 and instructions for using a different mechanical ventilator. In one example, instructions for a mechanical ventilator in fourth fault mode 223 may include stopping ventilator operation until maintenance is performed. In another example, the operator can monitor sensor measurements and component behavior while the mechanical ventilator is in fourth fault mode, and the operator can determine when the mechanical ventilator can transition to normal mode 201. In one example, a prompt can be presented to the operator on the GUI display to allow the operator to indicate when the mechanical ventilator can transition to normal mode 201.

[0036] Turn now Figure 3 A schematic diagram 300 illustrates an exemplary method for reliability prediction and correction in a ventilator system. In this example, a deviation state is detected based on the difference between modeled ventilator parameters and sensed ventilator parameters, which indicates a change in the reliability of the mechanical ventilator's sensors, such as reference parameters. Figure 2 The state machine 200 describes this. It uses aggregated data from multiple mechanical ventilators connected in a cloud network to diagnose and correct sensor bias conditions. (See above reference.) Figure 2 As described, in some examples, mechanical ventilators may operate in environments with low levels of control, such as during patient transfer from the ICU to a surgical ward, where humidity varies considerably. Moisture in the air may be absorbed into the sensor cable of the oxygen probe, affecting the probe's capacitance during measurement. In some examples, moisture absorption may result in in-range or out-of-range malfunctions of the sensor probe. Oxygen bias can be diagnosed by comparing the actual oxygen level detected by one or more sensors on the mechanical ventilator with an oxygen estimate generated by a semi-empirical model. Bias or creep from one or more sensors can be diagnosed with respect to utilization and operating (e.g., environmental) conditions by analyzing oxygen sensor behavior as a function of various environmental conditions aggregated with data from other ventilator systems in a cloud network. Model outputs (such as information related to sensor reliability) and instructions (such as corresponding corrections to counteract bias or creep) can be displayed on the ventilator interface.

[0037] Schematic diagram 300 includes oxygen measurement 302. Oxygen measurement 302 is from one or more sensors of the mechanical ventilator (such as those referenced below). Figure 4 Examples of sensed ventilator parameters obtained by the described sensor. One or more sensors in the mechanical ventilator can provide oxygen measurement, such as inspiratory flow and / or pressure sensors.

[0038] Schematic diagram 300 includes oxygen estimate 304. Oxygen estimate 304 is an example of modeled ventilator parameters estimated through a semi-empirical, physics-based model. In one example, it may be based on the ventilator aerodynamic architecture (such as those referenced below). Figure 4 The model is developed using a digital model of the described ventilator aerodynamic architecture 400. In one example, the model uses a multidimensional linearized equation established by mass, energy, and momentum balance equations, where parameters and linearization are performed in real time via machine learning methods. In another example, a semi-empirical, physics-based model estimates the expected oxygen flow rate based on pressure, airflow, and / or other inputs, including data from the oxygen high-pressure channel, blower speed, and blower post-pressure values.

[0039] Oxygen measurement 302 and oxygen estimation 304 are input into the predictor-corrector (such as...) Figure 1 The predictor-corrector (116) described herein is used to estimate oxygen sensor bias using a pair of equations. Oxygen measurement (302) and oxygen estimation (304) are input into a random walk model and a Kalman update. In one example, the random walk model can be computed as follows: Where x k pred and x k est These refer to the deviations predicted and estimated using model and sensor data at time k, respectively. The Kalman update filter model estimate is as follows: (k+1) = (k+1) pred +K·(z meas -z pred ), where the oxygen sensor measures z meas And the oxygen value calculated from the physics-based model is z. pred In one example, the output is an oxygen sensor bias measurement. In another example, the oxygen sensor bias measurement could be the non-zero percentage difference between the measured time-averaged value and the model estimate. In this example, the amount of oxygen measured by 302 is less than a preset non-zero oxygen threshold.

[0040] It can be communicated with representatives in cloud networks (such as...) Figure 1The correction 306, determined by a predictor-corrector, is calculated by comparing oxygen sensor deviation data from multiple ventilators (e.g., ventilator groups) aggregated in the cloud network 108. Data representing ventilator groups (e.g., group-level data) can be accessed via the cloud network 108 by the edge computing box 101 and computed locally. Group-level data may include ventilator behavior information, such as sensor and / or actuator deviations, for various environmental and / or utilization conditions. In one example, the edge computing box 101 computes a group-range analysis algorithm to generate an oxygen deviation curve 308 as a function of humidity based on the group data. The correction 306 can be deployed in the field by the ventilator control 310 of the mechanical ventilator to adjust the mechanical ventilator settings and reduce patient discomfort.

[0041] When analyzing the correlation of oxygen sensor bias as a function of humidity (e.g., curve 308), the oxygen bias estimate of an exemplary ventilator can be compared to an oxygen bias threshold to determine whether the ventilator is nearing or already in a fault. In one example, the ventilator briefly transitions to the oxygen bias estimate state. The knowledge-based representation computed during correction 306 can then be fed back into the supervisory control algorithm to correct the oxygen distribution in the ventilator, thus allowing the ventilator to transition back to a normal state, as referenced. Figure 2 The state machine 200 is described.

[0042] Figure 4 The exemplary pneumatic architecture of the mechanical ventilator 400 is described in more detail below. The mechanical ventilator 400 prepares a mixture of ambient gas and oxygen gas. The prepared gas mixture is pressurized and passed through a series of lines shut off by valves and filters. The filtered gas mixture flows to the patient via a breathing tube. The mechanical ventilator 400 also includes various sensors and actuators to monitor and adjust the distribution of the gas mixture. The mechanical ventilator 400 may be based on one or more semi-empirical, physics-based models (such as...) Figure 1 Modeling of the ventilator 400 (model 102) can be performed. The exemplary methods described herein can be followed to monitor and control the operation of the mechanical ventilator 400.

[0043] The pneumatic architecture includes an ambient air line 403, a low-pressure line 405, a premixing line 407, an upper inhalation line 419a, a lower inhalation line 419b, and an exhalation line 443. The direction of gas flow is indicated by multiple arrows on the described lines. Ambient air 402 is filtered through a standard inlet filter 404. After filtration, ambient air 402 enters the mechanical ventilator 400 via ambient air line 403. Low-pressure (LP) oxygen 406 enters the low-pressure line 405 via port 409. In one example, LP oxygen 406 may flow at a rate of less than 15 liters per minute (LPM). High-pressure (HP) oxygen 408 enters the premixing line 407 via port 411 through a standard inlet filter 412 (e.g., a 20-micron filter). In one example, HP oxygen may have a lower pressure threshold of 242 kPa and an upper pressure threshold of 648 kPa.

[0044] After filtration, HP oxygen 408 continues through pressure regulator 414. In one example, pressure regulator 414 may limit the pressure to 30 pounds per square inch (PSI). After filtration and regulation, HP oxygen 408 continues through rotor flowmeter 416, where the oxygen flow rate is further restricted. In one example, rotor flowmeter 416 may limit HP oxygen 408 to 15 LPM. An LP check valve 410 is located on the low-pressure line 405 downstream of port 409 to prevent backflow of HP oxygen. Low-pressure line 405 is fluidly connected to premix line 407 at rotor flowmeter 416. LP oxygen 406 enters premix line 407 via rotor flowmeter 416.

[0045] Ambient air line 403 is fluidly connected at mixing chamber 418 to upper intake line 419a. Ambient air 402, LP oxygen 406, and HP oxygen 408 are combined in mixing chamber 418. In one example, mixing chamber 418 may have a capacity of 0.5 L. The gas mixture exits mixing chamber 418 and enters upper intake line 419a. The gas mixture is received by blower 420 (e.g., turbine) which operates at a fixed speed (e.g., 10 breaths / L). In one example, blower 420 may be operated by a mechanically coupled electric motor (not shown). The pressure of the gas mixture exiting the blower is regulated by flowing through proportional valve 422. In one example, proportional valve 422 is a voice coil-based proportional valve.

[0046] Multiple sensors measure gas pressure and oxygen along the flow path of the upper intake line 419a. A first pressure sensor 424 is located downstream of the blower 420 and proportional valve 422. In one example, the first pressure sensor 424 measures the pressure of the gas discharged from the blower 420. Subsequently, an oxygen sensor 426 and a check valve 428 are positioned along the upper intake line 419a. A free-breathing valve 430 is located downstream of the check valve 428 along the upper intake line 419a. A second pressure sensor 432 is located downstream of the free-breathing valve 430.

[0047] Following the first pressure sensor 424 and preceding the oxygen sensor 426, the lower intake line 419b separates from the upper intake line 419a. The lower intake line 419b is fluidly connected to the upper intake line 419a. The lower intake line includes a vent resistor 436. A third pressure sensor 439 is adjacent to the outlet 441 of the pneumatic exhalation valve 438. Opposite the pressure sensor 439 is a one-time flow limiter 437 that provides access to ambient air.

[0048] Following the second pressure sensor 432, the gas mixture flows into the inhalation tubing 442 via the upper inhalation line 419a through the bacteria / virus filter 434. The patient 446 can inhale the pressurized gas mixture from this inhalation tubing via a two-way breathing device 447. In one example, the ventilator 400 may be equipped with a 22 mm diameter inhalation tubing 442 for male patients and a 15 mm diameter inhalation tubing 442 for female patients.

[0049] Exhaled waste gas from patient 446 can be guided to exhalation tubing 444 via two-way breathing device 447. The waste gas is filtered through bacterial / viral filter 440. Exhalation tubing 444 is fluidly connected to exhalation line 443 after filter 440. Filtered exhaled gas flows through exhalation line 443 to pneumatic exhalation valve 438, passing through disposable flow limiter 437 to reach the atmosphere. In one example, the exhaled gas pressure can be measured by pressure sensor 439 opposite to the disposable flow limiter.

[0050] In one example, it can be targeted Figure 4One or more semi-empirical, physics-based models are developed for the ventilator pneumatic architecture 400, such as the models used in the methods described herein. In one example, an oxygen model can be developed as follows: The oxygen concentration of the gas entering the patient's lungs through the inhalation tubing 442 can be obtained via a numerical model, which calculates the oxygen concentration based on the pressure signal obtained from pressure sensor 424, blower flow rate 420, the oxygen gas purity levels distributed at 406 and 408, the gas inlet valve position before 418, and environmental information regarding the oxygen gas concentration. The numerical model can be further developed using gas flow dynamics knowledge, through mass balance, and initial adjustments to hyperparameters in the model to adjust the signal level based on the ventilator design.

[0051] Turn now Figure 5 A schematic diagram 500 illustrates a cloud-connected edge computing frame operating as part of a monitoring system for a mechanical ventilator. In this example, the reliability of the mechanical ventilator's sensors and actuators can be classified based on aggregated sensor bias data from multiple group models. These multiple group models can be considered as sensor bias models based on group data (e.g., group-level data), where each group model can distinguish between bias accumulations caused by environmental conditions (e.g., humidity variations) or utilization rates (e.g., misuse due to lack of experience). In some examples, the group-level data can distinguish between bias accumulations caused by environmental conditions or utilization rates to classify the reliability of one or more ventilator components over time. In some examples, the group-data-based sensor bias models can classify bias corrections according to environmental conditions or utilization rates and transmit operating status and fault thresholds back to the ventilator. Ventilator performance and ventilator lifespan can be tracked over time based on the group-level data, where operation and corrections are performed as needed, e.g., as... Figure 2 As described in state machine 200. In this way, group-level data can determine the health status of the ventilator, thereby notifying patient care and ventilator maintenance as needed.

[0052] The predictor-corrector output 502 (such as sensor bias estimation) can be input to the ventilator system (such as... Figure 1 The control component 504 of the ventilation system 100. In one example, the control component 504 may be... Figure 1 The edge computing box 101. In another example, the control component 504 may be... Figure 1 The control system 118. Control component 504 can be internally deployed and connected to cloud networks (such as...). Figure 1The control unit 504 can communicate bidirectionally via the cloud network 108. The control unit 504 can transmit low-resolution data (e.g., data about a single ventilator) via the cloud network 108. In one example, the low-resolution data may include data from one or more sensors of the ventilator, such as sensor bias estimates 506. Additionally or alternatively, the low-resolution data may include control or diagnostic information related to the ventilator, such as utilization 508.

[0053] In one example, a group-based sensor bias model 510 can track the reliability of sensor bias calculations for individual ventilators in response to changes in environmental conditions or utilization (e.g., humidity, historical log data). The group-based sensor bias model 510 can differentiate the cumulative bias caused by environmental conditions or utilization to classify the reliability of sensor components at various stages of their lifespan. For each ventilator, a humidity-based correction 512 and / or a utilization-based correction 514 can be determined based on the group-based sensor bias model 510. The output of the group-based sensor bias model 510 (e.g., humidity and / or utilization corrections) can be visualized in a graph (such as a group model graph 516). The group model graph 516 shows the status of individual ventilators on a three-dimensional plot of utilization (x-axis), environmental conditions (y-axis), and humidity and / or utilization-based corrections (x). The group model graph 516 shows hollow circles indicating a group of ventilators in the group that are close to failure based on environmental conditions and utilization. Hollow circles indicate that the ventilators in a group will soon transition to fault-tolerant operating mode. Hollow circles represent ventilators and show moderate levels of sensor and / or actuator bias and correction due to changes in utilization and environmental conditions. Group model graph 516 shows black circles indicating ventilator groups progressing towards failure. Black circles represent ventilators and show low to moderate levels of sensor and / or actuator bias and correction due to changes in utilization and environmental conditions. Hollow squares represent new ventilator groups with low utilization levels, minimal environmental condition fluctuations, minimal or no bias correction.

[0054] In one example, a group-data-based sensor bias model 510 can monitor and adjust the operation of a cloud-connected ventilator in response to changing environmental variations and utilization levels. The ventilator's sensor bias calculations can be visualized in the context of other ventilators based on the group-data-based sensor bias model, providing information about the reliability of the exemplary ventilator. The group-data-based sensor bias model separates bias correction from environmental conditions and utilization. A group-data-based sensor bias model 516 can determine the ventilator's operating status and / or fault thresholds, which can be transmitted back to the control unit 504 and used to adjust the operation of the exemplary ventilator. In this way, ventilator performance and ventilator lifespan can be tracked over time based on group-level data, thereby simplifying operating protocols and / or maintenance accordingly.

[0055] Figure 6 A schematic diagram 600 of an exemplary condition-based maintenance for tracking reliability forecasting of a ventilator group is shown. In some examples, the methods disclosed herein can be performed on the existing computational footprint of the ventilators (including those without cloud networking capabilities) and therefore involve minimal additional investment. In other examples, existing computing and connectivity capabilities can be upgraded and / or modified. In some examples, establishing bidirectional cloud connectivity for the ventilator group can enhance ventilator reliability forecasting, thereby enabling a streamlined supply chain for ventilator component replacement, significantly offsetting upfront investment and reducing downstream costs associated with group management.

[0056] The examples shown in schematic diagram 600 include Figure 1 Components of the ventilation system 100 are shown. Figure 1 Ventilator 115. Ventilator 115 communicates electronically with controller 114. Schematic diagram 600 includes a ventilator 115. Figure 4 The oxygen sensor 426 and the first pressure sensor 424 are used in the ventilator 115. The sensed ventilator parameters can be compared with, for example, using... Figure 1The modeled ventilator parameters formed by ventilator model 102 are compared, with reliability predictions performed internally using edge computing box 101. In one example, ventilator 115, controller 114, and edge computing box 101 represent a single ventilator system in a cloud-networked ventilator system group within cloud network 108. One or more reliability predictions for the ventilator system can be transmitted to cloud network 108 and aggregated with one or more reliability predictions transmitted from multiple ventilators. In one example, an oxygen sensor bias correction 603a can be determined for oxygen sensor 426 and transmitted to controller 114 to adjust the operation of ventilator 115. In another example, a blower flow bias correction 603b can be determined for a first pressure sensor 424 and transmitted to controller 114 to adjust the operation of ventilator 115. Sensor bias correction data (e.g., oxygen or pressure) for the ventilator system can be aggregated with data representing other ventilator systems in the cloud-networked group. Aggregated reliability predictions can be tracked over time and the modeled ventilator can be updated from them. In one example, the cloud network may include multiple ventilator systems distributed within a single public care center (e.g., a single hospital location). In another example, the cloud network may include ventilator systems distributed over a larger geographic area comprising multiple public care center locations, as well as the ventilator systems therein.

[0057] exist Figure 6 In this context, service / maintenance logs 602a and 602b can provide cloud network 108 with historical ventilator usage information about the ventilator system. Service / maintenance logs 602a and 602b may be accompanied by and / or support group data 604a and 604b, which represent sensor data and actuator data, respectively. Within the cloud network, group data 604a and 604b, combined with group service and / or maintenance logs 602a and 602b, can provide an indication of the remaining useful life (RUL) of sensors and / or actuators. In one example, an oxygen sensor RUL indicator 606a may be supported by a combination of group data 604a and service and / or maintenance log 602a. In another example, a blower RUL indicator 606b may be supported by a combination of group data 604b and service and / or maintenance log 602b.

[0058] For example, service and maintenance plans can be developed based on the RUL indications of a group of ventilators in the network. Maintenance can be scheduled for a group of ventilators based on a first RUL indication threshold. Replacement of a group of ventilators can be guided based on a second, shorter RUL indication threshold. In this way, cloud networking of group-level data provides an added value to enhance ventilator failure prediction, thereby enabling a smooth supply chain process for ventilator maintenance and replacement.

[0059] In this way, mechanical ventilation systems can use semi-empirical, physics-based models to predict the reliability of the ventilator over time. The difference between the modeled parameters and sensor signals can be interpreted as a confidence metric indicating changes in the reliability of the sensor signals. By generating reliability predictions for the ventilator's sensors, actuators, and other components, and comparing these predictions with those of other ventilators, failure thresholds can be determined and ventilator operation adjusted accordingly. Through bidirectional communication with a cloud-connected ventilator group, the model can adapt to real-time environmental conditions and usage data to improve the reliability of individual ventilators and reduce reliance on redundant sensors. The technical benefits of fault-tolerant mechanical ventilation systems include more efficient ventilator maintenance, easier service workflows to ensure serviceability, reduced ventilator downtime, and a lower likelihood of patient discomfort.

[0060] This disclosure also provides support for a method comprising: predicting the reliability of a ventilator during ventilator operation based on time-tracked differences between modeled ventilator parameters and sensed ventilator parameters, and ventilator utilization. In a first example of the method, the method further comprises: displaying the predicted reliability to an operator of the ventilator, wherein the predicted reliability is based on predicted sensor bias and sensor failure. In a second example of the method (optionally including the first example), the method further comprises: transmitting the predicted reliability to a network that aggregates predicted reliability transmitted from multiple ventilators. In a third example of the method (optionally including one or both of the first and second examples), the method further comprises: generating an action for the operator and displaying the generated action to the operator. In a fourth example of the method (optionally including one or more of the first to third examples), the predicted reliability of the ventilator follows a state machine that includes states that allow restricted ventilator operation to continue even if a fault is detected, by including corrections to enable fault tolerance, and includes a state where ventilator operation is deactivated. In a fifth example of the method (optionally including one or more or each of the first to fourth examples), the modeled ventilator parameters are further based on a predictor-corrector architecture, and the severity of the discrepancy determines the state of the state machine. In a sixth example of the method (optionally including one or more or each of the first to fifth examples), the modeled ventilator parameters are based on a digital model embedded in the ventilator. In a seventh example of the method (optionally including one or more or each of the first to sixth examples), the method further includes: the ventilator communicating with a monitoring system via a bidirectional cloud connection, the monitoring system tracking ventilator performance and ventilator lifespan based on group-level data of a group of ventilators including the ventilator, the group-level data adjusting the ventilator for environmental conditions. In an eighth example of the method (optionally including one or more or each of the first to seventh examples), the method further includes: displaying a model output to an operator, the model output including information related to the reliability of the ventilator, and corresponding corrections for offsetting the reliability. In a ninth example of the method (optionally including one or more or each of the first to eighth examples), the ventilator group includes ventilators in a public care facility. In a tenth example of the method (optionally including one or more or each of the first to ninth examples), the method further includes: generating condition-based maintenance for the ventilator group based on the predicted reliability for each ventilator in the group tracked over time.

[0061] This disclosure also provides support for a method comprising: predicting the reliability of a ventilator during operation based on the difference between modeled ventilator parameters and sensed ventilator parameters, the modeled ventilator parameters being based on a digital model of the ventilator and sensor readings of the ventilator; tracking the ventilator in a state machine, the state of which is determined based on the predicted reliability, the state machine including a normal operating state, a degraded operating state in which the ventilator continues to operate under a warning indicating degradation, and a shutdown state in which the ventilator cannot continue to operate; and sending the output of the digital model to an edge processor, the edge processor communicating with a supervisory control system monitoring the ventilator. In a first example of the method, the modeled ventilator parameters are further based on a predictor-corrector architecture, and wherein the severity of the difference determines the state of the ventilator. In a second example of the method (optionally including the first example), the degraded operating state includes an actuator deviation state and an oxygen estimation failure state, and wherein the shutdown state includes a gas leak state.

[0062] This disclosure also provides support for a system comprising: a mechanical ventilator; a ventilator model running in real time on a processor; a predictor-corrector running in real time on the processor, the predictor-corrector receiving sensed outputs from the mechanical ventilator and outputs from the ventilator model; a GUI generating a display based on the output of the predictor-corrector; an edge processor; and a ventilator supervisory control system running in real time on the system, the ventilator supervisory control system receiving the output of the predictor-corrector, the ventilator supervisory control system communicating with the mechanical ventilator and the edge processor, the edge processor further communicating with the ventilator model and a cloud network. In a first example of the system, the mechanical ventilator includes a state machine controlling the operation of the mechanical ventilator, the state machine including: a normal operating state, a pressure deviation state, an oxygen estimation state, an oxygen standby state, and a gas leak state. In a second example of the system (optionally including the first example), the transition from the normal operating state to the pressure deviation state is determined based on the system's detection of pressure deviations. In a third example of the system (optionally including one or both of the first and second examples), the transition from the normal operating state to the oxygen estimation state is determined based on the system's detection of oxygen deviation. In a fourth example of the system (optionally including one or more of the first to third examples), the transition from the oxygen estimation state to the oxygen standby state is determined based on the detection of oxygen deviation exceeding a first threshold. In a fifth example of the system (optionally including one or more of the first to fourth examples), the transition from the normal operating state to the gas leak state is determined based on the system's detection of gas leak.

[0063] As used herein, the term "module" can include hardware and / or software systems that operate to perform one or more functions. For example, a module can include a computer processor, controller, or other logic-based means that performs operations based on instructions stored on a tangible, non-transitory computer-readable storage medium, such as computer memory. Alternatively, a module can include a hardwired means that performs operations based on the hardwired logic of the means. The various modules illustrated in the figures may represent hardware that operates based on software or hardwired instructions, software that instructs the hardware to perform operations, or a combination thereof.

[0064] A “module” may include or represent hardware and associated instructions (e.g., software stored on a tangible, non-transitory computer-readable storage medium, such as a computer hard disk drive, ROM, RAM, etc.) that perform one or more of the operations described herein. The hardware may include electronic circuitry comprising and / or connected to one or more logic-based devices, such as microprocessors, processors, controllers, etc. These devices may be readily available devices appropriately programmed or instructed to perform the operations described herein according to the instructions described above. Additionally or alternatively, one or more of these devices may be hardwired to logic circuitry to perform these operations.

[0065] As used herein, elements or steps listed in the singular and beginning with the word "a" or "an" should be understood to not exclude a plurality of said elements or steps unless such exclusion is explicitly stated. Furthermore, references to "one embodiment" of the invention are not intended to be construed as excluding the existence of additional embodiments that also include the referenced features. Moreover, unless explicitly stated to the contrary, embodiments that "comprise," "include," or "have" elements or multiple elements having a particular characteristic may include additional such elements that do not have that characteristic. The terms "comprise" and "in..." are used as concise linguistic equivalents to the corresponding terms "comprising" and "wherein". Furthermore, the terms "first," "second," and "third," etc., are used merely as notations and are not intended to impose numerical requirements or a particular order of position on their objects.

[0066] This written description uses examples to disclose the invention, including the best mode, and also enables those skilled in the art to practice the invention, including making and using any apparatus or system and performing any included methods. The scope of patentability of the invention is defined by the claims and may include other examples that would occur to those skilled in the art. Such other examples are intended to fall within the scope of the claims if they have structural elements that are not indistinguishable from the literal language of the claims, or if they include equivalent structural elements that have minor differences from the literal language of the claims.

Claims

1. A ventilator system, the ventilator system comprising a controller, the controller being configured to: The reliability of the ventilator during operation is predicted based on the time-tracked differences between modeled ventilator parameters and sensed ventilator parameters, as well as ventilator utilization. The modeled ventilator parameters are based on a digital model embedded in the ventilator. The output of the digital model is sent to an edge processor, which communicates with a supervisory control system monitoring the ventilator and a cloud network, wherein the cloud network includes multiple group models for measuring conditions of a group of ventilators including the ventilator, and wherein each of the multiple group models is configured to distinguish and determine a classification of deviation accumulation based on proximity to a triggering fault, the deviation accumulation being caused by environmental conditions or the utilization rate of the ventilator, wherein the multiple group models distinguish the deviation accumulation by classifying deviation corrections with environmental conditions and utilization rates, and adjust the operation of the ventilator via the cloud network in response to the deviation corrections; wherein The edge processor adjusts the operation of the ventilator based on communication with the cloud network, which provides data on a group of ventilators including the ventilator.

2. The ventilator system of claim 1, wherein the controller is further configured to display the predicted reliability to the operator of the ventilator, wherein the predicted reliability is based on predicted sensor bias and sensor failure.

3. The ventilator system of claim 1, wherein the controller is further configured to transmit the predicted reliability to a network, the network aggregating the predicted reliability transmitted from multiple ventilators.

4. The ventilator system according to claim 2, wherein the controller is further configured to generate actions for the operator and display the generated actions to the operator.

5. The ventilator system of claim 1, wherein the predicted reliability of the ventilator follows a state machine, the state machine including states that allow restricted ventilator operation to continue even if a fault is detected by including corrections to enable fault tolerance, and including states that ventilator operation is disabled.

6. The ventilator system of claim 5, wherein the modeled ventilator parameters are further based on a predictor-corrector architecture, and wherein the severity of the discrepancy determines the state of the state machine.

7. The ventilator system of claim 1, wherein the ventilator communicates with a monitoring system via a bidirectional cloud connection, the monitoring system tracking ventilator performance and ventilator lifespan based on group-level data of a group of ventilators including the ventilator, the group-level data adjusting the ventilator according to the environmental conditions of the ventilator.

8. The ventilator system of claim 1, wherein the controller is further configured to display a model output to an operator, the model output including information related to the reliability of the ventilator, and corresponding corrections for offsetting the reliability.

9. The ventilator system of claim 7, wherein the ventilator group comprises ventilators in a public care facility.

10. The ventilator system of claim 9, wherein the controller is further configured to generate condition-based maintenance for the ventilator group based on the reliability predictions for each ventilator in the group over time.

11. A ventilator system, the ventilator system comprising a controller, the controller being configured to: The reliability of the ventilator during operation is predicted based on the difference between modeled ventilator parameters and sensed ventilator parameters, wherein the modeled ventilator parameters are based on a digital model of the ventilator and sensor readings of the ventilator; and The ventilator is tracked in a state machine, the state of which is determined based on predicted reliability. The state machine includes a normal operating state, a degraded operating state where the ventilator continues to operate despite a warning indicating degradation, and a shutdown state where the ventilator can no longer operate. The output of the digital model is sent to an edge processor, which communicates with a supervisory control system monitoring the ventilator and a cloud network, wherein the cloud network includes multiple group models for measuring conditions of a group of ventilators including the ventilator, and wherein each of the multiple group models is configured to distinguish and determine a classification of deviation accumulation based on proximity to a triggering fault, the deviation accumulation being caused by environmental conditions or the utilization rate of the ventilator, wherein the multiple group models distinguish the deviation accumulation by classifying deviation corrections with environmental conditions and utilization rates, and adjust the operation of the ventilator via the cloud network in response to the deviation corrections; wherein The edge processor adjusts the operation of the ventilator based on communication with the cloud network, which provides data on a group of ventilators including the ventilator.

12. The ventilator system of claim 11, wherein the modeled ventilator parameters are further based on a predictor-corrector architecture, and wherein the severity of the discrepancy determines the state of the ventilator.

13. The ventilator system of claim 11, wherein the degraded operating state includes an actuator deviation state and an oxygen estimation failure state, and wherein the shutdown state includes a gas leakage state.

14. A ventilator system, the ventilator system comprising: Ventilation unit group, including mechanical ventilators; The cloud network includes multiple group models for measuring conditions including the ventilator group, wherein each of the multiple group models is configured to distinguish and determine a classification of deviation accumulation based on proximity to a triggering fault, the deviation accumulation being caused by environmental conditions or the utilization rate of the ventilator, wherein the multiple group models distinguish the deviation accumulation by classifying deviation correction with environmental conditions and utilization rate, and adjust the operation of the ventilator via the cloud network in response to the deviation correction; The mechanical ventilator model is run in real time on the processor; A predictor-corrector, which runs in real time on the processor, receives the sensed output of the mechanical ventilator and the output of the ventilator model; A GUI is generated and displayed based on the output of the predictor-corrector. Edge processor; as well as A ventilator monitoring and control system, which operates in real time on the system, receives the output of the predictor-corrector, communicates with the mechanical ventilator and the edge processor, and the edge processor further communicates with the ventilator model and the cloud network. The edge processor adjusts the operation of the mechanical ventilator based on communication with the cloud network, which provides data for the ventilator group.

15. The ventilator system of claim 14, wherein the mechanical ventilator includes a state machine for controlling the operation of the mechanical ventilator, the state machine comprising: Normal operating status; Pressure deviation status; Oxygen estimated state; Oxygen is in standby status; And gas leak status.

16. The ventilator system of claim 15, wherein the transition from the normal operating state to the pressure deviation state is determined based on the system's detection of the pressure deviation.

17. The ventilator system of claim 15, wherein the transition from the normal operating state to the oxygen estimation state is determined based on the system's detection of oxygen deviation.

18. The ventilator system of claim 15, wherein the transition from the oxygen estimation state to the oxygen standby state is determined based on the detection of an oxygen deviation beyond a first threshold.

19. The ventilator system of claim 15, wherein the transition from the normal operating state to the gas leakage state is determined based on the system's detection of gas leakage.

Citation Information

Patent Citations

  • Pressure sensor evaluation for respiratory apparatus

    CN104540539A

  • Computer aided mechanical ventilation systems and methods

    US20170232214A1

  • Systems and methods for predicting alertness

    WO2021171266A1