Medical equipment fault detection method based on multiple indicators

Abstract

The present invention relates to the field of medical device fault detection methods, and specifically to a multi-index-based medical device fault detection method, comprising: establishing an overall output observation signal model of the medical device using the medical device as an overall system; first, based on the characteristic intervals of each state determined by a medical device, measuring multiple medical devices of the same type under the same fault, calculating the recognition rate of each parameter under the fault and matching the weight value; calculating the total fault recognition rate for each fault state based on the recognition rate and weight value of the single fault indicator; adjusting the total fault recognition rate by adjusting the weight value or the fault characteristic interval until the set condition is met; and when the set condition is met, establishing a fault rule library corresponding to multiple preset fault indicators of the medical device. The present invention can accurately identify specific fault information of the medical device when the fault characteristic interval overlaps with the normal interval.

Description

Technical Field

[0001] The present invention relates to the field of fault detection methods for medical equipment, and in particular to a medical equipment fault detection method based on multiple indicators. Background Art

[0002] Medical devices continuously monitor a patient's physiological data in real time, allowing medical staff to quickly and intuitively view the corresponding vital signs and determine the patient's physical condition. Medical devices can malfunction in a variety of ways. Failure to detect medical device malfunctions during use can lead to misdiagnosis of the patient's condition and even delay monitoring.

[0003] Current fault detection for medical devices relies on measuring the voltage or current of a specific component during operation. This determines whether the device is faulty based on the voltage or current. This requires installing additional detection sensors at the corresponding locations on the medical device, which requires disassembling the device. If signal detection is performed on the medical device as a whole, and then fault diagnosis is made based on the detected signal, medical device failures typically occur in actuators or sensors. This current method, which relies on voltage or current to determine faults, only indicates that the medical device is faulty, but cannot distinguish which component of the device is faulty and which type of fault is occurring. Summary of the Invention

[0004] The present invention aims to provide a medical equipment fault detection method based on multiple indicators to solve the problem that existing methods cannot distinguish the specific fault types of medical equipment.

[0005] The multi-index-based medical equipment fault detection method in this solution includes the following steps:

[0006] Step 1: Taking the medical device as an overall system, establish an overall output observation signal model of the medical device, and collect normal ranges of various preset fault indicators when the medical device is working normally;

[0007] Step 2: Based on the observed signal model from step 1, measure the fault characteristic interval of a faulty medical device for each preset fault indicator. Then, introduce several new medical devices, measure their fault data under the same fault indicator, and calculate the single fault indicator recognition rate of the fault indicator of the new medical devices within the fault characteristic interval.

[0008] Step 3: Match the weight value of each preset fault indicator from the preset weight rule table according to the single fault indicator recognition rate, and calculate the total fault recognition rate of each fault state under multiple preset fault indicators based on the single fault indicator recognition rate and weight value;

[0009] Step 4: Compare the total fault recognition rate of each fault state with the set value to obtain a comparison result, and adjust the weight value or fault feature interval according to the comparison result until the set condition is met;

[0010] Step 5: When the set conditions are met, a fault rule library corresponding to multiple preset fault indicators of the medical equipment is built;

[0011] Step 6: Continuously test the existing faults, repeating steps 2, 3, and 4 until all the faults are tested and the total fault recognition rate is greater than the set value, completing an update.

[0012] Step 7: When the actual medical device has a fault, the value of the preset fault indicator of the faulty medical device is measured, and the value of the preset fault indicator is matched with the fault rule library, and the fault type of the medical device is determined according to the matching result.

[0013] The beneficial effects of this program are:

[0014] Taking the medical device as an overall system, normal ranges are set for each preset fault indicator collected by the medical device. The individual fault indicator recognition rates of the subsequently collected preset fault indicators under a single fault state are calculated. Based on the individual fault indicator recognition rates, the weight value of each preset fault indicator in identifying a fault state is matched. Based on the individual fault indicator recognition rates and weight values, the total fault recognition rate for each fault state under multiple preset fault indicators is calculated. Based on the performance of the total fault recognition rate, the weight value or fault feature interval is adjusted until the set conditions are met to establish a rule base. This allows the fault rule base to be used directly for fault matching during the next fault identification. The specific fault information of the medical device can be uniquely and accurately identified based on the matched fault feature area.

[0015] Furthermore, in step 1, the preset fault indicators include current, voltage and power, and in step 2, the real-time collection of the preset fault indicators is performed under the preset state of the medical device, and the preset state includes the power-on state and the power-off state.

[0016] The beneficial effect is that by collecting preset fault indicators of the medical device in the on and off states, the fault signal of the medical device can be represented to the greatest extent and accurately, which facilitates the analysis of the fault.

[0017] Furthermore, in step 1, a signal is input to the overall control system and superimposed on the entire output signal. The parameters and component failures generated by the overall system are recorded as f(t), and the observation signal output by the overall system is recorded as:

[0018] y(t)=Ax(t)+Bu(t)+R1f(t)+Cm(t);

[0019] Among them, x(t) is the state vector, u(t) is the control vector, and x(t)∈R n ,u(t)∈R p , both are the system operation vector and controller output signal under normal conditions; y(t) is the observed output signal vector; m(t) is the signal output added by the external acquisition sensor itself; f(t) is the abnormal operation state vector of a certain part of the system, f(t)∈R g , each element f i (t) (i = 1, 2, 3... g) corresponds to a specific fault form and is also the unknown time function that needs to be solved in our fault research; ABC is a constant matrix of corresponding dimension; R1 is the fault coefficient matrix, the number of detectable states determines the dimension of this matrix;

[0020] Using the residual measurement method, the discretely collected values ​​are substituted into the observation signal model, and the following fault characteristics of the preset fault indicators are obtained through deformation, where β represents current, χ represents power, and δ represents voltage:

[0021]

[0022] The beneficial effect is: the medical device is regarded as a whole system, and its fault analysis is carried out based on the output signal of the whole system. There is no need to set up a sensor for each component inside the medical device for signal acquisition, which reduces the analysis error caused by the external sensor devices during the analysis process, and the operation is simpler and more convenient.

[0023] Furthermore, in step 1, the fault feature after difference is multiplied by the amplification value and then used for comparison.

[0024] The beneficial effect is that since the characteristic value obtained by measurement is very small, at the milliampere level, comparison is facilitated by setting the amplification value.

[0025] Furthermore, in step 4, the comparison results include a total fault recognition rate greater than or equal to a set value and a total fault recognition rate less than a set value. When the comparison result is that the total fault recognition rate is less than the set value, each single fault indicator recognition rate is compared with the threshold to obtain a single result. The weight value or fault feature interval is adjusted according to the single result, and the total fault recognition rate is calculated again after the adjustment process is completed until the set conditions are met.

[0026] The beneficial effect is: when the total fault recognition rate is less than the set value, the recognition rate of the single fault indicator is compared, the weight value or the fault characteristic interval is adjusted according to the single result, and the total fault recognition rate is calculated again to finally meet the set conditions, which can distinguish various faults and improve the accuracy of fault judgment.

[0027] Further, in step 4, the single result includes that the recognition rate of at least one single fault indicator among the fault indicators is greater than or equal to a threshold value. When the single fault indicator recognition rate of a fault indicator in a fault state is greater than or equal to the threshold value, the weight value of the preset fault indicator is adjusted according to a first preset step size. When the single fault indicator recognition rates of the remaining fault indicators in a fault state are less than the threshold value, the weight values ​​of the preset fault indicators are adjusted according to a second preset step size. The total fault recognition rate is repeatedly calculated until the total fault recognition rate is greater than the set value.

[0028] The beneficial effect is: by comparing the single fault indicator recognition rate of each preset fault indicator with the threshold, and adjusting the weight value of the corresponding preset fault indicator according to the size of the recognition rate of some single fault indicators in each preset fault indicator, the overall fault recognition rate can be guaranteed to have good recognition ability.

[0029] Furthermore, the single result also includes that all single fault indicator recognition rates of all fault indicators are less than a threshold value. When the single fault indicator recognition rates of multiple fault indicators are less than the threshold value, first determine whether the current fault feature interval and the normal interval overlap. If the current fault feature interval and the normal interval do not overlap, the fault feature interval is amplified; if the fault feature interval and the normal interval overlap, the fault feature interval is shifted to a blank interval preset with other fault feature intervals, and the value of the fault feature interval of the preset fault indicator is compensated. The compensation includes positive compensation for increasing the shift upward and negative compensation for decreasing the shift downward. Then, the total fault recognition rate after compensation is calculated until the total fault recognition rate is greater than the set value and the set conditions are met, such as the fault feature interval of each fault state does not overlap with the normal interval.

[0030] The beneficial effect is that when the single fault indicator recognition rate of each fault indicator does not meet the requirements, different intervals are compensated according to whether the fault feature interval overlaps with the normal interval, thereby avoiding the overlap of the fault feature interval and the normal interval and improving the accuracy of fault identification.

[0031] Furthermore, in step 4, when the single fault indicator recognition rate of at least one fault indicator in any fault state is greater than a threshold and the fault feature interval does not overlap with the normal interval, the weight of the overlapping fault feature interval is reset to 0.

[0032] The beneficial effect is that when the fault characteristic interval of the fault indicator under at least one fault state has a high recognition rate, the weights of other fault characteristic intervals are set to 0, thereby reducing the amount of calculation and improving the calculation speed.

[0033] Furthermore, in step 3, the calculation formula of the total fault recognition rate is:

[0034]

[0035] Among them, p1, p2, and p3 represent the single fault indicator recognition rates of current, voltage, and power in each fault state, respectively, and θ1, θ2, and θ3 represent the weight values ​​of current, voltage, and power obtained by matching the single fault indicator recognition rates.

[0036] The beneficial effect is that, by calculating the total fault recognition rate, the corresponding fault states when the fault feature intervals overlap can be distinguished, so as to accurately locate the corresponding fault of the medical device.

[0037] Furthermore, in step 7, when the fault type of the medical device is not determined, it is judged to be a new fault, the normal range of the new fault medical device is re-determined, and steps 2 to 5 are repeated to enter the fault rules of the new fault.

[0038] The beneficial effect is that when the fault feature intervals in the fault rule base cannot identify the fault, the content in the fault rule base can be updated by re-establishing the fault judgment rules for the corresponding fault, thereby improving the recognition capability of the fault rule base. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a flowchart of a first embodiment of a medical device fault detection method based on multiple indicators of the present invention;

[0040] Figure 2 This is a diagram of a model applicable to medical equipment in Example 1 of the medical equipment fault detection method based on multiple indicators of the present invention. DETAILED DESCRIPTION

[0041] The following is further explained in detail through specific implementation methods.

[0042] Example 1

[0043] Medical equipment fault detection methods based on multiple indicators, such as Figure 1 As shown, the following steps are included:

[0044] Step 1: Take the medical device as an overall system and input a signal to the overall control of the system, which is superimposed on the overall output signal. The parameters and component failures generated by the overall system are recorded as f(t). The observation signal model of the overall output of the system is recorded as:

[0045] y(t)=Ax(t)+Bu(t)+R1f(t)+Cm(t);

[0046] Among them, x(t) is the state vector, u(t) is the control vector, and x(t)∈R n ,u(t)∈R p, both are the system operation vector and controller output signal under normal conditions; y(t) is the observed output signal vector; m(t) is the signal output added by the external acquisition sensor itself; f(t) is the abnormal operation state vector of a certain part of the system, f(t)∈R g , each element f i (t)(i=1,2,3...g) corresponds to a specific fault form and is also the unknown time function that needs to be solved in the fault study of this embodiment; ABC is a constant matrix of corresponding dimension; R1 is the fault coefficient matrix, and the number of detectable states determines the dimension of the matrix.

[0047] Using the residual measurement method, the discretely collected values ​​are substituted into the observed signal model, and the following fault characteristics of the preset fault indicators are obtained. β represents current, χ represents power, and δ represents voltage. The corresponding relationship of the fault characteristics is shown in Table 1. The fault characteristics after difference are multiplied by the amplification value and then used for comparison. The amplification value is set according to actual needs. For example, the amplification value is set to 1000:

[0048]

[0049] Table 1 Fault characteristics correspondence

[0050] Running status Fault phenomenon Status indicator Status Characteristics Running normally \ A <![CDATA[β0、χ0、δ0]]> ECG malfunction The ECG waveform is irregular or no waveform is detected during measurement <![CDATA[B1]]> <![CDATA[β1、χ1、δ1]]> Motherboard failure The main control system fails, cannot be turned on, cannot be controlled, or malfunctions <![CDATA[B2]]> <![CDATA[β2、χ2、δ2]]> Screen failure Black screen when powered on, no operation displayed <![CDATA[B3]]> <![CDATA[β3、χ3、δ3]]> Main cable failure Button failure, inability to turn on or off, and other operational controls <![CDATA[B4]]> <![CDATA[β4、χ4、δ4]]> Air pump failure Non-invasive blood pressure cannot be inflated <![CDATA[B5]]> <![CDATA[β5、χ5、δ5]]>

[0051] After establishing the observation signal model, the normal ranges of various preset fault indicators during normal operation of the medical device are collected. Assuming that the input fault f(t) = 0 indicates normal operation of the medical device, the characteristic range of the signal output when the medical device is normal is represented as a0±m0. The normal range is set by measuring the representative value range of each preset fault indicator during normal operation of the medical device, such as standby, shutdown, and working state 1. The HLW8032 metering chip module circuit can be used to collect electrical performance parameters. For the preset fault indicator of current, the measured representative values ​​range from 214mA to 238mA. Therefore, the normal range is set to 214-0.238. The preset fault indicators include current, voltage, and power.

[0052] Medical equipment is simulated as an overall system as the observation target. This method does not require the data collector to have rich professional knowledge and is easy to operate. It avoids adding too many sensors to monitor the failure of different components, and the superposition of multiple signals destroys the characteristics of the original output of the equipment. It can also highlight the changes in fault characteristics, making the characteristic changes more sensitive.

[0053] In step 2, based on the observed signal model from step 1, the fault characteristic interval of a faulty medical device is measured for each preset fault indicator. Several new medical devices are then introduced and their fault data is measured under the same fault indicator. The recognition rate of the single fault indicator for the fault data of the new medical devices within the fault characteristic interval is calculated. The recognition rate of the single fault indicator is the ratio of the number of fault data points within the fault characteristic interval to the total number of fault data points measured for a particular medical device fault. The fault indicator measurement is performed using the observed signal model from step 1 under preset states of the medical device, including both power-on and power-off states.

[0054] Taking medical device A as an example, we measure its output signal in a fault state. The output signal fault data is represented as a1±m1. The output signal data of another medical device B in the same fault state is then used to obtain a set A′1. The single fault indicator recognition rate of A′1 in the interval a1±m1 is calculated. The single fault indicator recognition rate is calculated for each of the multiple preset indicators under the same fault state, and the single fault indicator recognition rate is calculated for all indicators in each fault state.

[0055] Compared with the existing single-indicator judgment, the multi-indicator parameters used in this embodiment carry richer fault information. The higher the dimension of the characteristic parameters, the more accurate the judgment, and the more identifiable fault types.

[0056] Step 3: Based on the single fault indicator recognition rate, the weight value of each preset fault indicator is matched from the preset weight rule table, and the total fault recognition rate of each fault state under multiple preset fault indicators is calculated based on the single fault indicator recognition rate and weight value. Taking electrical characteristic parameters as an example, the three indicators of current, voltage and power are measured. The calculation formula for the total fault recognition rate of medical equipment under these three indicators is:

[0057]

[0058] Among them, p1, p2, and p3 represent the single fault indicator recognition rates of current, voltage, and power under each fault state, respectively. θ1, θ2, and θ3 represent the weight values ​​of current, voltage, and power obtained by matching their respective single fault indicator recognition rates. The weight rule table of weight values ​​is shown in Table 2. In Table 2, P represents the probability interval and θ represents the weight value.

[0059] Table 2 Weight rules table

[0060] p θ p θ p θ p θ p θ [90％，100％] 9 [80％，90％) 8 [70％，80％) 7 [60％，70％) 6 [50％，60％) 5 [40％，50％) 4 [30％，40％) 3 [20％，30％) 2 [10％，20％) 1 (0，10％) 0

[0061] By matching weights based on the recognition rate of a single fault indicator, the recognition rate can be flexibly adjusted to change. For a certain fault, one or more parameters may have a greater impact on the fault. In this case, the importance of the parameters can be defined through weights, thereby achieving a higher recognition rate and higher availability.

[0062] Step 4: Compare the total fault recognition rate of each fault state with the set value to obtain a comparison result. The set value is set according to actual needs. For example, the set value is set to 90%. The comparison results include the total fault recognition rate being greater than or equal to the set value and the total fault recognition rate being less than the set value. When the comparison result is that the total fault recognition rate is less than the set value, adjust the weight value or the fault feature interval, and calculate the total fault recognition rate again after the adjustment process is completed until the total fault recognition rate is greater than the set value and the fault feature interval of each fault state does not overlap with the normal interval.

[0063] When adjusting the weight value or fault feature interval, the recognition rate of each single fault indicator among multiple fault indicators is compared with the threshold to obtain a single result. The threshold is set according to actual requirements. For example, the threshold can be set to 90% or 80%. In this embodiment, the threshold takes 90% as an example, and the weight value or fault feature interval is adjusted according to the single result.

[0064] The single result includes that the recognition rate of at least one single fault indicator among the preset fault indicators is greater than or equal to the threshold. The process of adjusting the weight value or fault feature interval according to the single result is as follows:

[0065] The weight value of the preset fault indicator is adjusted according to the first preset step size and the second preset step size based on the single result. When the single fault indicator recognition rate of a fault indicator in a fault state is ≥90%, the first preset step size of the corresponding preset fault indicator is set to θ+1. When the single fault indicator recognition rate of the remaining fault indicators in a fault state is less than 90%, the second preset step size of the corresponding preset fault indicator is set to θ-1. The total fault recognition rate is repeatedly calculated until the total fault recognition rate is ≥90%.

[0066] The individual results also include that all individual fault indicator recognition rates for all fault indicators are less than the threshold;

[0067] When the recognition rates of individual fault indicators of multiple fault indicators are all less than 90%, that is, each fault indicator does not meet the preset recognition rate, first determine whether the current fault feature interval and the normal interval overlap. If the current fault feature interval and the normal interval do not overlap, the fault feature interval is enlarged; if the fault feature interval and the normal interval overlap, the fault feature interval is shifted to a preset blank interval that does not overlap with other fault feature intervals, and the value of the fault feature interval of the preset fault indicator is compensated. The compensation includes positive compensation for increasing the shift upward and negative compensation for decreasing the shift downward. After performing the above operations, the total fault recognition rate after compensation is calculated until the total fault recognition rate is greater than the set value and the fault feature interval of each fault state does not overlap with the normal interval.

[0068] Rapidly optimize weighting methods and weight optimization rules, through which the preset recognition rate can be achieved to ensure that faults can be accurately identified.

[0069] Step 5: When the total fault recognition rate is greater than the set value and the fault characteristic interval of each fault state does not overlap with the normal interval, a fault rule library corresponding to multiple preset fault indicators of the medical equipment is established. The fault rule library includes the fault characteristic intervals of multiple preset fault indicators in each fault state, and the characteristic intervals of each parameter under each fault at this time are recorded.

[0070] Build a fault rule library, associate fault events with characteristic rules and record them. After continuous updates in the subsequent process, the rule library will become richer and richer, and will become more and more perfect, eventually forming a usable advanced equipment fault rule library.

[0071] In step 6, the existing faults are continuously measured, and steps 2 and 3 are repeated (because there are existing rules, the judgment can be directly referenced). When the total fault recognition rate is lower than the preset recognition rate, the fault feature interval in step 4 is continuously amplified without compensation as much as possible until the current measurement is covered and the total fault recognition rate is greater than the set value. This completes an update.

[0072] By supplementing and updating the rule base, rolling updates can be achieved. The open fault rule base allows any new faults of the same model equipment to be entered at any time. If the same faults of the same model equipment conflict, they can be updated and entered at any time according to the patented method.

[0073] Step 7: When the medical device actually in use has a fault, the value of a preset fault indicator of the faulty medical device is measured, and the value of the preset fault indicator is matched with a fault rule library, and the fault type of the medical device is determined according to the matching result.

[0074] After completing the initial construction of the fault rule library, the rule library can be used to automatically identify on-site and remote equipment faults and provide repair suggestions with high accuracy, which is convenient for guiding new engineers entering the industry. For example, township health centers can carry out preliminary treatment for health institutions that do not have professional maintenance capabilities.

[0075] Taking a normal model A monitor as an example, the normal ranges of the three electrical properties of current, power, and voltage in the power-on and power-off states are shown in Table 3. When the model A monitor has an air pump failure, the measured air pump failure characteristic range is shown in Table 4.

[0076] Table 3 Normal ranges of the three electrical properties of current, power and voltage for model A monitors

[0077]

[0078] Table 4 Air pump fault characteristic intervals of three electrical properties: current, power, and voltage for model A monitor

[0079]

[0080] The method of this embodiment is compared with the fault detection method under a single fault feature, and a detection accuracy comparison table of the detection method of multiple fault features and the detection method of a single fault feature is obtained, as shown in Table 5.

[0081] Table 5 Comparison of the accuracy of multiple indicators and single indicator judgment

[0082]

[0083] As shown in Table 5, the method of this embodiment can accurately determine different fault types.

[0084] This embodiment collects data from the medical device as a whole system. If the collected data is simply used for fault diagnosis, it can only determine that the medical device is faulty, but cannot determine which component within the medical device has caused the fault. Therefore, based on the normal range of the medical device during normal operation, the system then calculates the individual fault indicator recognition rates for multiple preset fault indicators under a single fault state. Based on the individual fault indicator recognition rates, the system matches the weight value of each preset fault indicator in identifying a fault state. The total fault recognition rate for each fault state under multiple preset fault indicators is calculated based on the individual fault indicator recognition rates and weight values. The system then determines the total fault recognition rate and adjusts the weight value until the total fault recognition rate exceeds the set value and the fault feature area does not overlap with the normal range. A corresponding fault rule library is then established, so that the fault rule library can be directly used for fault matching during the next fault identification. This system can accurately identify specific fault information of the medical device when the fault feature area overlaps with the normal range.

[0085] Example 2

[0086] The multi-indicator-based medical device fault detection method differs from Example 1 in that, in step 4, when the single fault indicator recognition rate of at least one fault indicator under at least one fault state is greater than a threshold and the fault characteristic interval does not overlap with the normal interval, the weights of the fault characteristic intervals overlapping with other preset fault indicators are reset to 0. That is, the weights of the fault indicators corresponding to the overlapping normal intervals are reset to 0. For example, when the single fault indicator recognition rate of voltage is greater than a threshold, and the fault characteristic intervals of current and power overlap with the normal intervals, the weights corresponding to the fault characteristic intervals of current and power are set to 0. When the fault characteristic interval of a fault indicator under at least one fault state has a high recognition rate, the weights of the other fault characteristic intervals are set to 0, reducing the computational complexity and increasing computational speed for faster judgment.

[0087] Example 3

[0088] The multi-indicator-based medical device fault detection method differs from the first embodiment in that, in step 7, if the fault type of the medical device cannot be determined, it is determined to be a new fault, the normal range of the newly faulted medical device is re-determined, and steps 2 through 5 are repeated to enter the fault rule for the new fault. If none of the fault feature intervals in the fault rule library can identify the fault, the fault judgment rule for the corresponding fault is re-established, thereby updating the content of the fault rule library and improving its recognition capabilities.

[0089] The above is only an embodiment of the present invention, and the common knowledge such as the specific structure and characteristics of the scheme is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A multi-index-based medical equipment fault detection method, characterized in that: The following steps are involved: Step 1: Taking the medical device as an overall system, establish an overall output observation signal model of the medical device, and collect normal ranges of multiple preset fault indicators when the medical device is operating normally, wherein the preset fault indicators include current, voltage, and power; A signal is input to the overall control system and superimposed on the entire output signal. The parameters and component failures generated by the overall system are recorded as f ( t ), the observation signal output by the system as a whole is recorded as: ； in, is the state vector, are control vectors, respectively 、 , both are the system operation vector and controller output signal quantity under normal conditions of the equipment; is the observed output signal vector; It is the signal output added by the external acquisition sensor itself; is the abnormal operation state vector of a certain part of the system, , each element It corresponds to a specific fault form and is also the unknown time function that needs to be solved in our fault research; is a constant matrix of corresponding dimension; is the fault coefficient matrix, the number of detectable states determines the dimension of the matrix; Using the residual measurement method, the discretely collected values ​​are substituted into the observation signal model, and the fault characteristics of the following preset fault indicators are obtained by deformation: β represents the current, χ Indicates power, δ Indicates voltage: ； Step 2: Based on the output signal of the observation model in step 1, measure the fault characteristic interval of a faulty medical device for each preset fault indicator, then introduce several new medical devices, measure their fault data under the same fault indicator, and calculate the single fault indicator recognition rate of the fault indicator of the new medical devices within the fault characteristic interval. The real-time collection of the preset fault indicators is performed under preset states of the medical devices, which include power-on and power-off states. Step 3: Match the weight value of each preset fault indicator from the preset weight rule table according to the single fault indicator recognition rate, and calculate the total fault recognition rate of each fault state under multiple preset fault indicators based on the single fault indicator recognition rate and weight value; Step 4: Compare the total fault recognition rate of each fault state with the set value to obtain a comparison result, and adjust the weight value or fault feature interval according to the comparison result until the set condition is met; Step 5: When the set conditions are met, a fault rule library corresponding to multiple preset fault indicators of the medical equipment is built; Step 6: Continuously test the existing faults, repeating steps 2, 3, and 4 until all the faults are tested and the total fault recognition rate is greater than the set value, completing an update. Step 7: When the actual medical device has a fault, the value of the preset fault indicator of the faulty medical device is measured, and the value of the preset fault indicator is matched with the fault rule library, and the fault type of the medical device is determined according to the matching result.

2. The multi-index-based medical equipment fault detection method according to claim 1, characterized in that: In step 1, the fault feature after difference is multiplied by the amplification value and then used for comparison.

3. The multi-index-based medical equipment fault detection method according to claim 1, characterized in that: In step 4, the comparison results include the total fault recognition rate being greater than or equal to the set value and the total fault recognition rate being less than the set value. When the comparison result is that the total fault recognition rate is less than the set value, each single fault indicator recognition rate is compared with the threshold to obtain a single result, and the weight value or fault feature interval is adjusted according to the single result. After the adjustment process is completed, the total fault recognition rate is calculated again until the set conditions are met.

4. The multi-index-based medical equipment fault detection method according to claim 3, characterized in that: In step 4, the single result includes that the recognition rate of at least one single fault indicator among the fault indicators is greater than or equal to a threshold value. When the single fault indicator recognition rate of a fault indicator in a fault state is greater than or equal to the threshold value, the weight value of the preset fault indicator is adjusted according to a first preset step size. When the single fault indicator recognition rates of the remaining fault indicators in a fault state are less than the threshold value, the weight values ​​of the preset fault indicators are adjusted according to a second preset step size. The total fault recognition rate is repeatedly calculated until the total fault recognition rate is greater than the set value.

5. The multi-index-based medical equipment fault detection method according to claim 4, characterized in that: The single result also includes that all single fault indicator recognition rates of all fault indicators are less than the threshold. When the single fault indicator recognition rates of multiple fault indicators are less than the threshold, first determine whether the current fault feature interval and the normal interval overlap. If the current fault feature interval does not overlap with the normal interval, amplify the fault feature interval; If the fault characteristic interval overlaps with the normal interval, the fault characteristic interval will be shifted to the outside of the blank interval preset with other fault characteristic intervals, and the value of the fault characteristic interval of the preset fault indicator will be compensated. The compensation includes positive compensation for increasing the translation upward and negative compensation for decreasing the translation downward. The total fault recognition rate after compensation is then calculated until the total fault recognition rate is greater than the set value and the fault characteristic interval of each fault state does not overlap with the normal interval.

6. The multi-index-based medical equipment fault detection method according to claim 5, characterized in that: In step 4, when the single fault indicator recognition rate of at least one fault indicator in any fault state is greater than a threshold and the fault feature interval does not overlap with the normal interval, the weight of the overlapping fault feature interval is reset to 0.

7. The multi-index-based medical equipment fault detection method according to claim 1, characterized in that: The calculation formula of the total fault recognition rate is: ， in, Respectively represent the single fault indicator recognition rate of current, voltage and power under each fault state, They represent the weight values ​​of current, voltage, and power respectively obtained by matching the recognition rate of single fault indicators.

8. The multi-index-based medical equipment fault detection method according to claim 1, characterized in that: In step 6, when the fault type of the medical device is not determined based on the working status value, it is determined to be a new fault, the normal range is readjusted, and steps 2 to 5 are repeated.

Images