CT tube reliability prediction method, device, electronic equipment and storage medium

Abstract

The present invention relates to equipment reliability technology, and provides a CT tube reliability prediction method, device, electronic device and storage medium, the method comprising: obtaining a reference tube usage interval for predicting the reliability of a target CT tube, the minimum value and maximum value of the reference tube usage interval being respectively the minimum value and maximum value of the tube usage of a plurality of sample CT tubes, each sample CT tube and the target CT tube being of the same model; extracting a plurality of reference usages from the reference tube usage interval; inputting each reference usage into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each reference usage; and predicting the reliability of the target CT tube based on all reference failure probabilities. The present invention can improve the accuracy of CT tube reliability prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment reliability, and in particular to a CT tube reliability prediction method, device, electronic equipment and storage medium. Background Art

[0002] With the widespread application of IoT technology, medical equipment can achieve real-time data collection and remote monitoring through sensors and network connections. For key medical equipment such as CT tubes, the IoT platform can continuously collect their operating status and fault data, providing a large amount of time series data for reliability prediction. However, how to make full use of these complex data for accurate prediction is still a technical challenge.

[0003] Traditional statistical methods, such as the Weibull distribution model, although stable in specific applications, usually rely on assumptions about data distribution and are difficult to cope with dynamically changing failure data. The accuracy of reliability prediction often fails to meet expectations. Summary of the invention

[0004] The present invention aims to provide a CT tube reliability prediction method, device, electronic equipment and storage medium.

[0005] The embodiments of the present invention can be implemented as follows:

[0006] In a first aspect, the present invention provides a method for predicting the reliability of a CT tube, the method comprising:

[0007] Obtaining a reference tube usage interval for predicting the reliability of a target CT tube, wherein a minimum value and a maximum value of the reference tube usage interval are respectively a minimum value and a maximum value of tube usage of a plurality of sample CT tubes, and each of the sample CT tubes and the target CT tube are of the same model;

[0008] Extracting a plurality of reference dosages from the reference bulb dosage interval;

[0009] Each of the reference usage is input into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each of the reference usage, wherein the LSTM model is trained using the tube usage of the multiple sample CT tubes as sample data and the failure probability when the used tube usage of each of the sample CT tubes reaches the corresponding sample data as a label of each of the sample data;

[0010] The reliability of the target CT tube is predicted based on all the reference failure probabilities.

[0011] In an optional implementation manner, the reliability of the target CT tube is the remaining tube usage of the target CT tube, and the step of predicting the reliability of the target CT tube based on all the reference failure probabilities includes:

[0012] Obtaining a failure probability threshold and a current tube usage of the target CT tube;

[0013] Using the reference failure probability as the predicted failure probability and using the reference usage corresponding to the predicted failure probability as the predicted usage;

[0014] Determine a fault probability interval in which the fault probability threshold is located from all the predicted fault probabilities, wherein the minimum value of the fault probability interval is the maximum value of the predicted fault probability less than the fault probability threshold, and the maximum value of the fault probability interval is the minimum value of the predicted fault probability greater than the fault probability threshold;

[0015] If there is no target failure probability in the failure probability interval whose difference between the failure probability and the failure probability threshold satisfies the preset difference, determining the predicted usage interval according to the predicted usage corresponding to the minimum value and the maximum value of the failure probability interval;

[0016] Re-extracting a plurality of predicted usages from the predicted usage interval;

[0017] Input each newly extracted predicted usage into a pre-trained long short-term memory (LSTM) model to obtain a predicted failure probability corresponding to each newly extracted predicted usage, and return to the step of determining the failure probability interval in which the failure probability threshold is located from all the predicted failure probabilities until there is a target failure probability in the failure probability interval of the current iteration where the difference between the failure probability and the failure probability threshold satisfies the preset difference;

[0018] The difference between the predicted usage corresponding to the target failure probability and the current tube usage is calculated to obtain the remaining tube usage.

[0019] In an optional embodiment, the reliability of the target CT tube is a high failure usage of the target CT tube, the high failure usage represents the tube usage when the probability of failure of the target CT tube reaches a preset probability, and the step of predicting the reliability of the target CT tube based on all the reference failure probabilities includes:

[0020] Determine a failure rate function according to all the reference failure probabilities, wherein the failure rate function represents the relationship between the failure rate of the target CT tube and the tube usage;

[0021] The high failure usage is determined according to the tube usage corresponding to the peak value of the failure rate function.

[0022] In an optional implementation, the step of determining a failure rate function according to all the reference failure probabilities comprises:

[0023] generating a cumulative distribution function based on all of the reference failure probabilities;

[0024] Performing a derivative operation on the cumulative distribution function to obtain a probability density function;

[0025] The failure rate function is generated according to the cumulative distribution function and the probability density function.

[0026] In an optional implementation, the trained LSTM model is obtained as follows:

[0027] Obtaining the original tube usage of each of the sample CT tubes;

[0028] Determine the failure probability corresponding to each of the original tube usages;

[0029] The amount of each original bulb tube is standardized, and the amount of each original bulb tube after the standardized process is used as the amount of each bulb tube;

[0030] The usage of the bulb is taken as a sample data, and the failure probability corresponding to each sample data is taken as the label of each sample data. The pre-constructed LSTM model is trained using all the sample data, the labels and the preset loss function to obtain the trained LSTM model.

[0031] In an optional implementation, the step of determining the failure probability corresponding to each of the original tube usages includes:

[0032] Arrange all the original tube usages in ascending order to obtain a tube usage sequence;

[0033] According to the sequence number of each original tube usage in the tube usage sequence, the failure probability corresponding to each original tube usage is calculated.

[0034] In an optional embodiment, the step of standardizing the amount of each original bulb comprises:

[0035] Calculate the mean and standard deviation of all the original tube usages;

[0036] The original tube usage of each of the tubes is standardized according to the mean and the standard deviation to obtain the tube usage of each of the tubes.

[0037] In a second aspect, the present invention provides a CT tube reliability prediction device, the device comprising:

[0038] An acquisition module, used for acquiring a reference tube usage interval for predicting the reliability of a target CT tube, wherein the minimum value and the maximum value of the reference tube usage interval are respectively the minimum value and the maximum value of the tube usage of a plurality of sample CT tubes, and each of the sample CT tubes and the target CT tube are of the same model;

[0039] An extraction module, used for extracting a plurality of reference dosages from the reference bulb dosage interval;

[0040] A prediction module is used to input each of the reference usage into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each of the reference usage, wherein the LSTM model is trained using the tube usage of the multiple sample CT tubes as sample data and the failure probability when the used tube usage of each of the sample CT tubes reaches the corresponding sample data as a label of each of the sample data;

[0041] The prediction module is further used to predict the reliability of the target CT tube based on all the reference failure probabilities.

[0042] In a third aspect, the present invention provides an electronic device, including a processor and a memory, wherein the memory is used to store a program, and the processor is used to implement a CT tube reliability prediction method as described in any one of the aforementioned embodiments when executing the program.

[0043] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a CT tube reliability prediction method as described in any one of the aforementioned embodiments.

[0044] Compared with the prior art, the present invention has the following beneficial effects: the present invention determines a reference tube usage interval according to the minimum and maximum values ​​of the tube usage of multiple sample CT tubes, and uses a pre-trained LSTM model to obtain a reference failure probability corresponding to each reference usage extracted from the reference tube usage interval, and then predicts the reliability of the target CT tube with all reference failure probabilities. Since the LSTM model is fully utilized to effectively capture the long-term dependency and nonlinear characteristics in the failure data of the sample CT tubes of the same model as the target CT tube, the reliability of the target CT tube is predicted according to the reference failure probability predicted by the LSTM model, thereby improving the accuracy of the reliability prediction of the target CT tube. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 This is an example flowchart of the CT tube reliability prediction method provided in this embodiment.

[0047] Figure 2 This is an example diagram of the tube usage provided in this embodiment.

[0048] Figure 3 This is an example diagram of predicting the remaining tube usage in the cumulative distribution function provided in this embodiment.

[0049] Figure 4 This is an example diagram of the cumulative distribution function obtained by LSTM provided in this embodiment.

[0050] Figure 5 A comparison example diagram of the cumulative distribution function obtained by the LSTM model and Weibull distribution provided in this embodiment.

[0051] Figure 6 This is a block diagram of an example of a CT tube reliability prediction device provided in this embodiment.

[0052] Figure 7 An example block diagram of an electronic device provided in this embodiment.

[0053] Icons: 10 - electronic device; 11 - processor; 12 - memory; 13 - bus; 100 - CT tube reliability prediction device; 110 - acquisition module; 120 - extraction module; 130 - prediction module; 140 - training module. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0055] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0057] In the description of the present invention, it should be noted that if the terms "upper", "lower", "inside", "outside", etc. appear to indicate an orientation or position relationship, they are based on the orientation or position relationship shown in the accompanying drawings, or are the orientation or position relationship in which the product of the invention is usually placed when used. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0058] In addition, the terms “first”, “second”, etc., if used, are merely used to distinguish between the descriptions and should not be understood as indicating or implying relative importance.

[0059] It should be noted that, in the absence of conflict, the features in the embodiments of the present invention may be combined with each other.

[0060] Please refer to Figure 1 , Figure 1 The flowchart of the CT tube reliability prediction method provided in this embodiment includes the following steps:

[0061] Step S101, obtaining a reference tube usage interval for predicting the reliability of a target CT tube, wherein the minimum value and the maximum value of the reference tube usage interval are respectively the minimum value and the maximum value of the tube usage of a plurality of sample CT tubes, and each sample CT tube and the target CT tube are of the same model.

[0062] In this embodiment, the tube usage of multiple sample CT tubes with the same model as the target CT tube can be obtained through the CT equipment Internet of Things platform. The tube usage of each sample CT tube is the tube usage from the time the sample CT tube is installed in the CT equipment and started to be used until the time it is scrapped.

[0063] In this embodiment, since a CT device may use multiple CT tubes during its life cycle, each CT tube used by it, if its model is the same as the model of the target CT tube, can be used as a sample CT tube to record its tube usage. The sample CT tube may be CT tubes from multiple CT devices.

[0064] Please refer to Figure 2 , Figure 2 This is an example diagram of the tube usage provided in this embodiment. Figure 2 In the whole life cycle of a CT device, CT tube A will be turned on and off many times from the time of installation to the time of scrapping. Each time of turning on and off is a usage period. The tube usage in this usage period is calculated, and the tube usage of the CT tube in all usage periods in its life cycle is accumulated to obtain the tube usage of the CT tube. For the tube usage in any usage period, the calculation formula is: tube usage = tube voltage × tube current × exposure time, the unit is KWh, that is, kilowatt-hour, which means the energy consumed in one hour. Figure 2 The time periods shown in are k time periods, and the tube usage in each time period is t1, t2, ..., tk, then the tube usage of CT tube A is t1+t2+...+tk.

[0065] Step S102, extracting a plurality of reference dosages from the reference tube dosage interval.

[0066] In this embodiment, the number of reference dosages can be set according to actual needs. For example, the number of reference dosages is set to 10. The extraction principle can be to extract evenly from the reference tube dosage interval, or to extract according to a preset distribution.

[0067] In step S103, each reference usage is input into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each reference usage. The LSTM model is trained using the tube usage of multiple sample CT tubes as sample data and the failure probability when the used tube usage of each sample CT tube reaches the corresponding sample data as the label of each sample data.

[0068] In this embodiment, the reference failure probability corresponding to each reference usage is the probability of failure of the target CT tube when the tube usage of the target CT tube reaches the reference usage. The label of each sample data is the probability of failure of the sample CT tube when the used tube usage of the sample CT tube corresponding to the sample data reaches the sample data.

[0069] In this embodiment, the LSTM (Long Short-Term Memory) model can effectively capture the long-term dependence and nonlinear characteristics of the tube usage of multiple sample CT tubes, and adaptively learn the complex patterns therein, thereby providing higher prediction accuracy without relying on the assumption of the failure data distribution of multiple sample CT tubes. In this embodiment, as an implementation method, when training the LSTM model, the tube usage can be sequentially input into the LSTM model for training according to the order of length. In addition, in order to eliminate the magnitude difference in the tube usage, it can also be standardized first and then input into the LSTM model.

[0070] Step S104: predicting the reliability of the target CT tube based on all reference failure probabilities.

[0071] In this embodiment, the reliability of the target CT tube may be the remaining tube usage of the target CT tube, that is, the tube usage that can be used when the probability of failure of the target CT tube is a preset failure probability, or it may be the tube usage when the risk of failure of the target CT tube is higher.

[0072] The above method provided in this embodiment makes full use of the LSTM model to effectively capture the long-term dependence and nonlinear characteristics in the fault data of the sample CT tubes of the same model as the target CT tube, and then predicts the reliability of the target CT tube according to the reference failure probability predicted by the LSTM model, thereby improving the accuracy of the reliability prediction of the target CT tube.

[0073] In this embodiment, the prediction method for the remaining tube usage and the prediction method for the tube usage when the risk of failure of the target CT tube is high are different. This embodiment will describe these two prediction methods in turn.

[0074] Method 1: Predict the remaining tube usage

[0075] (1) Obtaining the failure probability threshold and the current tube usage of the target CT tube;

[0076] In this embodiment, the failure probability threshold can be set according to actual needs. If the reliability requirement has a higher priority, the failure probability threshold is set smaller. If the sufficiency of use has a higher priority, the failure probability threshold can be set larger.

[0077] (2) Taking the reference failure probability as the predicted failure probability and taking the reference usage corresponding to the predicted failure probability as the predicted usage;

[0078] (3) Determine the fault probability interval where the fault probability threshold is located from all predicted fault probabilities. The minimum value of the fault probability interval is the maximum value of the predicted fault probability that is less than the fault probability threshold, and the maximum value of the fault probability interval is the minimum value of the predicted fault probability that is greater than the fault probability threshold.

[0079] In this embodiment, the fault probability interval is the minimum interval of all predicted fault probabilities that covers the fault probability threshold, so that the fault probability threshold can be gradually approached through multiple iterations. For example, the predicted failure probabilities are 0.5, 0.7, 0.72, 0.8, and 0.9, and the fault probability threshold is 0.75, then the fault probability interval is (0.72, 0.8).

[0080] (4) If there is no target failure probability in the failure probability interval whose difference between the failure probability and the failure probability threshold satisfies the preset difference, the predicted usage interval is determined based on the predicted usage corresponding to the minimum and maximum values ​​of the failure probability interval;

[0081] In this embodiment, if there is no target failure probability in the failure probability interval whose difference between the failure probability and the failure probability threshold satisfies the preset difference, it means that the target failure probability that meets the preset accuracy requirement has not been approached, and the next iteration is required for further approximation.

[0082] In this embodiment, the preset difference can be set according to the accuracy requirement. For example, if the preset difference is 0.01, the target failure probability satisfies |failure probability threshold - target failure probability|<0.01.

[0083] (5) Re-extracting multiple predicted usages from the predicted usage range;

[0084] In this embodiment, the re-extraction method is the same as the method of extracting multiple reference dosages in step S102, which will not be repeated here. The number of predicted dosages can also be set according to actual needs, and the number of predicted dosages can be the same as or different from the number of reference dosages in step S102.

[0085] (6) Each newly extracted predicted usage is input into the pre-trained long short-term memory (LSTM) model to obtain the predicted failure probability corresponding to each newly extracted predicted usage, and return to (3) until the difference between the failure probability and the failure probability threshold in the failure probability interval of the current iteration satisfies the target failure probability of the preset difference;

[0086] In this embodiment, if the tube usage is standardized before training the LSTM model, the newly extracted predicted usage also needs to be standardized in the same way before being input into the pre-trained LSTM model to obtain the corresponding predicted failure probability.

[0087] (7) Calculate the difference between the predicted usage corresponding to the target failure probability and the current tube usage to obtain the remaining tube usage.

[0088] In this embodiment, in order to more intuitively display the remaining tube usage, please refer to Figure 3 , Figure 3 This is an example diagram of predicting the remaining tube usage in the cumulative distribution function provided in this embodiment, Figure 3 The remaining tube usage RUL is the target failure probability P fail Corresponding predicted usage t RUL Current tube usage t The difference between 0, t 0=2000, failure probability threshold P fail is 0.9, corresponding to t RUL It is about 7798.37, so the remaining tube usage RUL= t RUL - t 0, which is 5798.37.

[0089] Method 2: Predicting high failure usage

[0090] (1) Determine a failure rate function based on all reference failure probabilities, where the failure rate function represents the relationship between the failure rate of the target CT tube and the tube usage;

[0091] In this embodiment, all reference failure probabilities are a set of discrete data. In order to obtain a continuous failure rate function, this embodiment provides an implementation method:

[0092] First, a cumulative distribution function is generated based on all reference failure probabilities;

[0093] Secondly, the cumulative distribution function is differentiated to obtain the probability density function;

[0094] Finally, the failure rate function is generated based on the cumulative distribution function and the probability density function.

[0095] For example, represents the cumulative distribution function, t represents the tube usage, then the probability density function It can be obtained by the following formula: , further, the failure rate function It can be obtained by the following formula: .

[0096] Please refer to Figure 4 , Figure 4 This is an example diagram of the cumulative distribution function obtained by LSTM provided in this embodiment. Figure 4 In the figure, the original data is the data used to train the LSTM model, the discrete LSTM prediction data is the corresponding failure probability obtained by inputting each original data into the LSTM model, and the continuous LSTM prediction data is an example of the continuous cumulative distribution function obtained based on the discrete LSTM prediction data.

[0097] (2) Determine the high failure usage based on the tube usage corresponding to the peak value of the failure rate function.

[0098] In an optional implementation, the high failure usage represents the tube usage when the probability of failure of the target CT tube reaches a preset probability.

[0099] In this embodiment, the time duration corresponding to when the failure rate function reaches a peak value is the high failure duration, which means that the failure rate of the target CT tube is the largest at this time. The preset time period before the high failure duration can be set as high failure usage. The target CT tube in high failure usage has an extremely high risk of failure, which requires special attention from maintenance engineers, or timely maintenance in advance to prevent problems before they occur.

[0100] In this embodiment, in order to make full use of the LSTM model to extract the long-term dependency and nonlinear characteristics in the fault data of the sample CT tube, this embodiment also proposes a training method of the LSTM model:

[0101] First, the original tube usage of each sample CT tube was obtained;

[0102] Secondly, determine the failure probability corresponding to each original tube usage;

[0103] In this embodiment, the method for determining the failure probability corresponding to the original tube usage is:

[0104] (1) Arrange all the original tube usage in ascending order to obtain the tube usage sequence;

[0105] (2) Calculate the failure probability corresponding to each original tube usage based on the serial number of each initial tube usage in the tube usage sequence.

[0106] In this embodiment, for example, the number of original tube usages is n, and the n original tube usages are arranged from small to large to obtain a tube usage sequence. For any original tube usage in the tube usage sequence, its failure probability can be calculated using the following formula: , where i is the original tube usage The serial number in the tube usage sequence, is the failure probability corresponding to the i-th original tube usage in the tube usage sequence.

[0107] Third, the amount of each original tube is standardized, and the amount of each standardized original tube is used as the amount of each tube;

[0108] In this embodiment, a method for implementing the standardization process is as follows:

[0109] (1) Calculate the mean and standard deviation of all original tube usage;

[0110] (2) Standardize the original tube usage according to the mean and standard deviation to obtain the tube usage.

[0111] In this embodiment, for any original tube dosage The following formula can be used for standardization: ,in, is the original tube dosage, is the standardized data, is the standard deviation of all original tube usage, It is the average of all original tube usage.

[0112] It should be noted that when using the pre-trained LSTM model for reliability prediction, in order to maintain the data processing flow consistent with the training phase, the same standard deviation and mean are used in the reliability prediction and training phases to standardize the input data of the LSTM model.

[0113] Fourth, take the usage of a tube as a sample data, take the failure probability corresponding to each sample data as the label of each sample data, use all sample data and labels and the preset loss function to train the pre-built LSTM model to obtain the trained LSTM model.

[0114] In this embodiment, the preset loss function may adopt MSE (Mean Squared Error), and its formula is expressed as: , among which is the i-th fault probability value predicted by the LSTM model, is the failure probability of the label of the i-th sample data, and n is the number of sample data.

[0115] In order to more intuitively illustrate the effect of the method provided in this embodiment, this embodiment compares it with the "gold standard" Weibull distribution in terms of three dimensions: coefficient of determination, mean square error, and mean absolute error.

[0116] Coefficient of determination The formula is expressed as: ,in, is the predicted value, is the true value, n is the number of observations, is the mean of the true value. The closer the coefficient of determination is to 1, the better the fit.

[0117] The expression formula of mean square error MSE is: , the smaller the MSE, the more accurate the prediction.

[0118] The formula for the mean absolute error (MAE) is: .

[0119] The comparison results are shown in Table 1.

[0120] Table 1

[0121]

[0122] It can be seen from Table 1 that the method provided in this embodiment is more accurate than the results obtained by Weibull distribution in these three dimensions.

[0123] For a more intuitive comparison of the two, please refer to Figure 5 , Figure 5 A comparison example diagram of the cumulative distribution function obtained by the LSTM model and Weibull distribution provided in this embodiment. Figure 5 In the figure, the hollow circles represent the original data, the solid line is the curve fitted by the method of the LSTM model provided in this embodiment, and the dotted line is the cumulative distribution function obtained by using the Weibull distribution. Figure 5 It can be seen that the curve fitted by the LSTM model provided in this embodiment is closer to the original data, which means that the reliability prediction result of the method provided in this embodiment is more accurate.

[0124] In order to execute the corresponding steps in the above embodiment and each possible implementation method, a method for implementing the CT tube reliability prediction device 100 is given below. Figure 6 , Figure 6 This is a block diagram of the CT tube reliability prediction device provided in this embodiment. It should be noted that the basic principle and technical effects of the CT tube reliability prediction device 100 provided by the present invention are the same as those of the corresponding above-mentioned embodiments, and are not mentioned in this embodiment for the sake of brief description.

[0125] The CT tube reliability prediction device 100 includes an acquisition module 110 , an extraction module 120 , and a prediction module 130 .

[0126] An acquisition module 110 is used to acquire a reference tube usage interval for predicting the reliability of a target CT tube, wherein a minimum value and a maximum value of the reference tube usage interval are respectively a minimum value and a maximum value of tube usage of a plurality of sample CT tubes, and each sample CT tube and the target CT tube are of the same model;

[0127] An extraction module 120, for extracting a plurality of reference dosages from a reference tube dosage interval;

[0128] The prediction module 130 is used to input each reference usage into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each reference usage. The LSTM model is trained using the tube usage of multiple sample CT tubes as sample data and the failure probability when the used tube usage of each sample CT tube reaches the corresponding sample data as a label of each sample data.

[0129] The prediction module 130 is further configured to predict the reliability of the target CT tube based on all reference failure probabilities.

[0130] In an optional implementation, the reliability of the target CT tube is the remaining tube usage of the target CT tube, and the prediction module 130 is specifically used to:

[0131] Obtain the failure probability threshold and the current tube usage of the target CT tube;

[0132] Taking the reference failure probability as the predicted failure probability and taking the reference usage corresponding to the predicted failure probability as the predicted usage;

[0133] Determine a fault probability interval where the fault probability threshold is located from all predicted fault probabilities, the minimum value of the fault probability interval is the maximum value of the predicted fault probability less than the fault probability threshold, and the maximum value of the fault probability interval is the minimum value of the predicted fault probability greater than the fault probability threshold;

[0134] If there is no target failure probability in the failure probability interval whose difference between the failure probability and the failure probability threshold satisfies the preset difference, the predicted usage interval is determined according to the predicted usage corresponding to the minimum and maximum values ​​of the failure probability interval;

[0135] Re-extract multiple forecast usages from the forecast usage range;

[0136] Input each newly extracted predicted usage into the pre-trained long short-term memory (LSTM) model to obtain the predicted failure probability corresponding to each newly extracted predicted usage, and return to the step of determining the failure probability interval in which the failure probability threshold is located from all predicted failure probabilities until there is a target failure probability in the failure probability interval of the current iteration where the difference between the failure probability and the failure probability threshold satisfies the preset difference;

[0137] The difference between the predicted usage corresponding to the target failure probability and the current tube usage is calculated to obtain the remaining tube usage.

[0138] In an optional implementation, the reliability of the target CT tube is the high failure usage of the target CT tube, and the high failure usage represents the usage of the tube when the probability of failure of the target CT tube reaches a preset probability. The prediction module 130 is further specifically used to:

[0139] Determine a failure rate function according to all reference failure probabilities, wherein the failure rate function represents the relationship between the failure rate of the target CT tube and the tube usage;

[0140] The high failure usage is determined based on the tube usage corresponding to the peak value of the failure rate function.

[0141] In an optional embodiment, when used to determine the failure rate function according to all reference failure probabilities, the prediction module 130 is further used to:

[0142] Generate a cumulative distribution function based on all reference failure probabilities;

[0143] Derivative operation is performed on the cumulative distribution function to obtain the probability density function;

[0144] Generate a failure rate function based on the cumulative distribution function and the probability density function.

[0145] In an optional implementation, the CT tube reliability prediction device 100 further includes a training module 140, and the training module 140 is used to train the LSTM model. The method of obtaining the trained LSTM model is:

[0146] Obtain the original tube usage of each sample CT tube;

[0147] Determine the failure probability corresponding to each original tube usage;

[0148] The amount of each original bulb tube is standardized, and the amount of each original bulb tube after the standardized treatment is used as the amount of each bulb tube;

[0149] The usage of a bulb is taken as a sample data, and the failure probability corresponding to each sample data is taken as the label of each sample data. The pre-built LSTM model is trained using all sample data, labels and a preset loss function to obtain a trained LSTM model.

[0150] In an optional embodiment, the training module 140 is specifically used to determine the failure probability corresponding to each original tube usage:

[0151] Arrange all original tube usages in ascending order to obtain a tube usage sequence;

[0152] According to the sequence number of each original tube usage in the tube usage sequence, the failure probability corresponding to each original tube usage is calculated.

[0153] In an optional embodiment, the training module 140 is further used to perform standardization on each original tube usage:

[0154] Calculate the mean and standard deviation of all original tube usage;

[0155] The amount of each original tube was standardized according to the value and standard deviation to obtain the amount of each tube.

[0156] The embodiment of the present invention also provides a block diagram of an electronic device 10, which implements the CT tube reliability prediction method of the above embodiment. Figure 7 , Figure 7 This is a block diagram of an electronic device 10 provided in this embodiment. The electronic device 10 includes a processor 11, a memory 12 and a bus 13. The processor 11 and the memory 12 are connected via the bus 13.

[0157] The processor 11 may be an integrated circuit chip with signal processing capability. In the implementation process, each step of the CT tube reliability prediction method of the above embodiment may be completed by hardware integrated logic circuits or software instructions in the processor 11. The above processor 11 may be a general processor, including a CPU (Central Processing Unit), a NP (Network Processor), etc.; it may also be a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Logic Gate Array) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0158] The memory 12 is used to store a program for implementing the CT tube reliability prediction method. The program may be a software function module stored in the memory 12 in the form of software or firmware or solidified in the OS (Operating System) of the electronic device 10 .

[0159] After receiving the execution instruction, the processor 11 executes the program to implement the CT tube reliability prediction method of the above embodiment.

[0160] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the CT tube reliability prediction method as described in the above-mentioned embodiment is implemented.

[0161] In summary, the embodiments of the present invention provide a CT tube reliability prediction method, device, electronic device and storage medium, the method comprising: obtaining a reference tube usage interval for predicting the reliability of a target CT tube, the minimum value and the maximum value of the reference tube usage interval being respectively the minimum value and the maximum value of the tube usage of a plurality of sample CT tubes, and each sample CT tube and the target CT tube are of the same model; extracting a plurality of reference usages from the reference tube usage interval; inputting each reference usage into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each reference usage, the LSTM model being trained with the tube usage of a plurality of sample CT tubes as sample data and the failure probability when the used tube usage of each sample CT tube reaches the corresponding sample data as the label of each sample data; and predicting the reliability of the target CT tube based on all reference failure probabilities. Compared with the prior art, this embodiment has at least the following advantages: (1) The LSTM model is fully utilized to effectively capture the long-term dependency and nonlinear characteristics in the fault data of the sample CT tube of the same model as the target CT tube, and then the reliability of the target CT tube is predicted according to the reference fault probability predicted by the LSTM model, thereby improving the accuracy of the reliability prediction of the target CT tube. (2) When using the pre-trained LSTM model for reliability prediction, in order to maintain the data processing flow consistent with the training stage, the same standard deviation and mean are used in the reliability prediction and training stages to standardize the input data of the LSTM model.

[0162] The above are only various embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A CT tube reliability prediction method, characterized in that: The method comprises: Obtaining a reference tube usage interval for predicting the reliability of a target CT tube, wherein a minimum value and a maximum value of the reference tube usage interval are respectively a minimum value and a maximum value of tube usage of a plurality of sample CT tubes, and each of the sample CT tubes and the target CT tube are of the same model; Extracting a plurality of reference dosages from the reference bulb dosage interval; Each of the reference usage is input into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each of the reference usage, wherein the LSTM model is trained using the tube usage of the multiple sample CT tubes as sample data and the failure probability when the used tube usage of each of the sample CT tubes reaches the corresponding sample data as a label of each of the sample data; The reliability of the target CT tube is predicted based on all the reference failure probabilities.

2. The CT tube reliability prediction method according to claim 1, characterized in that: The reliability of the target CT tube is the remaining tube usage of the target CT tube, and the step of predicting the reliability of the target CT tube based on all the reference failure probabilities includes: Obtaining a failure probability threshold and a current tube usage of the target CT tube; Using the reference failure probability as the predicted failure probability and using the reference usage corresponding to the predicted failure probability as the predicted usage; Determine a fault probability interval in which the fault probability threshold is located from all the predicted fault probabilities, wherein the minimum value of the fault probability interval is the maximum value of the predicted fault probability less than the fault probability threshold, and the maximum value of the fault probability interval is the minimum value of the predicted fault probability greater than the fault probability threshold; If there is no target failure probability in the failure probability interval whose difference between the failure probability and the failure probability threshold satisfies the preset difference, determining the predicted usage interval according to the predicted usage corresponding to the minimum value and the maximum value of the failure probability interval; Re-extracting a plurality of predicted usages from the predicted usage interval; Input each newly extracted predicted usage into a pre-trained long short-term memory (LSTM) model to obtain a predicted failure probability corresponding to each newly extracted predicted usage, and return to the step of determining the failure probability interval in which the failure probability threshold is located from all the predicted failure probabilities until there is a target failure probability in the failure probability interval of the current iteration where the difference between the failure probability and the failure probability threshold satisfies the preset difference; The difference between the predicted usage corresponding to the target failure probability and the current tube usage is calculated to obtain the remaining tube usage.

3. The CT tube reliability prediction method according to claim 1, characterized in that: The reliability of the target CT tube is the high failure usage of the target CT tube, and the high failure usage represents the tube usage when the probability of failure of the target CT tube reaches a preset probability. The step of predicting the reliability of the target CT tube based on all the reference failure probabilities includes: Determine a failure rate function according to all the reference failure probabilities, wherein the failure rate function represents the relationship between the failure rate of the target CT tube and the tube usage; The high failure usage is determined according to the tube usage corresponding to the peak value of the failure rate function.

4. The CT tube reliability prediction method according to claim 3, characterized in that: The step of determining the failure rate function according to all the reference failure probabilities comprises: generating a cumulative distribution function based on all of the reference failure probabilities; Performing a derivative operation on the cumulative distribution function to obtain a probability density function; The failure rate function is generated according to the cumulative distribution function and the probability density function.

5. The CT tube reliability prediction method according to claim 1, characterized in that: The way to get the trained LSTM model is: Obtaining the original tube usage of each of the sample CT tubes; Determine the failure probability corresponding to each of the original tube usages; The amount of each original bulb tube is standardized, and the amount of each original bulb tube after the standardized process is used as the amount of each bulb tube; The usage of the bulb is taken as a sample data, and the failure probability corresponding to each sample data is taken as the label of each sample data. The pre-constructed LSTM model is trained using all the sample data, the labels and the preset loss function to obtain the trained LSTM model.

6. The CT tube reliability prediction method according to claim 5, characterized in that: The step of determining the failure probability corresponding to each of the original tube usages comprises: Arrange all the original tube usages in ascending order to obtain a tube usage sequence; According to the sequence number of each original tube usage in the tube usage sequence, the failure probability corresponding to each original tube usage is calculated.

7. The CT tube reliability prediction method according to claim 5, characterized in that: The step of standardizing the amount of each original tube comprises: Calculate the mean and standard deviation of all the original tube usages; The original tube usage of each of the tubes is standardized according to the mean and the standard deviation to obtain the tube usage of each of the tubes.

8. A CT tube reliability prediction device, characterized in that: The device comprises: An acquisition module, used for acquiring a reference tube usage interval for predicting the reliability of a target CT tube, wherein the minimum value and the maximum value of the reference tube usage interval are respectively the minimum value and the maximum value of the tube usage of a plurality of sample CT tubes, and each of the sample CT tubes and the target CT tube are of the same model; An extraction module, used for extracting a plurality of reference dosages from the reference bulb dosage interval; A prediction module is used to input each of the reference usage into a pre-trained long short-term memory (LSTM) model to obtain a reference failure probability corresponding to each of the reference usage, wherein the LSTM model is trained using the tube usage of the multiple sample CT tubes as sample data and the failure probability when the used tube usage of each of the sample CT tubes reaches the corresponding sample data as a label of each of the sample data; The prediction module is further used to predict the reliability of the target CT tube based on all the reference failure probabilities.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory is used to store a program, and the processor is used to implement the CT tube reliability prediction method according to any one of claims 1 to 7 when executing the program.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the CT tube reliability prediction method according to any one of claims 1 to 7 is implemented.

Images