Method for predicting a failure of at least one component of an x-ray tube

Abstract

The invention relates to a method for predicting a failure of at least one component of an X-ray tube (11), wherein at least one reference value from the X-ray tube (11) is stored in an evaluation unit of a control system for the X-ray tube (11), wherein the X-ray tube (11) comprises a cathode (16) with a filament (18) and at least one filament current of the cathode (16), an anode current and / or an anode voltage of an anode (14) of the X-ray tube (11) are each stored as a reference value, wherein during operation of the X-ray tube (11) actual values of at least the filament current, the anode current and / or the anode voltage are measured and fed to the evaluation device, wherein in the evaluation device at least a filament failure prediction and a vacuum leakage prediction in a housing (12) of the X-ray tube (11) are determined based on at least a comparison of the reference values with the actual measured values during operation of the X-ray tube (11), and based on the determination of the filament failure prediction and the vacuum leakage prediction in the housing (12) of the X-ray tube (11) at least one of a failure of the filament (18) and a leakage of the housing (12) of the X-ray tube (11) is output.

Description

Technical Field

[0001] The present invention relates to a method for predicting failure of at least one component in an X-ray tube. Background Technology

[0002] US 10,753,969 B2 discloses a method and system for predicting failures during the lifespan of an X-ray tube. In this method, data is used to train an X-ray tube model. This data may include both the voltage and current of the X-ray tube. Based on this predetermined data, an algorithm can be used to predict failures or estimate the remaining lifespan. The method provides a representation of the X-ray tube current in the X-ray tube model as a function of the filament current and the X-ray tube voltage. Based on this, failures of the cathode filament can be predicted. Other possible sources of failure in the X-ray tube, such as leaks in the X-ray tube housing—which impede vacuum maintenance and could therefore lead to premature failure—are not predicted.

[0003] DE 10 2009 036 940 A1 discloses a method for determining the aging condition of an X-ray tube cathode. The method involves determining a reference value for the current flowing through the cathode based on the voltage applied to it, and using this data to determine the aging condition based on the relationship between the actually measured cathode current and the applied voltage. If the detected deviation exceeds the allowable tolerance, an alarm message is issued, indicating that the X-ray tube is nearing the end of its service life. Furthermore, the method is designed to detect electric arcs occurring within the X-ray tube. These arcs occur particularly during vacuum degradation within the X-ray tube and also contribute to aging. These arcs are indirectly measured via the internal pressure of the X-ray tube by measuring the pump current of the vacuum ion pump. Direct detection of leaks within the X-ray tube is not provided.

[0004] DE 103 38 693 B3 discloses a method for estimating the remaining lifetime of an X-ray source. Under specified test conditions, measurements suitable for determining the remaining lifetime of an operating X-ray source are determined at regular intervals and stored in a storage device. Based on the current measurements and past measurements stored in the storage device—which were determined under the same test conditions—a predicted trend of future measurements is then estimated, and the predicted remaining lifetime of the X-ray tubes is determined based on this predicted trend and thresholds assigned to individual X-ray tubes and stored in the storage device.

[0005] DE 103 14 538 A1 discloses a method for measuring pressure present in an X-ray tube.

[0006] DE 10 2012 204 138 A1 discloses a method for predicting the lifespan of an X-ray generator. For this purpose, data on at least one physical parameter affecting the generator's lifespan are stored at regular intervals. The stored data is evaluated using a predefined statistical model to determine the expected remaining lifespan of the X-ray generator. Summary of the Invention

[0007] The present invention is based on the task of proposing a method for predicting X-ray tube failures, wherein multiple components of the X-ray tube are monitored simultaneously, and at least one failure prediction is determined for each monitored component of the X-ray tube.

[0008] This task is accomplished by the method described in claim 1. The method stores reference values ​​from the X-ray tube (preferably a new or first-time-operation X-ray tube) in the memory of the X-ray tube control system, wherein at least the cathode filament current, the anode current of the X-ray tube anode, and / or a predetermined accelerating voltage or the anode voltage of the X-ray tube are stored. During X-ray tube operation, particularly in an X-ray fluorescence analyzer, current values ​​from the X-ray tube are measured, wherein at least the filament current, anode voltage, and / or anode current are measured and fed to the evaluation device. In the evaluation device, based on a comparison of the reference values ​​and the measured values, at least one of a filament failure prediction and a vacuum leak prediction in the X-ray tube housing is determined, and at least one of the filament failure and the vacuum leak failure of the X-ray tube is evaluated; and preferably, the result of at least one next failure is output.

[0009] The advantage of this method is that it allows for the prediction of potential failures in multiple components of the X-ray tube, and at least the ability to distinguish between filament failure and vacuum leaks in the X-ray tube housing. This improves the quality assurance of equipment or systems operating with X-ray tubes. Furthermore, predictive maintenance can be implemented to avoid unexpected downtime of the X-ray tube.

[0010] Preferably, in order to predict at least one of a filament fault or a vacuum leak in the housing, the trend of the filament current is assessed, and preferably the ratio between the filament current and the anode voltage is assessed, with a decrease in the current filament current indicating a filament fault and an increase in the current filament current indicating a vacuum leak. This change in filament current, particularly under otherwise constant operating conditions, can generally serve as a basis for distinguishing between filament faults and housing vacuum leaks.

[0011] To refine fault prediction, at least for filament and vacuum leaks, it is preferable to additionally measure the filament resistance and feed it to the evaluation apparatus. During X-ray tube operation, the filament resistance increases monotonically over time. By incorporating the filament resistance and its changes into fault prediction, and preferably into determining which fault will occur first, fault trends can be detected very early, or it can be determined which component of the X-ray tube will be affected by a fault.

[0012] Specifically, a method is provided to normalize the values ​​recorded in the evaluation apparatus for predicting faults during X-ray tube operation before comparing them with stored reference values. This has the advantage that the evaluation of the recorded values ​​for predicting faults is independent of the user-defined X-ray tube operating conditions.

[0013] According to an advantageous implementation of the method, a filament fault is detected when the anolyte current in the X-ray tube is constant and a decrease in filament current and an increase in filament resistance are detected. Surprisingly, it was found that both the decrease in filament current and the increase in filament resistance are closely related to the prediction of filament faults.

[0014] Furthermore, it is preferably provided that, when the anode current is constant, a vacuum leak in the X-ray tube is detected when an increase in the filament current is detected and the filament resistance is constant or unchanged. Surprisingly, it has also been found that, when the filament resistance is constant and the filament current increases, an impending failure due to a vacuum leak in the X-ray tube can be predicted.

[0015] To perform fault prediction, the filament current is preferably stored in the evaluation device during the preparatory step, and preferably as a function of anode current and anode voltage. Here, multiple values ​​of the filament current are preferably stored as a function of anode voltage and anode current that can be set by the user. This allows baseline values ​​of unused or soon-to-be-operational X-ray tubes to be used as a basis; based on these, variations during X-ray tube operation can be determined, and faults can be predicted accordingly.

[0016] As a first step in determining X-ray tube faults, a divergence calculation is preferably performed. In this process, actual operating data of the X-ray tube is recorded and compared with stored reference values, thereby evaluating the device and calculating a divergence value. Based on this divergence value, subsequent steps can then determine which fault is more likely to have occurred earlier. Preferably, an averaging curve is provided for calculating the divergence value based on the filament current, and more preferably based on the current value of the filament current and / or based on the filament resistance determined by the current values ​​of the filament current and filament voltage, the determination of the averaging curve being independent of user-set anode operating parameters. This allows for more accurate prediction of the corresponding faults.

[0017] Furthermore, it is preferable to use a "conditional random field" or an autoencoder neural network as an evaluation algorithm to determine the divergence value and use it as the basis for further fault prediction.

[0018] In the second step of the prediction used to determine X-ray tube faults, a selection process is preferably performed. This selection is based on the difference determined in the previous step through divergence calculation. During this process, a preliminary selection can be made for the next fault that might lead to failure. At least filament fault prediction and vacuum leak prediction are evaluated and initiated.

[0019] When predicting filament failure, it is preferable to compare the measured value with reference data to determine the expected failure time; preferably, the logarithmic inverse curve of the divergence value is used together with regression techniques as the basis for calculating the failure time, thereby dynamically selecting first-order, second-order, etc. regression techniques.

[0020] When predicting vacuum leaks, it is preferable to use actual measurements of the filament current and compare them with a threshold to determine the timing of the failure. Advantageously, the timing of the filament failure can be determined by fitting a curve based on the recorded filament current values.

[0021] In subsequent steps, the output multiplexer selects data (particularly the data determined in the second step) for use in the output results. During this process, the amount of information to be output can be determined. For example, only the first predicted fault can be output. Alternatively, the weighted sum of the determined predictions can be output, making the order of successive faults, and in particular, the corresponding expected times, apparent.

[0022] The steps described above for predicting X-ray tube failures not only allow for differentiation of failure types or identification of at least two components of the X-ray tube, but also enable the determination of the expected failure time for each corresponding component. This allows for the timely initiation of predetermined measures, enabling the necessary replacement of affected components to be performed shortly before or at the time of failure. This helps prevent uncontrolled operational downtime. Attached Figure Description

[0023] The invention, along with other advantageous embodiments and variations thereof, is described and explained in more detail below with reference to the examples shown in the accompanying drawings. Features disclosed in the specification and drawings may be applied individually or in any combination thereof according to the invention. The drawings show: Figure 1 Schematic diagram of an X-ray tube. Figure 2 A schematic diagram illustrating the steps involved in predicting X-ray tube failure is shown. Figure 3 A schematic diagram for detecting filament faults in X-ray tubes, and Figure 4A schematic diagram used to detect vacuum leaks in X-ray tubes. Detailed Implementation

[0024] Figure 1 A schematic diagram of an X-ray tube 11 is shown. The X-ray tube 11 includes a housing 12, within which a vacuum is maintained. An anode 14 and a cathode 16 opposite the anode are arranged within the housing 12. The anode 14 preferably includes a cooling system 17. The cathode 16 is preferably designed as a thermocathode and includes a filament 18. A filament voltage—also called a heating voltage—is applied to drive the cathode 16. An anode voltage is applied between the cathode 16 and the anode 14. This is also called an accelerating voltage because it accelerates electrons traveling from the cathode 16 to the anode 14. Electrons emitted by the cathode 14 or the filament 18 are emitted as an electron beam 19 and strike the anode 14. The material of the anode 14 emits fluorescent radiation 21 and bremsstrahlung radiation, also known as X-rays. This X-ray radiation can be used in the medical field. This X-ray radiation can also be used to analyze the material of an object and / or at least one layer of material on an object and / or the material of the object to which the layer is applied.

[0025] During the operation of cathode 16, the cathode voltage (or filament voltage) and filament current are measured. Simultaneously, during the operation of X-ray tube 11, the anode voltage and anode current are measured. The anode current is set by the user for operating X-ray tube 11. Due to the user's adjustment of the anode current, the filament voltage is automatically adjusted, ensuring that the filament current generated by the filament resistance produces the correct anode current under the influence of the set anode voltage.

[0026] During operation of the X-ray tube 11, the filament 18 may change due to emitted electrons. Furthermore, leaks may occur in the housing 12 of the X-ray tube 11, causing the vacuum to be unable to be maintained, or perhaps only maintained to a limited extent. As will be described below, at least these two components 16, 18 of the X-ray tube 11 should be considered in terms of fault prediction. Other components, such as the anode 14, may also be considered in fault prediction for potential failures.

[0027] Figure 2 A schematic diagram of the steps involved in predicting at least one failure of a component of an X-ray tube is shown.

[0028] In a method for predicting failures of at least one component 16, 18 of an X-ray tube 11, reference data is stored in an evaluation device of the control system (not shown in detail), specifically in a memory, before or during startup of the X-ray tube 11. These are preferably reference data for the corresponding filament current, and more specifically, reference data for the filament voltage of the filament 18 of the cathode 16, which depends on the adjustable anode voltage and anode current. To predict failures of at least one component 16, 18 of the X-ray tube, a divergence calculation is performed in a first step 31. In this divergence calculation, actual measurements of the X-ray tube 11 during operation are compared with reference values ​​in the evaluation device. The filament current depends on the anode current and anode voltage, whereby the anode voltage can be set by the user.

[0029] For divergence calculation, the filament current as a function of anode current and cathode voltage is first calculated and then normalized. This yields a smooth, continuously differentiable filament current curve. Preferably, the filament resistance is also determined in this step based on the measured filament current and filament voltage. This determined data is compared with reference data to subsequently determine the deviation. For the combination of normalization and deviation calculation, a conditional random number generator or an autoencoder neural network can be used, for example. Through this first step of divergence calculation, at least for the filament current, a divergence value is output.

[0030] In the subsequent or second step 34, the divergence value forms the basis of the selection process. This selection process determines which fault is more relevant, i.e., which is likely to occur first. If the divergence value is determined to increase, the system concludes that a leakage fault is likely. If the divergence value is determined to decrease, a filament fault is anticipated.

[0031] Based on the fault selection in step 34 described above, the expected failure time is calculated in step 36. This step 36 can be subdivided into several separate steps. The first separate step 37 aims to determine the service life of the filament 18 before its failure. The second separate step 38 aims to determine the service life of the housing 12 before a vacuum leak failure occurs. In the third separate step 39 or subsequent separate steps, other possible failures of components of the X-ray tube 11 can be determined individually.

[0032] Figure 3The diagram illustrates the filament current IF as a function of the lifespan t of filament 18. The filament current remains constant over a longer period and only begins to decrease exponentially after this longer period. This curve serves as the basis for determining the remaining lifespan. In addition to the filament current, the filament resistance based on measurements from X-ray tube 11 is also considered. To determine the time point or remaining lifespan, the values ​​from the logarithmic inverse curve of the divergence values ​​are combined with first-order, second-order, and other regression techniques. The determined time point is output in step 37.

[0033] Figure 4 A schematic diagram is shown, depicting the filament current IF and anode current IA over a predetermined duration t. Based on the curves of the filament current IF and anode current IA, the remaining service life of housing 12 with respect to leakage failure can be predicted. Near the end of the service life of housing 12, the filament current IF increases until it reaches its threshold at a constant anode current IA. From that point onward, both the filament current IF and the anode current IA immediately decrease due to the occurrence of the leakage failure. To determine the leakage failure, a fit is performed on the normalized filament current IF or the divergence value calculated from the normalized filament current IF to detect a rise from a nearly constant value, at which point the failure occurs. Step 38 outputs the determined time point.

[0034] exist Figure 2 In the fourth step 42 of the monitored component, the remaining service life determined in the third step 36 is output. This can be displayed on the screen of the evaluation device. In this step 42, according to the first embodiment, the faults selected by the selection process in the second step 34 can be calculated in the third step 36 and then output in the fourth step 42. No further calculations are performed. Alternatively, in parallel, at least one additional fault not selected in the selection process can also be calculated in step 36 and additionally output in step 42. This allows the user to obtain an overview of the expected faults and their timing, and furthermore, allows the user to decide whether to repair only the initially predicted faults or, for example, repair additional predicted faults simultaneously during maintenance.

Claims

1. A method for predicting failure of at least one component of an X-ray tube (11), in, At least one reference value from the X-ray tube (11) is stored in an evaluation unit for the control system of the X-ray tube (11), wherein the X-ray tube (11) includes a cathode (16) having a filament (18), and at least one filament current of the cathode (16), the anode current of the anode (14) of the X-ray tube (11), and / or the anode voltage are each stored as reference values. During the operation of the X-ray tube (11), the actual values ​​of at least the filament current, the anode current, and / or the anode voltage are measured and fed to the evaluation device. Specifically, at least the filament failure prediction and the vacuum leakage prediction in the housing (12) of the X-ray tube (11) are determined based on a comparison of the reference value with the actual measured value during the operation of the X-ray tube (11); and based on the determination of the filament failure prediction and the vacuum leakage prediction in the housing (12) of the X-ray tube (11), at least one of the filament (18) failure prediction and the X-ray tube (11) housing (12) leakage prediction is output.

2. The method according to claim 1, characterized in that, In order to predict the failure of the filament (18) and the vacuum leakage from the housing (12), the trend of the filament current is evaluated. Preferably, the ratio between the filament current and the anode voltage and / or the ratio between the filament current and the anode current is evaluated. If a decrease in the filament current is detected, it indicates a filament failure. If an increase in the filament current is detected, it indicates a leakage failure.

3. The method according to claim 1 or 2, characterized in that, For at least one of the filament fault prediction and leakage fault prediction, the filament resistance is additionally detected and provided to the evaluation device.

4. The method according to any one of the preceding claims, characterized in that, Before being compared with the reference value, the values ​​used for prediction detected during the operation of the X-ray tube (11) are normalized in the evaluation apparatus.

5. The method according to claim 3 or 4, characterized in that, When the anode current is constant and / or the anode voltage is constant, if a decrease in the filament current and an increase in the filament resistance are detected, a filament failure is predicted.

6. The method according to claim 3 or 4, characterized in that, When the anode current is constant and / or the anode voltage is constant, if an increase in the filament current and a constant filament resistance are detected, a vacuum leak in the housing (12) is predicted.

7. The method according to any one of the preceding claims, characterized in that, In a preliminary step for determining a fault prediction of at least one component of the X-ray tube (11), the filament current is stored in the evaluation device, preferably as a function of the anode current and the anode voltage, wherein multiple values ​​of the filament current are stored as a function of the anode voltage and / or the anode current that can be adjusted by the user.

8. The method according to any one of the preceding claims, characterized in that, In the first step, a divergence calculation is performed for at least one of the fault predictions, wherein, for the first step, the current value of the X-ray tube (11) is recorded, the current value including at least the anode current, anode voltage and / or filament voltage, and an evaluation algorithm is used to calculate the divergence value based on the reference value.

9. The method according to claim 8, characterized in that, An average value curve is determined based on the current value of the filament current and / or based on the filament resistance determined by the current values ​​of the filament current and the filament voltage, the determination of the average value curve being independent of the set operating voltage and operating current of the anode.

10. The method according to claim 8 or 9, characterized in that, A conditional random generator or an autoencoder neural network is used as the evaluation algorithm to determine and output the difference.

11. The method according to any one of claims 8 or 9, characterized in that, In the second step (34) of the fault prediction for at least one fault prediction, fault selection is performed based on the difference determined in the first step, and it is determined whether at least the filament fault or the vacuum leak will occur as the next fault.

12. The method according to claim 11, characterized in that, In the subsequent step (36) for predicting the filament failure, the actual determined data is compared with the reference data, and the expected time of the failure is determined. Preferably, the expected time of the failure is calculated by combining the logarithmic inverse curve of the divergence value with first-order, second-order, and other regression techniques.

13. The method according to claim 11, characterized in that, In a subsequent step (36) for predicting vacuum leaks, the actual measured filament current data is compared with a threshold to determine the expected time of the fault, and preferably, a curve formed based on the recorded filament current data is fitted.

14. The method according to any one of the preceding claims, characterized in that, In the fourth step (42), the prediction of the first failure to occur is output, or a weighted average of the predictions of at least filament failure and vacuum leakage is performed, and the sequence and / or timing of the expected failures are output.

Images