Optical module fault detection method and device and electronic equipment

Abstract

The invention provides an optical module fault detection method and device and electronic equipment. According to the invention, under the condition that the fault detection result of the sample data of the N sample optical modules is not matched with the sample label of each sample optical module, the current weight of each detection rule is adjusted one by one in sequence, and after the current weight of one detection rule is adjusted each time, the fault detection result of the sample data of the N sample optical modules is obtained. Re-determining a reference detection result of each sample optical module according to the adjusted current weight, and if the number of unmatched sample optical modules is not less than N, indicating that the current weight adjustment is not appropriate, and re-adjusting the current weight of the detection rule; if the number of the unmatched sample optical modules is less than N, indicating that the current weight is properly adjusted, and adjusting the current weight of the next detection rule until a target weight is obtained; through adaptive adjustment of the weight corresponding to each detection rule, the obtained target weight can perform fault detection on the optical module more accurately.

Description

Technical Field

[0001] This application relates to the field of optical communication technologies, and particularly to an optical module fault detection method, apparatus, and electronic device. Background Art

[0002] An optical module is a key component for realizing signal conversion and data transmission between network devices and optical fibers. In practical applications, the performance of an optical module may decay after long-term operation, resulting in unstable data transmission and affecting the integrity of data reception and transmission. However, the current fault detection accuracy of the fault detection scheme for optical modules is relatively low, and it is impossible to accurately detect faults in optical modules. Summary of the Invention

[0003] In view of this, this application provides an optical module fault detection method, apparatus, and electronic device to improve the fault detection accuracy of optical modules.

[0004] The technical solutions provided by this application are as follows: According to an embodiment of the first aspect of this application, an optical module fault detection method is provided, and the method includes: Obtain sample data of multiple sample optical modules; Determine whether the current detection result of each sample optical module matches the sample label of each sample optical module; the current detection result of any sample optical module is determined based on the current weights of at least two detection rules and the detection result obtained by performing fault detection on the sample data of the sample optical module by each detection rule; If the current detection results of N sample optical modules do not match the sample labels and N is greater than a set threshold, then sequentially adjust the current weights of each detection rule one by one, and after adjusting the current weight of each detection rule each time, determine the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; If it is found that the reference detection results of P sample optical modules do not match the sample labels, and P is greater than or equal to N, then the current weight of this detection rule is readjusted, and the step of determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule is returned; if it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N, and M is greater than the set threshold, then when this detection rule is not the last detection rule, the current weight of the next detection rule is adjusted in sequence, and when this detection rule is the last detection rule, the current weights of each detection rule are determined as the target weights of each detection rule, so as to determine the fault detection result of the optical module to be detected based on the target weights of each detection rule and the detection results obtained by performing fault detection on the optical module to be detected by each detection rule.

[0005] According to an embodiment of the second aspect of the present application, an optical module fault detection device is provided, and the device includes: A sample acquisition unit, configured to acquire sample data of a plurality of sample optical modules; A result judgment unit, configured to judge whether the current detection results of each sample optical module match the sample labels of each sample optical module; the current detection result of any sample optical module is determined based on the current weights of at least two detection rules and the detection results obtained by performing fault detection on the sample data of this sample optical module by each detection rule; A first determination unit, configured to, if the current detection results of N sample optical modules do not match the sample labels, adjust the current weights of each detection rule one by one in sequence, and after adjusting the current weight of each detection rule each time, determine the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; A second determination unit, configured to, if it is found that the reference detection results of P sample optical modules do not match the sample labels, and P is greater than or equal to N, then readjust the current weight of this detection rule, and return the step of determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; if it is found that the reference detection results of M sample optical modules do not match the sample labels, and M is less than N, then when this detection rule is not the last detection rule, adjust the current weight of the next detection rule in sequence, and when this detection rule is the last detection rule, determine the current weights of each detection rule as the target weights of each detection rule, so as to determine the fault detection result of the optical module to be detected based on the target weights of each detection rule and the detection results obtained by performing fault detection on the optical module to be detected by each detection rule.

[0006] According to an embodiment of the third aspect of the present application, an electronic device is provided, including: a processor and a machine-readable storage medium storing machine-executable instructions that can be executed by the processor; the processor is configured to execute the machine-executable instructions to implement the method described in the first aspect.

[0007] As can be seen from the above technical solutions, when the fault detection results of the sample data of N sample optical modules do not match the sample labels of the respective sample optical modules, the current weights of the respective detection rules are adjusted one by one in sequence. After adjusting the current weight of one detection rule each time, the reference detection results of the respective sample optical modules are re-determined according to the adjusted current weight. If the number of sample optical modules with mismatches is not less than N, it indicates that the current weight adjustment is inappropriate, and then the current weight of this detection rule is adjusted again; if the number of sample optical modules with mismatches is less than N, it indicates that the current weight adjustment is appropriate, and then the current weight of the next detection rule is adjusted until the target weight is obtained; through the adaptive adjustment of the weights corresponding to the respective detection rules, the obtained target weight can more accurately detect faults in the optical modules. Description of the Drawings

[0008] The drawings here are incorporated into the description and form a part of the description, showing embodiments consistent with the present application and used together with the description to explain the principles of the present application.

[0009] Figure 1 It is a flowchart of the optical module fault detection method provided by the embodiment of the present application; Figure 2A It is a schematic diagram of the statistical characteristics of the transmission power of a normal optical module provided by the embodiment of the present application; Figure 2B It is a schematic diagram of the statistical characteristics of the transmission power of a faulty optical module provided by the embodiment of the present application; Figure 3 It is a schematic diagram of the overall application of the optical module fault detection method provided by the embodiment of the present application; Figure 4 It is a structural diagram of an optical module fault detection device provided by the embodiment of the present application; Figure 5 It is a schematic diagram of the structure of an electronic device provided by the embodiment of the present application. Detailed Embodiments

[0010] In order to enable those skilled in the art to better understand the technical solutions provided by the embodiments of the present application and make the above objects, features, and advantages of the embodiments of the present application more obvious and understandable, the technical solutions in the embodiments of the present application will be further described in detail below with reference to the drawings.

[0011] An optical module is a key component for realizing signal conversion and data transmission between network devices and optical fibers. In practical applications, the performance of the optical module will decay after long-term operation, resulting in unstable data transmission and affecting the integrity of data reception and transmission. Therefore, it is necessary to detect faults in the optical module to identify the optical modules that may malfunction, so as to discover and eliminate potential faults in advance.

[0012] Among them, the fault detection of the optical module can be to detect various types of index data of the optical module, and determine whether the optical module is abnormal according to whether the index data is abnormal.

[0013] As an embodiment, the state of the optical module can be evaluated based on index data such as the transmitted optical power, received optical power, bias current, voltage, temperature, CRC (Cyclic Redundancy Check) error packet rate, and operating state (Up / Down) of the optical module.

[0014] However, the current fault detection accuracy of the fault detection scheme for optical modules is relatively low, and it is impossible to accurately detect faults in optical modules.

[0015] Based on this, the present application proposes a method for detecting faults in optical modules to improve the accuracy of fault detection of optical modules.

[0016] Please refer to Figure 1 , Figure 1 which is a flowchart of the method for detecting faults in optical modules provided by the embodiments of the present application.

[0017] As Figure 1 shown, the method may include the following steps: Step 101, obtaining sample data of multiple sample optical modules.

[0018] In this embodiment, sample data of multiple sample optical modules can be obtained in advance. Among them, in order to ensure correct anomaly detection results for both normal optical modules and faulty optical modules, the multiple sample optical modules can include at least one normal optical module and at least one faulty optical module.

[0019] As an embodiment, the sample data of each sample optical module can be a time series data, and the time series data can include the recorded values of at least one type of index data of the sample optical module within a specified period. For example, the sample data of a sample optical module can be a set of the transmitted optical power of the sample optical module within 30 days, and the present application does not limit this.

[0020] For the sample data of multiple acquired sample optical modules, their corresponding sample labels can be normal or faulty. For example, the sample label of the sample data of a normal optical module is normal, that is, when performing a fault detection on the sample data of a normal optical module, the correct detection result should be normal. The sample label of the sample data of a faulty optical module is faulty, that is, when performing a fault detection on the sample data of a faulty optical module, the correct detection result should be faulty.

[0021] As an example, different numerical values can be used to represent different detection results. For example, 0 represents normal and 1 represents faulty. Of course, other numerical values can also be used to characterize different detection results, and the present application does not limit this.

[0022] So far, the description of step 101 is completed. Next, step 102 is executed.

[0023] Step 102: Determine whether the current detection result of each sample optical module matches the sample label of each sample optical module.

[0024] Among them, the current detection result of any sample optical module is determined based on the current weights of at least two detection rules and the detection results obtained by performing fault detection on the sample data of this sample optical module by each detection rule.

[0025] In this embodiment, for the sample data of any sample optical module, the sample data can first be subjected to fault detection according to different detection rules to obtain an initial detection result. Further, based on the initial detection results corresponding to each detection rule and the current weights of each detection rule, the current detection result of this sample optical module can be determined.

[0026] As an example, the specific method for determining the current detection result of this sample optical module based on the initial detection results corresponding to each detection rule and the current weights of each detection rule can be to perform a weighted operation on the initial detection results corresponding to each detection rule based on the current weights of each detection rule, and determine the current detection result of this sample optical module according to the weighted operation result.

[0027] For example, taking 0 to represent normal and 1 to represent a fault as an example, first, a judgment threshold can be preset (such as 0.4). If the preset weighted operation result is greater than or equal to this judgment threshold, it is determined that the current detection result is a fault. If the weighted operation result is less than this judgment threshold, it is determined that the current detection result is normal. Suppose a sample optical module is detected for faults through two different detection rules (denoted as detection rule 1 and detection rule 2, and their corresponding current weights are 0.4 and 0.6 respectively). For detection rule 1, the initial detection result obtained is normal, and for detection rule 2, the initial detection result obtained is a fault. Then, the weighted operation can be performed on the initial detection results corresponding to each detection rule based on the current weight of each detection rule to obtain the weighted operation result: 0 * 0.4 + 1 * 0.6 = 0.6. Since 0.6 is greater than the above judgment threshold of 0.4, it can be determined that the current detection result of this sample optical module is a fault.

[0028] In this embodiment, multiple detection rules for detecting faults in the optical module can be preset in advance. The sample data of the sample optical module is respectively detected for faults using the above detection rules. For each detection rule, the initial detection result corresponding to this rule can be obtained. Further, based on the initial detection result corresponding to each detection rule and the current weight of each detection rule, the current detection result of this sample optical module is determined.

[0029] Among them, the detection rules for detecting faults in the optical module can include but are not limited to threshold discrimination detection rules, trend discrimination detection rules, abnormal point quantity detection rules, and model detection rules. This application does not limit this.

[0030] Several representative detection rules are introduced below.

[0031] (1) Threshold discrimination detection rule This detection rule can include: performing fault detection on the index data of the optical module according to the determined actual detection range; if the index data of the optical module is not within the actual detection range, it is determined that the detection result obtained by performing fault detection on the index data of this optical module by this detection rule is a fault; if the index data of the optical module is not within the actual detection range, it is determined that the detection result obtained by performing fault detection on the index data of this optical module by this detection rule is no fault.

[0032] In this embodiment, the normal fluctuation range of each type of index data can be determined according to the statistical characteristics of the actually detected various types of index data in the normal optical module; according to the normal fluctuation range and the alarm range configured for each type of index data at the factory, the actual detection range of each type of index data is determined.

[0033] Further, fault detection is performed on the optical module according to the determined actual detection range.

[0034] Specifically, a large number of normal optical modules can be actually detected in advance, and various types of index data of a large number of normal optical modules are collected. For each index data, its mean value "mean" and standard deviation "std" are statistically obtained. Then, its normal fluctuation range [Value_lo, Value_hi] = [mean + 6 * std, mean - 6 * std], where Value_lo refers to the minimum value of the normal fluctuation range, and Value_hi refers to the maximum value of the normal fluctuation range. Here, [mean + 6 * std, mean - 6 * std] is only an exemplary range value, indicating that this range is determined by the statistical characteristics of each index data (here, the mean value and the standard deviation). The present application does not limit the specific determination method of this range.

[0035] Meanwhile, for each type of optical module, when leaving the factory, an alarm range is usually set for each index data, that is, if the index data exceeds this alarm range, it is considered that the optical module may have a fault. This alarm range configured at the factory is usually relatively wide, and the above-determined normal fluctuation range is usually included in this alarm range configured at the factory. Here, the alarm range configured at the factory is denoted as [Fac_lo, Fac_hi], where Fac_lo refers to the minimum value of the alarm range configured at the factory, and Fac_hi refers to the maximum value of the alarm range configured at the factory.

[0036] In this embodiment, the actual detection range of each type of index data can be determined according to the normal fluctuation range and the alarm range configured at the factory for each type of index data.

[0037] For example, the determined actual detection range can be [Warn_lo, Warn_hi], where Warn_lo refers to the minimum value of the actual detection range, and Warn_hi refers to the maximum value of the actual detection range; Warn_lo = (Value_lo + Fac_lo) / 2; Warn_hi = (Value_hi + Fac_hi) / 2. That is, the mean value of the normal fluctuation range and the minimum value of the alarm range configured at the factory can be used as the minimum value of the actual detection range, and the mean value of the normal fluctuation range and the maximum value of the alarm range configured at the factory can be used as the maximum value of the actual detection range.

[0038] After determining the actual detection range [Warn_lo, Warn_hi], this actual detection range can be used as a new discrimination criterion to detect the optical module data through this range. If the index data of the optical module is not within the actual detection range, it is determined that the initial detection result of the optical module is a fault.

[0039] It should be noted that the above method for determining the actual detection range is only exemplary, and is only used to illustrate that the actual detection range is determined based on the normal fluctuation range and the alarm range configured at the factory. In actual applications, the actual detection range can also be obtained by weighted averaging the normal fluctuation range and the alarm range configured at the factory. The present application does not limit this.

[0040] It can be seen that the actual detection range obtained by the above method combines the normal fluctuation range and the alarm range configured at the factory, and can more accurately judge the fault situation of the optical module.

[0041] In addition, it should be noted that since some index data, such as the number of mispackets at the input port and the number of FECs, do not have an alarm range configured at the factory, for these index data, the detection can be performed by setting an empirical value (Threshold). For example, if it is greater than the empirical threshold, it is considered that an anomaly has been detected.

[0042] (2) Trend discrimination detection rule The detection rule may include: if it is detected that the change of any index data of the optical module to be detected shows a monotonic trend within a specified time period, and the change amount of the index data within the specified time period is greater than the first threshold, it is determined that the detection result obtained by the detection rule for the index data of the optical module is a fault; otherwise, it is determined that the detection result obtained by the detection rule for the index data of the optical module is no fault.

[0043] In this embodiment, considering that the normal optical module index data fluctuates randomly, if the index data of the optical module shows an obvious monotonic change trend and reaches a certain change amount, it is often considered that there is a suspicion of a fault.

[0044] Specifically, please refer to Figure 2A , Figure 2A which is a schematic diagram of the statistical characteristics of the transmission power of a normal optical module provided by an embodiment of the present application.

[0045] As Figure 2A shown, Figure 2A the upper left image in it is the change trend of the mean of the transmission power within 16,000 samplings; Figure 2A the lower left image in it is the change trend of the standard deviation (std) of the transmission power within 16,000 samplings; Figure 2A the upper right image in it is the change trend of the maximum value (max) of the transmission power within 16,000 samplings; Figure 2A the lower right image in it is the change trend of the minimum value (min) of the transmission power within 16,000 samplings.

[0046] It should be noted that the sampling value for each sampling refers to the statistical characteristics of the transmission power within a specified time duration. For example, sampling is performed once every minute. Figure 2A If the ordinate of the data point corresponding to the abscissa 12000 in Figure 2A is 9900, it means that in the 12000th sampling, the average value of the transmission power of the optical module within the sampling time period (i.e., one minute from the previous sampling to this sampling) is 9900 milliwatts.

[0047] Through Figure 2A It can be seen from the four images included in Figure 2A that for the index data of a normal optical module, the changes in its mean, standard deviation, maximum value, and minimum value within a certain time range show a random state and there is no obvious monotonic trend.

[0048] Please refer to Figure 2B , Figure 2B which is a schematic diagram of the statistical characteristics of the transmission power of the faulty optical module provided by the embodiment of the present application.

[0049] As Figure 2B shown, Figure 2B In the upper left image in Figure 2B is the change trend of the mean of the transmission power within 16000 samplings; Figure 2B In the lower left image in Figure 2B is the change trend of the standard deviation of the transmission power within 16000 samplings; Figure 2B In the upper right image in Figure 2B is the change trend of the maximum value of the transmission power within 16000 samplings; Figure 2B In the lower right image in Figure 2B is the change trend of the minimum value of the transmission power within 16000 samplings.

[0050] Among Figure 2B the four images included, for the index data of the faulty optical module, the changes in its mean and maximum value within this time range show an obvious monotonic decreasing trend, and the decrease amount is relatively large, exceeding 15%.

[0051] Based on this, for the above example, the following trend discrimination detection rule can be set: If any index data shows a monotonic change trend within 30 days and the difference before and after exceeds 10%, it is considered that the initial detection result of the optical module under this detection rule is faulty.

[0052] It should be noted that this trend discrimination detection rule can detect the statistical characteristics (such as mean, standard deviation, maximum value, minimum value) of each index data, or directly detect each index data itself. At the same time, the specified time duration and the magnitude of the change can also be set according to the actual situation, and the present application does not limit this.

[0053] (3) Abnormal point quantity detection rule The detection rule may include: counting the abnormal points in the index data of the optical module that exceed the reference data range. If the number of abnormal points is greater than or equal to the second threshold, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is a fault. If the number of abnormal points is less than the second threshold, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is no fault; the reference data range is determined according to the statistical characteristics of the index data of the optical module.

[0054] In this embodiment, considering that when performing fault detection on the optical module through the reference data range of the set index data (that is, if any data in the index data within the specified time period exceeds the reference data range, it is determined that the optical module is abnormal), there will also be abnormal points in the normal optical module, and it cannot be guaranteed that every data point under each index data is within the reference data range. Determining that the optical module is abnormal directly through the reference data range will cause a large number of false detections. Since there will be abnormal points in the normal optical module, but the number is relatively small, while the number of abnormal points in the faulty optical module is relatively large. Therefore, a detection rule based on the number of abnormal points is proposed.

[0055] Specifically, abnormal detection and analysis can be performed on the index data of multiple normal optical modules and multiple faulty optical modules. For example, the 4-sigma method can be used for detection. This method assumes that the index data conforms to a normal distribution. For the index data of the optical module, the data within a specified time period, such as 30 days, is taken, and its mean value mean and standard deviation std are calculated to obtain the reference data range [mean - 4*std, mean + 4*std]. Then, the points outside the range [mean - 4*std, mean + 4*std] are considered abnormal points. Abnormal detection is performed on all types of index data of the optical module to determine the reference data range corresponding to the index data. Further, for each type of index data, the number of abnormal points is calculated, and the number of abnormal points of all types of index data of the optical module is summed up.

[0056] For the above example, abnormal detection is performed on the index data of the optical module for 30 days through the 4-sigma algorithm, and the sum of the abnormal points of each index data is calculated. If the number of abnormal points is greater than or equal to the second threshold (such as 50), it is considered that the initial detection result of the optical module under this detection rule is a fault.

[0057] (4) Model detection rule The detection rule may include: detecting each index data of the optical module according to the trained anomaly detection model to obtain the reconstruction error of each index data of the optical module; if the reconstruction error of any index data is greater than or equal to the third threshold, determining that the detection result obtained by the detection rule for the index data of the optical module is a fault; if the reconstruction errors of all index data are less than the third threshold, determining that the detection result obtained by the detection rule for the index data of the optical module is no fault.

[0058] In this embodiment, each index data of the optical module can be detected according to the trained anomaly detection model to obtain the reconstruction error of each index data of the optical module. This reconstruction error can be used to determine whether the index data matches the normal sample data during training. If the reconstruction error of an index data is high, the index data may be regarded as an outlier because normal data can be reconstructed well by the model, while abnormal data may not be reconstructed well, resulting in a large reconstruction error.

[0059] Specifically, in the actual scenario, since the number of faulty optical modules is very small, it is very difficult to collect a large amount of data of faulty optical modules. It is necessary to monitor a large-scale network containing a large number of optical modules for a long time. Therefore, a detection method that only uses normal data needs to be designed.

[0060] As an embodiment, the anomaly detection model can be a classification model based on the VAE (Variational Autoencoder) algorithm. VAE is a generative model used to learn the probability distribution of training data, so as to generate new samples similar to the training data. VAE combines the characteristics of generative models and autoencoders and is an unsupervised learning method.

[0061] In anomaly detection, VAE mainly learns the distribution of normal data. Therefore, for normal samples, VAE can reconstruct well, while for abnormal samples, since they do not conform to the distribution learned during training, the reconstruction effect is poor, so the reconstruction error will be large. By setting the threshold of the reconstruction error, samples can be classified as normal or abnormal. If the reconstruction error of a certain sample exceeds the predetermined threshold, the sample may be abnormal. Similar to the application of anomaly detection, the reconstruction error can be used to distinguish between two types of samples. If the reconstruction error of one category is significantly higher than that of another category, the reconstruction error can be used as a feature for binary classification. By setting an appropriate threshold, the category can be directly determined based on the reconstruction error.

[0062] In addition, in this embodiment, models such as Isolation Forest, Local Outlier Factor (LOF), and K-means clustering can also be used as anomaly detection models, and this application does not limit this.

[0063] It can be seen that in the method of fault detection through the model detection rule, the training of the model only needs to use the data of normal modules, thus avoiding the problem of scarce fault data.

[0064] So far, the description of several representative detection rules is completed.

[0065] After obtaining the current detection results of each sample optical module, it can be determined whether the current detection results of each sample optical module match the sample labels of each sample optical module.

[0066] So far, the description of step 102 is ended, and then step 103 is executed.

[0067] Step 103, if the current detection results of the N sample optical modules do not match the sample labels and N is greater than the set threshold, then sequentially adjust the current weights of each detection rule one by one. After adjusting the current weight of each detection rule each time, determine the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module using each detection rule.

[0068] In this embodiment, if it is found that the current detection results of the N sample optical modules do not match the sample labels and N is greater than the set threshold, it indicates that the current weight settings of each detection rule are inappropriate and a good detection effect cannot be achieved. At this time, the current weights of each detection rule need to be adjusted.

[0069] It should be noted that the accuracy of anomaly detection and the degree of computing resource occupancy can be balanced by adjusting the above-set threshold. It is easy to understand that the smaller the set threshold, the fewer the current detection results that do not match the sample labels, and it will be considered that the current weight settings of each detection rule are inappropriate, and the current weights of each detection rule will be further adjusted. Such a set threshold can make the accuracy of anomaly detection relatively high, but the computing resource occupancy is also relatively large, and the determination speed of the target weight is slower; while the larger the set threshold, the more current detection results that do not match the sample labels are required to consider that the current weight settings of each detection rule are inappropriate, and the current weights of each detection rule will be further adjusted. Such a set threshold will result in a relatively low accuracy of anomaly detection (with a large error), but the computing resource occupancy is also smaller, and the determination speed of the target weight is faster. Therefore, the above-set threshold can be adjusted according to the actual accuracy requirements of anomaly detection to balance the accuracy of anomaly detection and the degree of computing resource occupancy.

[0070] As an embodiment, the method for adjusting the current weight of any detection rule may include: increasing the current weight of the detection rule or decreasing the current weight of the detection rule.

[0071] Specifically, for the method of increasing the current weight w i , it can be set that w i = w i + w i / 4, where w i refers to the current weight of the i-th detection rule, that is, increasing by 1 / 4 on the basis of the current weight. This is only an exemplary method. It is also possible to increase the current weight by a specified value or multiply the current weight by a specified coefficient to obtain the increased current weight. The present application does not limit this.

[0072] Similarly, for the method of decreasing the current weight w i , it can be set that w i = w i - w i / 4, where w i refers to the current weight of the i-th detection rule, that is, decreasing by 1 / 4 on the basis of the current weight. This is only an exemplary method. It is also possible to decrease the current weight by a specified value or multiply the current weight by a specified coefficient to obtain the decreased current weight. The present application does not limit this.

[0073] After adjusting the current weight of a detection rule, the reference detection results of each sample optical module can be determined based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule.

[0074] So far, the description of step 103 ends. After completing one adjustment of the current weight, the reference detection results of each sample optical module can be determined again based on the adjusted weight to determine whether the adjustment of the current weight is effective, that is, to determine whether the reference detection results of each sample optical module match the sample labels of each sample optical module. If it is found that the reference detection results of P sample optical modules do not match the sample labels, and P is greater than or equal to N, then step 104 is executed; if it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N, and M is greater than the set threshold, then step 105 is executed.

[0075] Step 104, if it is found that the reference detection results of P sample optical modules do not match the sample labels, and P is greater than or equal to N, then readjust the current weight of this detection rule, and return to the step of determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule.

[0076] In this embodiment, after adjusting the current weight of a certain detection rule through step 103 to obtain the reference detection results of each sample optical module, the reference detection results of each sample optical module can be compared with the sample labels. If it is found that the reference detection results of P sample optical modules do not match the sample labels, and P is greater than or equal to N, that is, the number of sample optical modules that do not match has not decreased, it indicates that the adjusted current weight has not achieved a good effect. At this time, the previous adjustment of the current weight of this detection rule can be abandoned, and the current weight before the adjustment of this detection rule can be readjusted.

[0077] As an embodiment, if the current weight of this detection rule was previously adjusted in the way of increasing the current weight, then the current weight of this detection rule is readjusted in the way of decreasing the current weight; if the current weight of this detection rule was previously adjusted in the way of decreasing the current weight, then the current weight before the adjustment of this detection rule is readjusted in the way of increasing the current weight.

[0078] For example, the way to adjust the current weight of this detection rule in step 103 is w i =w i + w i / 4, then when readjusting the current weight of this detection rule, it can be adjusted according to w i =w i - w i / 4.

[0079] After completing the readjustment of the current weight of the detection rule, the step of determining the reference detection result of each sample optical module based on the current weight of each detection rule and the detection result obtained by performing fault detection on the sample data of each sample optical module in step 103 can be returned to re-determine the effect of the current weight adjustment this time.

[0080] So far, the description of step 104 ends.

[0081] Step 105, if it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N and M is greater than the set threshold, then when this detection rule is not the last detection rule, sequentially adjust the current weight of the next detection rule, and when this detection rule is the last detection rule, then determine the current weights of each detection rule as the target weights of each detection rule, so as to determine the fault detection result of the optical module to be detected based on the target weights of each detection rule and the detection results obtained by performing fault detection on the optical module to be detected by each detection rule.

[0082] In this embodiment, after adjusting the current weight of any detection rule through step 103 to obtain the reference detection results of each sample optical module, the reference detection results of each sample optical module can be compared with the sample labels. If it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N and M is greater than the set threshold, that is, the number of sample optical modules with mismatches decreases and the number of successfully matched sample optical modules increases, but it has not reached the error range, it indicates that the effect of the adjusted current weight is better than that of the pre-adjusted current weight. At this time, the current weight of the next detection rule can be further adjusted sequentially, and the adjustment method has been detailed above and will not be repeated here.

[0083] When the adjusted current weight is the current weight of the last detection rule, the current weights of each detection rule can be determined as the target weights of each detection rule to perform fault detection on the optical module through the target weights of each detection rule.

[0084] As an embodiment, determining the current weights of each detection rule as the target weights of each detection rule may specifically include: Determining the target detection result of each sample optical module based on the current weight of each detection rule and the detection result obtained by performing fault detection on the sample data of each sample optical module by each detection rule; If it is found that the target detection results of L sample optical modules do not match the sample labels, and L is greater than the set threshold, then return the step of sequentially adjusting the current weights of each detection rule one by one, and after adjusting the current weight of each detection rule each time, determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module using each detection rule; if it is found that the target detection results of L sample optical modules do not match the sample labels, and L is less than or equal to the set threshold, then use the current weights of each detection rule as the target weights of each detection rule.

[0085] In this embodiment, when the adjusted current weight is the current weight of the last detection rule, the target detection results of each sample optical module can be determined based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module using each detection rule.

[0086] If the target detection results of L sample optical modules do not match the sample labels, and L is greater than the set threshold, it indicates that the adjustment of the current weight in this round (regarding adjusting the current weight once for each detection rule as one round) still does not achieve the required effect. At this time, the step of sequentially adjusting the current weights of each detection rule one by one can be returned, and after adjusting the current weight of each detection rule each time, determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module using each detection rule, so as to perform the next round of adjustment.

[0087] If the target detection results of L sample optical modules do not match the sample labels, and L is less than or equal to the set threshold, it indicates that the adjustment of the current weight in this round has achieved the required effect. At this time, the current weights of each detection rule can be used as the target weights of each detection rule to perform fault detection on the optical module through the target weights of each detection rule.

[0088] As an embodiment, if the current detection results of N sample optical modules do not match the sample labels, and N is less than or equal to the set threshold, then use the current weights of each detection rule as the target weights of each detection rule; if it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N, and M is less than or equal to the set threshold, use the current weights of each detection rule as the target weights of each detection rule.

[0089] In this embodiment, if it is found that N is less than or equal to the set threshold or M is less than or equal to the set threshold, it indicates that the current detection results are already within the allowable error range set in advance. At this time, the current weights of each detection rule can be directly used as the target weights of each detection rule.

[0090] Thus, the description of step 105 ends.

[0091] After obtaining the target weights of each detection rule through steps 101 to 105, the optical module to be detected can be abnormally detected according to the target weights, that is, based on the target weights of each detection rule and the detection results obtained by performing fault detection on the optical module to be detected by each detection rule, the fault detection result of the optical module to be detected is determined.

[0092] Specifically, please refer to Figure 3 , Figure 3 which is an overall application schematic diagram of the optical module fault detection method provided in the embodiments of the present application.

[0093] As Figure 3 shown, first, the optical module to be detected and each index data of the optical module within a specified time period can be determined according to the actual detection requirements.

[0094] In this embodiment, the optical module can be detected based on the obtained index sequence. Specifically, since the fault condition of the optical module is detected based on each index data included in a specified time period (such as 30 days), the above-mentioned specified time period can be determined first according to the actual detection requirements.

[0095] For example, if it is necessary to detect the fault condition of the optical module at the current moment, the time period including 30 days before the current date can be determined as the specified time period, and each index data within the specified time period is obtained to perform fault detection on the optical module.

[0096] For another example, if it is necessary to predict the fault condition of the optical module 10 days later, the time period including 30 days before the cut-off date with the date 10 days later as the cut-off date can be determined as the specified time period, that is, 20 days before the current date and the next 10 days after the current date. It should be noted that the index data of the optical module for the 20 days before the current date can be directly obtained from the recorded data, and the index data for the next 10 days after the current date needs to be predicted according to the index data before the current date. The specific prediction method can use time series prediction methods such as through LSTM (Long Short-Term Memory), Random Forest, Support Vector Regression (SVR), etc. The present application does not limit this.

[0097] It is easy to understand that the specified duration can be set according to actual needs. The longer the specified time period is set, the more index data can be obtained for prediction during detection, but it will also be more time-consuming.

[0098] In this embodiment, the index data for abnormal detection in the optical module to be detected (Figure 3 After the index sequence in Figure 3 Detection Rule 1, Detection Rule 2, Detection Rule n, etc. in Figure 3 The detection results can be used to perform anomaly detection on the index data of the optical module. For each detection rule, the initial detection result corresponding to the detection rule can be obtained (

[0099] Detection Result 1, Detection Result 2, Detection Result n, etc. in Figure 3 Each initial detection result can be identified by 0 (normal) or 1 (abnormal). Further, based on the determined target weights of the detection rules (

[0100] Weight w1, Weight w2, Weight wn, etc. in Figure 3 ), a weighted operation can be performed on each initial detection result to obtain a weighted operation result, and the anomaly detection result can be determined based on the weighted operation result.

[0101] Up to this point, the description of the optical module fault detection method in

[0102] For flexibility, in this embodiment, it is not required that the sum of all target weights is equal to 1, nor is it required that the weighted operation result must be less than or equal to 1.

[0103] In addition, in this embodiment, two judgment thresholds can be specified, such as 0.2 and 0.7. If it is determined that the weighted operation result is less than or equal to 0.2, it indicates that the optical module is a normal optical module and there is no need to repair or adjust the optical module; if it is determined that the weighted operation result is greater than 0.2 and less than 0.7, it indicates that the optical module is a sub - healthy optical module, which means that there may be some risks of failure. At this time, relevant personnel can be prompted to further check the optical module to detect and eliminate potential failure hazards in advance; if it is determined that the weighted operation result is greater than 0.7, it indicates that the optical module is a faulty optical module and needs to be repaired or replaced.

[0104] It should be noted that when the detection rules do not change, after obtaining the target weights corresponding to the current detection rules by executing the above optical module fault detection method only once, when detecting the optical module to be detected subsequently, the determined target weights can be directly used without re - determining the target weights.

[0105] Up to this point, the description ofFigure 1 Description.

[0106] When the fault detection results of the sample data of N sample optical modules in this application do not match the sample labels of each sample optical module, the current weights of each detection rule are adjusted one by one in sequence. After adjusting the current weight of each detection rule each time, the reference detection results of each sample optical module are re-determined according to the adjusted current weight. If the number of sample optical modules with mismatches is not less than N, it indicates that the current weight adjustment is inappropriate, and then the current weight of this detection rule is adjusted again; if the number of sample optical modules with mismatches is less than N, it indicates that the current weight adjustment is appropriate, and then the current weight of the next detection rule is adjusted until the target weight is obtained; through the adaptive adjustment of the corresponding weights of each detection rule, the obtained target weight can more accurately detect faults in the optical module.

[0107] Please refer to Figure 4 , Figure 4 which is a structural diagram of an optical module fault detection device proposed in an embodiment of this application. As Figure 4 shown, the device may include a sample acquisition unit 401, a result judgment unit 402, a first determination unit 403, and a second determination unit 404. Specifically, the device includes: The sample acquisition unit 401 is configured to acquire sample data of multiple sample optical modules; The result judgment unit 402 is configured to judge whether the current detection results of each sample optical module match the sample labels of each sample optical module; the current detection result of any sample optical module is determined based on the current weights of at least two detection rules and the detection results obtained by performing fault detection on the sample data of this sample optical module by each detection rule; The first determination unit 403 is configured to, if the current detection results of N sample optical modules do not match the sample labels and N is greater than a set threshold, adjust the current weights of each detection rule one by one in sequence, and after adjusting the current weight of each detection rule each time, determine the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; The second determination unit 404 is configured to, if it is found that the reference detection results of P sample optical modules do not match the sample labels, where P is greater than or equal to N, readjust the current weight of the current detection rule, and return the step of determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; if it is found that the reference detection results of M sample optical modules do not match the sample labels, where M is less than N and M is greater than a set threshold, when the detection rule is not the last detection rule, sequentially adjust the current weight of the next detection rule, and when the detection rule is the last detection rule, determine the current weights of each detection rule as the target weights of each detection rule, so as to determine the fault detection result of the to-be-detected optical module based on the target weights of each detection rule and the detection results obtained by performing fault detection on the to-be-detected optical module by each detection rule.

[0108] Optionally, the second determination unit 404 is specifically configured to: Determine the target detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; If it is found that the target detection results of L sample optical modules do not match the sample labels, where L is greater than the set threshold, return the step of sequentially adjusting the current weights of each detection rule one by one, and after adjusting the current weight of each detection rule each time, determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; if it is found that the target detection results of L sample optical modules do not match the sample labels, where L is less than or equal to the set threshold, use the current weights of each detection rule as the target weights of each detection rule.

[0109] Optionally, the second determination unit 404 is further configured to: If the current detection results of N sample optical modules do not match the sample labels and N is less than or equal to the set threshold, use the current weights of each detection rule as the target weights of each detection rule; If it is found that the reference detection results of M sample optical modules do not match the sample labels, where M is less than N and M is less than or equal to the set threshold, use the current weights of each detection rule as the target weights of each detection rule.

[0110] Optionally, the detection rule includes a threshold discrimination detection rule; the threshold discrimination detection rule includes: Perform fault detection on the indicator data of the optical module according to the determined actual detection range; if the indicator data of the optical module is not within the actual detection range, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is a fault; if the indicator data of the optical module is not within the actual detection range, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is no fault; Among them, the actual detection range is determined according to the following method: According to the statistical characteristics of the indicator data of various types actually detected in the normal optical module, determine the normal fluctuation range of the indicator data of various types; According to the normal fluctuation range and the alarm range configured at the factory for the indicator data of various types, determine the actual detection range of the indicator data of various types.

[0111] Optionally, the detection rule includes a trend discrimination detection rule; the trend discrimination detection rule includes: If it is detected that the change of any indicator data of the optical module to be detected shows a monotonic trend within the specified time period, and the change amount of the indicator data within the specified time period is greater than the first threshold, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is a fault; otherwise, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is no fault.

[0112] Optionally, the detection rule includes an outlier quantity detection rule; the outlier quantity detection rule includes: Count the outliers in the indicator data of the optical module that exceed the reference data range. If the number of outliers is greater than or equal to the second threshold, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is a fault. If the number of outliers is less than the second threshold, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is no fault; the reference data range is determined according to the statistical characteristics of the indicator data of the optical module.

[0113] Optionally, the detection rule includes a model detection rule; the model detection rule includes: Detect the indicator data of the optical module according to the trained anomaly detection model to obtain the reconstruction error of the indicator data of the optical module; If the reconstruction error of any indicator data is greater than or equal to the third threshold, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is a fault; if the reconstruction errors of all indicator data are less than the third threshold, determine that the detection result obtained by performing fault detection on the indicator data of the optical module by this detection rule is no fault.

[0114] Optionally, the multiple sample optical modules include at least one normal optical module and at least one faulty optical module.

[0115] Optionally, adjusting the current weight of each detection rule includes increasing the current weight or decreasing the current weight; the second determination unit 404 is specifically configured to: If the current weight of this detection rule was previously adjusted in a manner of increasing the current weight, then readjust the current weight of this detection rule in a manner of decreasing the current weight; If the current weight of this detection rule was previously adjusted in a manner of decreasing the current weight, then readjust the current weight of this detection rule in a manner of increasing the current weight.

[0116] Thus far, the description of the optical module fault detection device in Figure 4 is completed.

[0117] The embodiments of the present application further provide Figure 4 a description of the hardware structure of the device shown. This hardware structure is Figure 5 the structure in the electronic device shown in Figure 5 , Figure 5 which is the structural diagram of the electronic device provided by the embodiments of the present application. As shown in Figure 5 , this hardware structure may include: a processor and a machine-readable storage medium, where the machine-readable storage medium stores machine-executable instructions that can be executed by the processor; the processor is used to execute the machine-executable instructions to implement the method disclosed in the above examples of the present application.

[0118] Based on the same inventive concept as the above method, the embodiments of the present application further provide a machine-readable storage medium, on which several computer instructions are stored. When the computer instructions are executed by a processor, the method disclosed in the above examples of the present application can be implemented.

[0119] Exemplarily, the above machine-readable storage medium can be any electronic, magnetic, optical or other physical storage device that can contain or store information, such as executable instructions, data, etc. For example, the machine-readable storage medium can be: RAM (Random Access Memory), volatile memory, non-volatile memory, flash memory, storage drives (such as hard disk drives), solid-state drives, any type of storage disk (such as optical discs, DVDs, etc.), or similar storage media, or a combination thereof.

[0120] The above are only the embodiments of the present application and are not used to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for detecting optical module faults, characterized in that, The method includes: Obtaining sample data of multiple sample optical modules; Determining whether the current detection result of each sample optical module matches the sample label of each sample optical module; the current detection result of any sample optical module is determined based on the current weights of at least two detection rules and the detection results obtained by performing fault detection on the sample data of the sample optical module by each detection rule; If the current detection results of N sample optical modules do not match the sample labels and N is greater than a set threshold, then sequentially adjust the current weights of each detection rule one by one, and after adjusting the current weight of each detection rule each time, determine the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; If it is found that the reference detection results of P sample optical modules do not match the sample labels and P is greater than or equal to N, then readjust the current weight of this detection rule, and return to the step of determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; If it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N, and M is greater than the set threshold, then when this detection rule is not the last detection rule, sequentially adjust the current weight of the next detection rule, and when this detection rule is the last detection rule, then determine the current weights of each detection rule as the target weights of each detection rule, so as to determine the fault detection result of the optical module to be detected based on the target weights of each detection rule and the detection results obtained by performing fault detection on the optical module to be detected by each detection rule.

2. The method according to claim 1, characterized in that, The determining the current weights of each detection rule as the target weights of each detection rule includes: Determining the target detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; If it is found that the target detection results of L sample optical modules do not match the sample labels and L is greater than the set threshold, then return to the step of sequentially adjusting the current weights of each detection rule one by one, and after adjusting the current weight of each detection rule each time, determine the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; if it is found that the target detection results of L sample optical modules do not match the sample labels and L is less than or equal to the set threshold, then use the current weights of each detection rule as the target weights of each detection rule.

3. The method according to claim 1, characterized in that The method further includes: If the current detection results of N sample optical modules do not match the sample labels and N is less than or equal to the set threshold, then use the current weights of each detection rule as the target weights of each detection rule; If it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N, and M is less than or equal to the set threshold, use the current weights of each detection rule as the target weights of each detection rule.

4. The method according to claim 1, wherein The detection rules include threshold discrimination detection rules; the threshold discrimination detection rules include: Perform fault detection on the index data of the optical module according to the determined actual detection range; if the index data of the optical module is not within the actual detection range, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is a fault; if the index data of the optical module is not within the actual detection range, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is no fault; Among them, the actual detection range is determined according to the following method: Determine the normal fluctuation range of each type of index data according to the statistical characteristics of the actually detected index data in the normal optical module; Determine the actual detection range of each type of index data according to the normal fluctuation range and the alarm range configured at the factory for each type of index data.

5. The method according to claim 1, wherein The detection rules include trend discrimination detection rules; the trend discrimination detection rules include: If it is detected that the change of any index data of the optical module to be detected shows a monotonic trend within the specified time period, and the change amount of this index data within the specified time period is greater than the first threshold, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is a fault; otherwise, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is no fault.

6. The method according to claim 1, characterized in that, The detection rules include abnormal point quantity detection rules; the abnormal point quantity detection rules include: Count the abnormal points exceeding the reference data range in each index data of the optical module. If the number of abnormal points is greater than or equal to the second threshold, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is a fault. If the number of abnormal points is less than the second threshold, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is no fault; the reference data range is determined according to the statistical characteristics of each index data of the optical module.

7. The method according to claim 1, characterized in that, The detection rules include model detection rules; the model detection rules include: Detect each index data of the optical module according to the trained anomaly detection model to obtain the reconstruction error of each index data of the optical module; If the reconstruction error of any index data is greater than or equal to the third threshold, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is a fault; if the reconstruction errors of all index data are less than the third threshold, it is determined that the detection result obtained by performing fault detection on the index data of the optical module by this detection rule is no fault.

8. The method according to claim 1, wherein Adjusting the current weight of each detection rule includes increasing the current weight or decreasing the current weight; Re-adjusting the current weight of this detection rule includes: If the current weight of this detection rule was previously adjusted in the way of increasing the current weight, then re-adjust the current weight of this detection rule in the way of decreasing the current weight; If the current weight of the detection rule was previously adjusted in a way that decreases the current weight, then readjust the current weight of the detection rule in a way that increases the current weight.

9. An optical module fault detection device, characterized in that, The device includes: A sample acquisition unit, configured to acquire sample data of a plurality of sample optical modules; A result determination unit, configured to determine whether the current detection result of each sample optical module matches the sample label of each sample optical module; the current detection result of any sample optical module is determined based on the current weights of at least two detection rules and the detection results obtained by performing fault detection on the sample data of the sample optical module by each detection rule; A first determination unit, configured to, if the current detection results of N sample optical modules do not match the sample labels and N is greater than a set threshold, sequentially adjust the current weights of each detection rule one by one, and after adjusting the current weight of each detection rule each time, determine the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; A second determination unit, configured to, if it is found that the reference detection results of P sample optical modules do not match the sample labels and P is greater than or equal to N, readjust the current weights of the detection rules, and return to the step of determining the reference detection results of each sample optical module based on the current weights of each detection rule and the detection results obtained by performing fault detection on the sample data of each sample optical module by each detection rule; if it is found that the reference detection results of M sample optical modules do not match the sample labels, M is less than N, and M is greater than the set threshold, then when the detection rule is not the last detection rule, sequentially adjust the current weight of the next detection rule, and when the detection rule is the last detection rule, determine the current weights of each detection rule as the target weights of each detection rule, so as to determine the fault detection result of the optical module to be detected based on the target weights of each detection rule and the detection results obtained by performing fault detection on the optical module to be detected by each detection rule.

10. An electronic device, characterized in that, It includes: A processor and a machine-readable storage medium, where the machine-readable storage medium stores machine-executable instructions that can be executed by the processor; The processor is configured to execute the machine-executable instructions to implement the method according to any one of claims 1 to 8.

Images