MXPA97004636A - Ave inspection - Google Patents

Abstract

The present invention relates to a method for inspecting a plurality of elements of a telecommunication system for faults and generating alarms in response thereto, the method comprising the steps of inspecting the elements for the occurrences of faults, measuring the duration of each occurrence of failure and, for each occurrence of failure, if the duration exceeds a predetermined value, activate an alarm indicator, if the duration does not exceed the predetermined value, increase a stored value

Description

INSPECTION OF FAILURE DESCRIPTION OF THE INVENTION This invention relates to the inspection of faults in a telecommunications system. Telecommunications networks have a large number of components distributed over a wide area, and it is important to be able to identify faults when they occur and treat them quickly before they manifest themselves to the users of the network as a service interruption. The network user, in turn, can have a service level agreement with the network operator, which specifies the contractual penalties if the service interruptions exceed a predetermined limit. Many methods are known to identify faults and alert operators thereof so that remedial action can be presented. Included in the term "breakdown" for the purpose of this specification, not only are the equipment or service interruptions found, but also cases such as overloads, which are due to external causes but require immediate remedial action. Different users of a system need notification in different circumstances. A telecommunications network operator, responsible for the inspection of the equipment, needs to know the faults of the individual equipment. However, it is possible for an item of equipment in the telecommunications network to fail without any immediate effect on the level of service provided to the customer, if, for example, alternative routes are available, or if the consumer is not using the full capacity of the customer. system. Conversely, a customer may experience a service interruption when no item of equipment has failed, for example, if the total demand for system use by all customers exceeds the capacity of the system. In order to inspect the service provided to the customers, the network operator also needs to be informed of such service interruptions. For example, in a telecommunications network, if a channel is running at full capacity, any of the call attempts on that channel could fail. If the rate of occurrence of such call failures increases, this indicates that the network is at full capacity for an increasing proportion of time. Remedial action may be possible, for example, by reassigning the capacity of any part (for example, a second channel in a time division multiple access system). In the case of service interruptions for the customer, they can be registered. The nature according to the level of service will determine what information is required, but in most cases, a guaranteed minimum level is agreed. The minimum level can be an agreed proportion of time in which the service is available. In other circumstances, for example, when reset procedures are complex, the number of interruptions separately can be used as an additional or alternative criterion. There is a difficulty in the inspection of faults, the phenomenon that can be of a temporary nature can be indicative of the presence of an imminent real or underlying problem, but rather no more than unimportant statistical fluctuations. A transient fault, which in itself seems transparent, may mean that the cause has ceased or may mean that there is an underlying problem and that the fault will recur under certain conditions. A particular special case of this problem occurs when an alarm is triggered through a system property exceeding a predetermined threshold value. If the threshold value is set too high, the alarm indicator will not be activated until the situation is already critical, leaving no margin where remedial action can be taken. However, if the threshold is set too low there will be an excessive number of false alarm indications. The importance of temporary faults may differ according to the individual requirements of the customer. For a customer whose terminal equipment includes error correction installations, intermittent, temporary faults may be important. However, for a customer whose interruption, however short, needs a restart operation involving complex security procedures or a site visit to reset the computer, a large number of temporary interruptions are much more bad than a single long-term interruption . The passenger faults have to be handled differently from non-fleeting faults. The passenger faults are canceled by themselves, so you do not need to take any remedial action to cancel them manually. However, because they cancel out on their own, it is difficult to investigate the cause of a transient fault, or to identify patterns that may indicate an underlying problem. A prior art fault inspection system is known from IBM Technical Disclosure Bulletin No. 7 (December 1962). Every second, this system registers if a fault is present or not, detecting transmission errors. An alert is sent if the number X of individual seconds, where a fault is detected in a period of 15 minutes, exceeds a first threshold L, or if in a period of 24 hours the number Z of individual seconds, where it is detected a breakdown, exceeds a second threshold which and -H is, in proportion to the size of the measurement period, much less than L. This arrangement allows statistically significant changes in the fault regime to be detected for both long and short periods, allowing rapid changes to be detected. sudden large, and also detecting smaller long-term changes which, due to the larger sample size, are statistically significant, without false alerts caused by normal short-term statistical fluctuations. This prior art system only inspects the number of seconds where a fault is present. It does not take into account the duration of individual breakdowns. For example, a report of nine "seconds of error" in a period of 15 minutes can be caused by nine individual failures, each less than a second, or an individual failure of 9 seconds. Nor is provision made for separate faults that occur simultaneously, or of overlapping duration; only a maximum of one fault per second can be recorded. According to a first aspect of the invention, there is provided a method for inspecting a plurality of elements of a telecommunications system for faults, and for generating alarms in response thereto, such a method comprising the steps of verifying elements for occurrences of breakdowns, measure the duration of each failure occurrence and, for each occurrence of failure; if the duration exceeds a predetermined value, activate an alarm indicator, if the duration does not exceed the predetermined value, increase a stored value. According to a second aspect, an apparatus is provided for inspecting a plurality of elements of a telecommunications system for faults, comprising an alarm indicator, detection means associated with each element for detecting occurrences of faults, means for controlling times for measuring the duration of each fault detected by any of the detection means, activation means for activating the alarm if the duration measured by the time control means exceeds a predetermined value, counting means for storing an account number, and increasing means for increasing the account number stored in the counting means if the duration of a fault measured by the time control means does not exceed the predetermined value. In this way, a fault, which does not cancel spontaneously within a predetermined period and must be triggered, can be easily distinguished from reports of passenger faults, which can be stored for subsequent analysis, without the operator of the system having to be alerted for each individual passenger fault. Preferably, the start and override times of faults are recorded, and after the start time of a fault has been recorded, a delay period is initiated, and if the breakdown of the fault is recorded before the delay period ends, the stored value is increased, and if the delay period ends before the cancellation of the fault has been recorded, the alarm indicator is activated. Preferably, the system is also inspected for the occurrences of system disability, or a system function, to a user. In a preferred arrangement, an alarm indicator is activated if the stored value reaches a value equal to or greater than a threshold value within a predetermined range. This additional aspect is the subject of the PCT co-pending application filed on the same day of this case, and claiming the same priority, Agent Reference A25113. This arrangement measures the frequency of occurrence of actual faults, rather, as discussed in the previous provision, the proportion of time that a fault or breakdowns are present. Generating alarm indications in response to the frequency of occurrence of individual faults, a better indication is available for the system operator so that if the fault requires attention if the transient occurrences are reported directly to the operator. For a user, the system can be used to identify the number of times the service is not available. Preferably, the method comprises the steps of: establishing a period of analysis; inspect the system continuously for occurrences of faults; at the end of a scan interval shorter than the analysis period, count the number of occurrences of a fault during the analysis period, which ends at the end of the scan interval; and activate an alarm indicator if the number of occurrences of the fault in the analysis period is equal to or greater than a threshold value. Preferably, the start times and the cancellation of faults are recorded, and at the end of each scan interval, faults, for which a time longer than the analysis period has elapsed, since the cancellation time has been exceeded. registered, are not counted. Preferably, also the alarm indicator is kept in its real state if the number of occurrences of the fault is between the first and second threshold values.

In a preferred arrangement, the alarm indicator is deactivated if the number of occurrences of the fault in the analysis period is equal to or less than a second threshold value, the second threshold value being less than the first threshold value. This alarm indicator may be the same as that which is activated if the duration of a fault exceeds a predetermined value. It can be arranged to remain activated whenever any of the stored values exceeds the second threshold or a fault of duration greater than the predetermined value, remains without cancellation. The alarm may have different activation states depending on whether the stored value remains above its threshold, or a long-term fault remains un-canceled, or both. In another arrangement, an additional alarm indicator is also activated if the number of occurrences of the fault in the analysis period is equal to or greater than a threshold value, and the additional alarm indicator remains activated until it is recognized by an operator. The counting means may comprise a memory, arranged to store the number of occurrences of the fault condition within each of a plurality of scanning intervals, whose total duration is that of the period of analysis. The durations of the predetermined analysis period and the scanning interval can be selectable. Preferably, the equipment includes an update processor arranged, at the end of each scan interval, to output the data stored in the memory and to supply it to the counting means, and to instruct the memory to delete the data in relation to the data. to the oldest scan interval for which the data is stored. Preferably, discrimination means are provided for distinguishing occurrences of the fault condition having a first predetermined characteristic from those having a second predetermined characteristic, the counting means being arranged to count those occurrences that have each characteristic separately, or to count only one of the types. The means of counting can also be arranged to be canceled, thus suspending its operation. This allows routine tests and interruptions pre-arranged in the service to be disconnected from the analysis, to avoid the generation of false alarms. An embodiment of the invention will now be described by way of example only, with reference to the drawings, in which: Figure 1 is a diagrammatic representation of a sequence of faults occurring in a telecommunications system; Figure 2 is a functional block diagram showing the various components of an apparatus according to the invention for inspecting a telecommunications network for intermittent fault conditions, together with elements of the network to be inspected; and Figure 3 shows the general construction of a computer suitable for the operation of the invention. Figure 1 shows a time sequence of passenger faults of a particular type occurring in a telecommunications system, which will be inspected through an apparatus modeled by the invention. For example, the system can be a telecommunications network and faults can be overloads in a communication channel, and interruptions in the service to a specific client. In the inspection apparatus, which will be described below with reference to Figure 2, the time is divided into a number of scanning intervals t-j_, t2, etc., each of length t. In this mode, the scanning interval is substantially longer than the duration of the faults, and it is possible for more than one fault to occur in the same scanning interval. An analysis period T is defined. In this illustrative example, the analysis period T is three times the length of the scanning interval t. Nevertheless, during the practice, the period of analysis can be much longer than this one. In this invention, distinction is made between passenger faults of longer duration. When a fault is identified, a period of delay begins. The length of the delay period can be configurable for different clients and for different services. If the fault is not canceled during the delay period, an alarm indicator is activated. This alarm indicator can be the same as activated if it exceeds the threshold number, but preferably the alarms are different between them, or the alarm has different activation states, so that long interruptions and a series of short interruptions can be distinguished , allowing the operator to prioritize their actions in response to alarms. However, if the fault is not canceled during the period of delay, the fault is classified as a passing alarm and a passing service level fault mark for the actual scan interval is increased by one. If the predetermined fault occurs and is canceled more than once in the same scan interval, the final transient fault rating for the scan interval will be the number of times the fault occurs in that interval. A separate alarm can be provided, which remains activated after a transient fault has been reported, until the alarm is canceled being recognized by the operator. In this example, temporary faults A, B, C occur, D, E, in exploration intervals t2, t4 (double), tg and ty respectively. At the end of each scan interval, the device counts the number of faults that have occurred during the analysis period T, in this example comprising the last three scan intervals, and an alarm indicator is activated if the number reaches or exceeds a threshold value, in this example is set to 3. Only when the number of the analysis period falls or is below a second threshold value, in this example 1, the alarm indicator is canceled. Different thresholds are selected to avoid the occurrence of intermittent alarms, which could otherwise occur when the transient fault occurrence regime is close to an individual threshold value. For example, at the end of the scan interval t ^, only one fault is counted, A, in the analysis period t1 + t2 + t3. At the end of the next scan interval t ^, three faults are counted, A, B, C in the analysis period t2 + t3 + t4, and the alarm indicator is activated. At the end of the next scan interval t5, only two faults are counted, since fault A is now outside the analysis period t3 + t4 + t5- Although two faults are below the alarm threshold value of 3, the alarm it is not canceled since the number of faults in the analysis period has not fallen to a second threshold value. At the end of the scan interval tg, three faults reappear; B, C, D, within the analysis period. At the end of the scan interval t7, only two faults are found, D and E, because although fault E is added (in the scan interval?), Faults B and C are now outside the analysis period t5 + tg + t ?. However, the alarm indicator again remains on, because the "cancellation" threshold has not been passed. At the end of the period tg, only the fault E is within the analysis period t7 + tg + tg, and thus the number of faults has fallen towards the threshold and the alarm indicator is canceled. However, when fault "E" has been followed in the period tg or tg by a group of one or more additional faults, the second threshold could have been reached and the alarm indicator could have not been canceled. It should be noted that the fault E, although it starts in the period tg, is counted as being in period t7, the period in which it is canceled. This avoids a double counting of the fault E. The cancellation time is used since at the end of the scanning interval tg, the duration of the fault is indeterminate, and it can be a long-term fault (see fault F in scanning intervals). tg, t10). Referring to Figure 2, a functional block diagram of the functional components of an apparatus 20 for inspecting faults in a telecommunications network 21 is shown, the network includes elements 2 to 5. These elements can be functional components of the network, or they can be elements of the service provided by the network to a client. The apparatus 20 is in a conventional construction computer, as shown in Figure 3, and comprising a memory 220, a display 222 and a board 224, a central processing unit 226 and an interface 228. The memory 220 can be implemented as a combination of a hard disk, a random access memory (RAM) and a Read Only Memory (ROM). The computer has a program stored in its memory 220 and the program includes a group of program modules corresponding to the functional components 1, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 and 16 shown in Figure 2. The apparatus 20 comprises a fault monitor 1 for inspecting the network 21 for faults in the various elements 2, 3, 4, 5 of the network 21. The monitor 1 provides a first output to a duration discriminator. failure 6, which identifies if the failure has a long duration or a short duration. If the fault has been canceled within a predetermined time, the discriminator 6 causes the information with respect to the fault that is going to be transmitted to the memory 6. The information stored on each fault includes information with respect to the network elements 2, 3, 4 or 5 to which it refers, the time of start of the fault and the time of cancellation of the fault. The start and cancellation times can be reported by the relevant network element 2, 3, 4 or 5, or through the monitor 1. If the fault has not been canceled within the predetermined time, the discriminator causes the activation of an alarm 7, to indicate a non-transient fault to the user on the screen 222. For some purposes, it may also be necessary for the user to alter the passenger faults as they occur, for example, if the equipment needs to be reset after an interruption. . If such an arrangement is required, alarm 7 is activated, each time monitor 1 detects a fault. An update processor 9 periodically outputs the data from the memory 8, at a scan interval t controlled by a time controller 10. The update processor 9 instructs the memory 8 to erase the information in relation to any failure during a period T (the period of analysis) that has elapsed even if these have been canceled. The period of analysis T and the exploration interval t both are selectable, subject to the exploration interval that does not exceed the period of analysis. The analysis period T is typically several times the scanning interval t, so that at the end of each scanning interval, the faults in the previous T / t scanning intervals can be removed. The processor 9 then passes the data to a counter 11, which counts the number of failures in relation to each element 2, 3, 4, 5, which are canceled within the previous analysis period. The obtained values are passed to a comparator 12, which compares the values with the activation and deactivation threshold values stored in the threshold memory 13, and the result is passed to an alarm controller 14. The controller of alarm 14 also receives an input from an alarm status monitor 15, and causes the alarm indicator 7 to be operated (activated or bypassed) according to its pre-existing state and the results of comparator 12 are in accordance with the following true table .

In this way, when the count value is equal to or less than the deactivation threshold, the alarm indicator goes off. If it rises to the deactivation threshold, the alarm indicator 7 remains off until, if it rises to or above the activation threshold, the alarm indicator lights up. If it is between the thresholds, the alarm indicator remains in the "on" condition. If it falls further, below or at the deactivation level, the alarm goes off. The alarm indicator can be maintained in its active state until a number of scanning intervals have passed due to its activation, even if the deactivation threshold is passed, in order to allow sufficient time for it to draw attention operator. Alarms can also be manually overridden. A separate alarm can be provided, which remains active after a transient fault has been reported, until the alarm is canceled by recognition by the operator. A user input 16 associated with the board 224 or other input device (eg, a "mouse") allows the fault duration used by the fault duration discriminator 6, the scan interval used by the time controller 10. , the analysis period used by the update processor 9, and the threshold values stored in the threshold memory 13 that are selected, and allows the selection of those of the elements 2, 3, 4, 5 which are going to be monitored by the monitor 1. The alarm indicator 7 provides information for the user, which relates which of the elements 2, 3, 4, 5, has caused the alarm indicator 7 to be activated and if the activation is due to a failure of prolonged duration or an accumulation of shorter faults. The operation of the apparatus will now be described with reference to the sequence of faults shown in Figure 1. For the purpose of illustration, only one individual element 2 is monitored, the long / short failure threshold is set 5 minutes, the interval of scan at 20 minutes, the analysis period at 1 hour, the activation threshold of the alarm indicator at 3 and the deactivation threshold of the alarm indicator at 1. These values have been selected to illustrate purposes only and are not necessarily representative of the appropriate values for the practical system. It is assumed that no failures were recorded in the memory 8 at the beginning of the illustrated time period, and that the alarm indicator 7 is no longer activated. The synchronizer 10 activates the update process 9 to operate every 20 minutes. In the first of such operation, at the end of the scanning interval t- ^, the memory 8 does not contain data, until the updating processor 9 does not recover the data from the memory 8. The counter 11 passes a result of zero to the comparator 12, which compares this result with the activation and deactivation thresholds. Since the count of zero is less than several thresholds, the indicator of alarm 7 is not activated. During the second scanning interval t2, the A fault is detected by the monitor 1. This fault is cleared before the 5-minute threshold established by the discriminator 6 and thus the data that is related to the A fault is passed to the memory 8. At the end of the scan interval, the update processor 9 recovers the data in fault A, but does not eliminate fault A in memory 8, because a time shorter than the analysis period T has elapsed since its presentation . The data is passed to the counter 11 and the value 1 is compared in the comparator 12 with the values stored in the threshold memory 13. The counter is now in the deactivation threshold, but the alarm indicator 7 is no longer activated and the count is still less than the activation threshold in such a way that the indicator of alarm 7 is not activated. Similarly, at the end of scan interval t3 no faults have been added to memory 8 and no faults have been in memory for a period longer than for the period of analysis T, updated and counted and compared as for the previous exploration interval t. During the scanning interval t4 two short faults B, C occur. These are added to the memory 8 in the same way that the fault A was recorded in the scanning interval t. At the end of the scanning interval t4, the counter 11 now counts three faults. This value 3 is compared to the value stored in the threshold memory 13 and is found to be in the activation threshold value. The output of the comparator 12 is input to an alarm indicator control 14 together with an input of the alarm indicator status monitor 15, which indicates that the alarm indicator is actually not activated. This causes the alarm control 14 to activate the alarm indicator 7. During the scanning interval t5, no fault occurs. At the end of the scanning interval t ^, the update processor identifies the fault A as having occurred more than one time T before the present, having occurred in the scanning interval t2 - The fault A in this way is removed from the memory 8 The counter 11, therefore, only counts two faults (faults B and C) and this value is compared with the threshold values stored in the memory 13. Although the value has now fallen below the activation threshold, it remains above the deactivation threshold, so that the alarm indicator 7 remains in its present (activated) state. During the scan interval tg, another fault D occurs, and another fault E is occurring as the scan interval is terminated. Since the E fault has not yet been canceled, it can not be determined if it exceeds the long / short discrimination threshold. At the end of the scan interval tg, the update processor, therefore, finds three faults, B, C and D stored in memory 8, none of these has been stored for more than the total period of analysis T. Of this In this manner, they are not eliminated from the memory 8, and they are all counted by the counter 11. The value of 3 determined by the counter 11 is compared with the threshold values stored in the threshold memory 13 and found to be in the activation threshold. Since the alarm indicator status monitor 15 identifies the alarm indicator 7, already activated, a new alarm indication is not generated. If the fault E of the scanning interval t7 is canceled within the long / short discrimination threshold then it is stored in the memory 8. At the end of the scanning interval t7, the update processor updates the memory 8 eliminating the faults B and C , since the analysis period T has now elapsed as it occurs. The counter 11, therefore, only counts two faults, D and E, and this value of two is compared through the comparator 12 with the threshold values. Although the value has again fallen below the activation threshold, it remains above the deactivation threshold so that the alarm indicator 7 remains in its present state (activated). Similarly, in the scanning interval tg, no new fault occurs and none is eliminated by the update processor 9, so that the number of faults counted by the counter 11 remains in two and the alarm indicator 7 remains on. Almost at the end of the scan interval tg another fault F is started, but it has not reached the long / short discrimination threshold or has been canceled before the end of the scan interval. At the end of the period tg, the fault D is eliminated from the memory 8 through the update processor 9 as the analysis period T is ending for this fault. This leaves only the fault E in the memory 8 to be counted by the counter 11. The output of the value of the counter 11 to the comparator 12 is, therefore 1, which is the deactivation threshold. The alarm control 14, in this way, deactivates the alarm indicator 7. During the scanning interval t- ^ Q the long / short discrimination threshold is terminated before the fault F is canceled. At the end of this period of discrimination, the discriminator 6 identifies the fault as a long fault and activates the alarm indicator 7 immediately. In certain circumstances, the system may receive an indication of "cancellation", when no indication of "failure" has been received. This may occur, for example, when the system is first established, or if the "fault" signal is not received due to a more general fault than the one formed. The system is arranged to not take into account any indication of "damaged" cancellation. The alarm indicator 7 can give information regarding the faults such as the elements 2, 3, 4 or 5, which are generating the faults, their total duration and the time of occurrence.

The apparatus 20 described above can be used to inspect the performance of a network at any level. For example, a network operator responsible for the maintenance of telecommunications equipment may wish to inspect failures of specific equipment items. Individual passenger interruptions may be insignificant, perhaps caused by an external cause. For example, if a failure occurs in the equipment at a point in the system, it will cause an interruption to the operation of many other items of equipment that are in communication with it. If the network has the so-called "auto-repair" capabilities, the calls can be redirected to avoid the use of a failed device, and only a single temporary interruption is detected for the other items. However, if a large number of passenger faults are detected in a single piece of equipment, this may be indicative of a development situation where the performance of that item is deteriorating. This information can be used to take a remedial action before the component fails completely, either by redirecting call traffic to avoid or minimize its use, or by repairing the component. A customer of a service provided by the network usually does not know of the individual failures of the equipment, since the same service is not interrupted. Such customers usually have service level agreements with the network operator, in which the maximum levels of service interruption are specified (either as a number of individual interruptions or as a proportion of the total time). The alarm indicator 7 may include means for recording the details of these interruptions so that the service is inspected and registered so that the network operator and / or the customer can have visibility into the performance of the system. The alarm indicator 7 can inspect the stored details to identify when the proportion of time, or the total time elapsed, for which the service has been interrupted exceeds a predetermined value, and the alarm is activated when this value is exceeded.

Claims (26)

1. A method for inspecting a plurality of elements of a telecommunications system to locate faults, and generate alarms in response to them, the method is characterized in that it comprises the steps of inspecting the elements for the occurrence of faults, measuring the duration of each occurrence of damage and, for each occurrence of damage, if the duration exceeds a predetermined value, activate an alarm indicator, if the duration does not exceed the predetermined value, increase a stored value.

2. The method in accordance with the claim 1, characterized in that the times of start and of cancellation of faults are recorded, and where after the time of start of a fault has been recorded, a period of delay is initiated, and where if the cancellation of the fault before the end of the period of delays, the stored value is increased, and if the period of delays is finished before the cancellation of the record of the fault, the alarm indicator is activated.

3. The method according to claim 1 or 2, characterized in that the system is inspected for occurrences of system inability, or a system function, to a user.

4. The method according to claims 1 to 3, characterized in that an alarm indicator is activated when the stored value reaches a value equal to or greater than a threshold value within a predetermined range.

5. The method according to claim 4, characterized in that it comprises the steps of: establishing a period of analysis; inspect the system continuously for occurrences of faults; at the end of a scan interval shorter than the period of analysis, count the number of occurrences of the fault during the analysis period, which ends at the end of the scan interval; and activate an alarm indicator if the number of occurrences of the fault in the analysis period is equal to or greater than a threshold value.

6. The method according to claim 5, characterized in that the start and breakdown times are recorded, and at the end of each scan interval, the faults, for which a longer time has elapsed than the analysis period due to that the registered cancellation time has not been counted.

7. The method according to claim 5 or 6, characterized in that it comprises the additional step of deactivating the alarm indicator if the number of occurrences of the fault in the analysis period is equal to or less than a second threshold value, the second value of threshold being less than the first threshold value.

8. The method in accordance with the claim 5, 6 or 7, characterized in that an additional alarm indicator is also activated if the number of occurrences of the fault in the analysis period is equal to or greater than a threshold value, and the additional alarm indicator remains activated until it is recognized by an operator.

9. The method according to claim 5, 6, 7 or 8, characterized in that the alarm indicator is activated if the number of faults in the analysis period exceeding the threshold is the same as that activated if the duration of a fault exceeds a predetermined value.

10. The method according to claim 9, characterized in that the alarm remains always activated as long as any stored value exceeds the second threshold or a fault of duration greater than the predetermined value remains unnumbered.

11. The method in accordance with the claim 9 or 10, characterized in that the alarm has different activation states depending on whether the stored value remains above its threshold, or whether a long-term fault, or both, remains without cancellation.

12. An apparatus for inspecting a plurality of elements of a telecommunications system to find faults, characterized in that it comprises an alarm indicator, detection means associated with each element for detecting occurrences of faults, time control means for measuring the duration of each fault detected by any of the detection means, activation means for activating the alarm if the duration measured by the time control means exceeds a predetermined value, counting means for storing an account number, and increment means for increasing the account number stored in the counting means if the duration of a fault measured by the time control means does not exceed the predetermined value.

13. The apparatus according to claim 12, characterized in that the detection means comprise means for detecting the start and the cancellation of individual faults, the time control means comprise time recording means for recording the start time and the cancellation of each failure, and delay time control means initiated by the detection of the start of a fault, and restored by the detection of the cancellation of the fault, to measure the time elapsed since the start of the fault, the activation means being ready to operate if the elapsed time measured by the delay time control means reaches a predetermined value, the increment means being arranged to operate if the detecting means detect the cancellation of the fault before the elapsed time measured by the delay time control means reach the predetermined value.

14. The apparatus according to claim 12 or 13, characterized in that it comprises additional alarm activation means that are operated if, within a predetermined analysis period, the account number stored in the counting means reaches or exceeds a predetermined value of activation threshold.

15. The apparatus in accordance with the claim 14, further characterized in that it includes a memory arranged to store the number of occurrences of the fault condition within each plurality of scanning intervals, whose total duration is that of the period of analysis.

16. The apparatus in accordance with the claim 15, characterized in that it includes time recording means for recording the start and cancellation times of each fault, and distribution means for distributing fault occurrences in predetermined analysis periods and scanning intervals according to their start times and / or cancellation.

17. The apparatus according to claim 15 or 16, characterized in that it includes an update processor, the update processor being arranged, at the end of each scan interval, to take out the data stored in the memory and supply it to the counting means, and to instruct the memory to delete the data in relation to the oldest scan interval for which the data was stored.

18. The apparatus according to any of claims 14 to 17, further characterized by comprising deactivation means that cause the alarm means to be deactivated if the number stored in the counting means is equal to or less than a deactivation threshold value, less than the activation threshold value.

19. The apparatus according to any of claims 14 to 18, characterized in that it comprises an additional alarm indicator, which is also activated if the number of occurrences of the fault in the analysis period is equal to or greater than a threshold value, and means that are controllable by the operator to deactivate the additional alarm indicator.

20. The apparatus according to any of claims 14 to 19, characterized in that the alarm indicator activated by the additional alarm activation means is equal to that activated if the duration of a fault exceeds a predetermined value.

21. The apparatus according to claim 20, characterized in that the alarm indicator is arranged to remain activated whenever any of the stored values exceeds the second threshold or a duration fault greater than the predetermined value, remains without cancellation.

22. The apparatus according to claim 20 or 21, characterized in that the alarm has different activation states depending on whether the stored value remains above its threshold, or if a long-term fault remains without cancellation, or both.

23. The apparatus according to any of claims 12 to 22, characterized in that the detection means comprises discriminating means for distinguishing the occurrences of the fault condition having a first predetermined characteristic, from those having a second predetermined characteristic, the means of counting being willing to count only those occurrences that have the first characteristic.

24. A telecommunications system comprising the inspection apparatus according to any of claims 12 to 23.

25. A method for inspecting a telecommunications system for faults substantially as described with reference to the drawings.

26. An apparatus for inspecting a telecommunications system for faults as described with reference to the drawings.