MXPA01002734A - Method and system for detecting loss of signal in wavelength division multiplexed systems - Google Patents

Abstract

A method and system for unambiguously detecting fiber (419-420) cuts in an optical network regardless of the number of EDFAs (499) that are located between the fiber cut and the monitor point. In accordance with our invention, the power of a marker wavelength is compared to the power of a nearby spectral region. Where the comparison indicates that the power ratio is approximately equal to unity a flag is raised indicating that there is fiber cut. Where the comparison indicates that the power ratio is much greater than unity the flag is not raised. The monitoring point includes circuitry for detecting the ratio between the marker wavelength and the nearby spectral band and for indicating when there is a fiber cut.

Description

METHOD AND SYSTEM FOR DETECTING LOSS OF SIGNAL IN MULTIPLEX TRANSMISSION SYSTEMS OF LENGTH DIVISION COOL FIELD OF THE INVENTION This invention relates to multiplex systems of Wavelength Division (WDM for its acronym in English) and specifically to the detection of flaws in WDM systems.

BACKGROUND OF THE INVENTION The WDM technology has provided an effective solution in cost to the exhausted fibers in the communications networks increasing the performance or flow of the data of the network without requiring the installation of new fibers. In a WDM system each of the various input signals enters a WDM node or network element and is assigned or converted to a specific wavelength, typically, in the 1550 nanometer (nm) band. After the wavelength conversion each individual signal wavelength or channel REF. : 127625 it is then multiplexed by multiplexing the division of the wavelength and transmitted over the same fiber. For WDM technology to be truly viable as a network solution, WDM systems must also be able to survive failures that occur in any network. The concept of network survival takes on additional importance in WDM systems since the loss of a fiber could be catastrophic and expensive given the huge amount of customer data, for example, multigigabit data, a WDM system is transported in a single fiber. In response to concerns regarding the survival of the WDM network, self-regenerating WDM rings and diverse point-to-point protection architectures have been proposed. A self-regenerating ring is a network architecture that connects to nodes in a physical ring topology with shared bandwidth and auto-regeneration capabilities to overcome network failures. For the purposes of this description each node in a ring is connected to another node via a fiber. If a fiber cut or other failure occurs, for example, the failure of a node, then the ring automatically switches to a fiber in wait and, in some cases, electronic circuits waiting. Various point-to-point protection systems similarly protect the network of fiber cable cuts by automatically switching customer data to a waiting fiber that is on the route along a different path. In any case, the automatic switching protection can be done optically, that is, by switching the received optical signal to a waiting fiber, or electrically, that is by switching the electrical representation of the received optical signal. The automatic protection switching in the WDM networks promises a considerable cost saving in relation to the protection of pure Synchronous Optical Networks (SONET). However, some fundamental points must be considered before the automatic protection switching in WDM systems is used. One such fundamental issue for WDM systems is the detection of fiber cuts in the optically amplified links. The detection of a fiber cut or a loss of signal has proven to be a difficult issue in WDM systems because the links between the nodes are usually amplified optically through Fiber Amplifiers that Contain Erbium (EDFAs). Typically, at each WDM node the signal is amplified by an EDFA after multiplexing and before transmission to the network facilities or fiber links. Similarly, after reception, at each WDM node the signal is again amplified by another EDFA before demultiplexing. Depending on the distance between a transmitter and the receiver, one or more additional EDFAs can also be placed at specific points along the path of the fiber. As the distance and the number of amplifiers between the cut of the fiber and the optical monitor or receiver increases, the amplified spontaneous emission of EDFA grows with each EDFA in the optical path. Specifically, when there is no optical input signal in a saturated EDFA, the amplified spontaneous emission can be increased sufficiently after several EDFAs so that fiber cutting could not be detected. In fact, due to the amplified spontaneous emission, the measurements of the total optical power or even of the optical power within a spectral band are insufficient to measure certain cuts of fibers.

The detection of the total optical power may fail to detect certain cuts of fibers depending on the location of the cut of the fiber in relation to the EDFAs and the detection threshold. In some fiber links or spacings there are no EDFAs beyond those in the nodes or elements of the network, while in others there may be more than one EDFA. FIG. 1 shows a pair of working fiber / protection fiber of the prior art in a WDM ring that includes the addition-drop elements of the network 120 having protection switches 121 and 122. Specifically, as exemplified in FIG. 1, in a link 110 there are four EDFAs 199 both in the clockwise and counterclockwise directions between two of the drop-add elements 120 (note here that although FIG. ring this discussion also belongs to point-to-point architectures). A fiber cut that occurs in a 111 link could easily be detected at a point 150? of the monitor because the total optical power at point 150 of the monitor drops to zero. However, for more remote fiber cuts, such as those occurring in the 112, 113, 114 and 115 sub-links, the amplified spontaneous emission provided through the intervention of the EDFAs 199 provides optical power for the point 150? of the monitor. The relationship between the power detected at point 150? of the monitor in relation to the intervention number of EDFAs 199 is shown in FIG. 2. FIG. 2 is a simulation of the wavelength domain that illustrates the problems with fiber cuts. The simulation assumes the specific characteristics and spacing of the EDFA. Although the results for other EDFA designs may differ quantitatively, the qualitative characteristics shown in FIG. 2 will be similar. As shown in FIG. 2, without a fiber cut, the level 201 of total optical power at point 150? of the monitor was approximately 18 dBm. If a fiber cut occurs in the 112 sublink, that is, with a EDFA 199? just before point 150? of the monitor, the total optical power 202 detected at point 150? drop it to approximately 4 dBm after 0.5 milliseconds (ms) On the other hand, where there are two or more EDFAs between the 150 point? and the fiber cut, ie a fiber cut in the sub-links 113, 114 or 115, the total optical power will return to within 2 dB of the total optical power when the fiber is intact. In fact, when there is already three or four EDFAs between point 150? and the fiber cut, the total optical power 204 or 205 will never vary more than 4 dB and will return to full optical power level 201 within less than 0.5 ms. As seen by the power level 203, with two EDFAs, the power level will return almost to the power level 201. The measurements made in the test bench have confirmed the results shown in FIG. 2. Based on the simulations and test bench measurements, the following conclusions have been drawn with respect to the simple monitoring of optical power to detect a fiber cut in a WDM system: when no EDFA is positioned between the fiber cut and the point of the monitor, the fiber cuts can be identified correctly; if an EDFA is between the fiber cut and the monitor, the correct identification of the fiber cut could not be completed without a careful selection of the detection threshold used to detect a fiber cut; and when more than two EDFAs were located between the point of the monitor and the fiber cut, a threshold could not be established that would allow detection of the fiber cut. It has also been investigated and found a unsatisfactory monitoring of power within a narrower spectral band to detect fiber cuts at point 150 of the monitor instead of detecting the total optical power in the fiber. In relation to this, an additional marker wavelength has been inserted into the fiber at the output of a network element. It was found that if the power in the marker was high enough, the simple detection of the marker is sufficient to indicate a fiber cut. However, the high power at the marker wavelength results in a lower EDFA gain for the wavelengths of the signal and is therefore undesirable. On the other hand, if the wave marker is at a power level comparable to the signal wavelengths, as shown in FIG. 3, then the marker allows detection of the change from level 301 of normal power to level 302 of lower power for a cut with only one EDFA before the monitor; but the marker alone will not give the required contrast to detect a fiber cut after more than two EDFAs, as shown by the power levels 303, 304 and 305. Although it was found that a contrast of 10 dB is possible if you use a very narrow band filter (filter width less than 0.2 nm) to generate the spectral band, such a filter places unrealistic requirements on the filtering of the marker wavelength. However, it will be noted that the width of the narrow band filter is dependent on the test switching. Other methods are known in the art. One such method is described by J. L. Zyskind, et al., In U.S. Patent Application No. 6,008,915, entitled "Method of identifying Faults in WDM Optical Networks." In this method Zyskind, et al., Uses an additional laser to insert an additional monitoring channel into the fiber of the WDM system along with the signal channels. The power in the monitoring channel and the spontaneous emission amplified by the EDFAs used along the fiber trajectory are then monitored and compared with the detected faults. That is, a change in the power in the same direction in the monitoring channel and the amplified spontaneous emission, for example, both the increase or decrease, is interpreted as signal channels either fall or addition. On the other hand, a change of power in the monitoring channel and the amplified spontaneous emission in the opposite direction is interpreted as an indication of the loss total of a fault. The method of Zynkind, et al., Requires additional components that include a laser monitoring, couplers, and narrow band filters to be implemented. More importantly, as the channels are added or decreased, the power level of the monitoring channel and the amplified spontaneous emission change thus change the threshold level to detect the faults. The Zynkind method also requires a fairly sophisticated detector that will require you to keep track of five different cases for the current loss upstream and the signal channels that could occur. This method, therefore, will probably require a set of programs to make a decision. In his article entitled "A Novel In-Service Surveillance Scheme for Optically Amplified Transmission Systems "(published in IEEE Photonics Technology Letters, Vol. 9, No. 11, November 1997) Chan, Chun-Kit, et al., Describes another methodology of the prior art to detect failures in the WDM systems Chan, et al., uses the amplified spontaneous emission spectra, not planes of the EDFAs as the light source to monitor or track the fiber channel to detect faults.

By the method of Chan, et al., The Bragg gratings of the fibers are placed near the inlet end of each EDFA, except the first EDFA after the transmitter, along the fiber path. Each of the Bragg gratings of the fibers then filters a different wavelength into the unused spontaneous emission spectrum. Each filtered wavelength is assigned to each immediately preceding amplifier to a Bragg grid of the fibers. Because the Bragg grating of the fibers operates as a suppression notch filter, a power loss upstream of the Bragg grid of the fibers occurs resulting in a spectral pulse that Bragg grating of the fibers with a length of different wave. By this method, fiber cuts can be located in the fiber space between any of the two amplifiers. While this method does not require the use of additional lasers, it nevertheless requires Bragg gratings of the fibers as additional components. This method will also require sophisticated spectral monitoring. This method may not be able to detect breaks in the fibers that occur between a Bragg grid of fibers and the input of its assigned amplifier, nor is it they would detect partial faults of certain amplifiers. All of the above approaches require either additional components or are not capable of detecting all fiber slices, without considering the location of the fiber cut in relation to an amplifier or a number of amplifiers.

BRIEF DESCRIPTION OF THE INVENTION The invention provides a method and a system for unambiguously detecting fiber slices in an optical network without considering the number of EDFAs that are located between the fiber cut and the point of the monitor. According to the invention, a marker wavelength is detected at the output of network elements that are part of a WDM network. The power level in the marker is then compared to the power level in a region of wavelength without signal. If the ratio of the power level in the spectral band of the signaling wavelength to the power level in the wavelength region without signal is high, then the fiber is intact. In contrast, if the ratio of the power level in the spectral band of the marker wavelength to that in the region of the wavelength without signal is approximately equal to unity then there is a cut in the fiber.

The invention adds only a moderate amount of cost and complexity to WDM networks while providing sufficient information for unambiguous identification of fiber slices, requiring no more than one additional laser in each network element in a WDM ring architecture. Furthermore, according to the invention, an additional laser is not necessarily required in certain rings and network configurations point to point. Furthermore, without considering the configuration of the WDM network, the invention requires only the circuitry that is capable of measuring the power difference in two nearby spectral regions to measure a fiber cut. According to the invention, a fiber cut can be detected without considering the EDFA number between the fiber cut and the point of the monitor, thereby eliminating the generation of false alarms or false negatives; false negatives are defined in the detection of a signal loss when there is no signal loss, BRIEF DESCRIPTION OF THE DRAWINGS These and other advantageous features of the invention will be understood from the following detailed description together with the accompanying drawings, in which: FIG. 1 illustratively represents a WDM ring of a prior art using EDFAs; FIG. 2 represents the results of measurements of the total output power at a point on the monitor for the fiber cuts for the WDM network shown in FIG. 1; FIG. 3 represents the results of the simulation of the total power in a 1 nanometer band at a marker wavelength for the fiber slices for the WDM network shown in FIG. 1; FIG. 4 illustratively represents a WDM ring architecture employing EDFA according to the invention; FIG. 5 illustratively represents a point-to-point WDM architecture employing EDFAs according to the invention; FIG. 6A represents the power spectrum of a complete optical network having a marker wavelength and a non-signal spectral region used to detect the fiber slices according to the invention when the fiber is intact; FIG. 6B represents the power spectrum in a complete optical network having a marker wavelength and a non-signal spectral region used to detect the fiber slices according to the invention after a fiber cut; FIG. 6C represents the power spectrum of a complete optical network having a non-signal spectral region used to detect fiber slices according to another aspect of the invention when the fiber is intact; and FIG. 7 represents the results of the simulation of the optical power before and after a fiber cut in a marker channel and a channel without near signal and the proportion of those powers according to the invention.

DETAILED DESCRIPTION Returning now to FIG. 4, there is illustrated an illustrative embodiment of the WDM ring having a monitor device or circuitry 450 for detecting a fiber cut without considering the number of EDFAs that is located between the fiber cut and the monitor circuitry 450. The ring includes a plurality of addition-drop network elements 400 having protection switches 421 and 422, connected by a closed circuit 419 of interior work and a closed circuit 420 of protection, each closed circuit includes a plurality of EDFAs 499. As shown in FIG. 4 when a signal enters a WDM node or 400x network element in the sublink 411, a portion of the signal branches or drifts and is fed to a monitoring device or circuitry 450 ?. The circuitry 450? is able to detect the power in two nearby spectral bands, compare the power difference between the two spectral bands and generate a flag if the comparison indicates that the power in both spectral is approximately equal to unity. In accordance with the invention, as seen in FIG. 4, the element 400? of WDM network include two circuits of 450 monitoring? and 4502. The circuitry 450? illustratively monitor the closed circuit 419 while the circuitry 4502 monitors the closed protection circuit 420. That portion of the signal that is not fed to the 450x circuit is demultiplexed and either dropped or fed through node 400? to sub-node node 4002. In addition to monitoring circuits 450? and 4502, the network elements 400 may also include an external laser 460 for inserting a marker wavelength or signal in the outer shield closed circuit 420; the signals in the outer closed circuit or protection ring 420 in FIG. 4 are illustratively propagating in a counter-clockwise direction. The laser 460 is necessarily in those ring architectures where the protection of the fibers does not carry a signal until there is a fault in the working fiber, for example, the rings switched in the unidirectional path. In other words, the laser 460 will necessarily be in the closed protection circuit 419 if during normal operation all the signals propagate clockwise only in the working fiber or the circuit inner closed 420 in FIG. 4. On the other hand, in ring architectures where both fibers carry a signal during normal operation, for example, a bi-directional line-switched ring, an additional laser would not be necessary. Returning now to FIG. 5, there is shown an illustrative embodiment of the invention having a diverse point-to-point routing architecture including endpoint network elements 500? and 5002 interconnected by work fibers 566 and 568 and protection fibers 567 and 569, each including a plurality of EDFAs 599. A monitoring circuit 550 in each network element 500 detects a fiber cut according to the invention. In this embodiment of the invention the WDM network elements 500 will not require an additional laser as in the case of ring architectures. This is the case because the same signal is transmitted in both a work fiber 566 and a protection fiber 567. As was the case in the ring architecture mode of FIG. 4, the monitoring circuitry 550 is able to detect the power in the two near spectral bands, compare the power difference between the two spectral bands and raise or generate a flag if the comparison indicates that the power in both spectral bands is approximately equal to unity. When the circuitry 550 in node 500? generates a flag, a fiber cut or signal loss in working fiber 566, node 500? switches the signal via an optical switch 577 to the protection fiber 567. Yes, contrary to what is shown in FIG. 5, a hot wait signal is not transmitted in the waiting channel, then an additional laser may be needed in the protection line during normal operation. Another aspect of the invention shown in FIG. 4 and in FIG. 5 is the use of fixed gain EDFAs or corrected with gain 490 and 590 along the fiber trajectories. A fixed gain or a corrected amplifier is an EDFA that has the output power maintained at a constant level by either inserting an extra channel, known as a stabilization or compensation channel, into the transmitted signal. The compensation channel is used to avoid having to count the number of channels transmitted and increase the power in the remaining channels each time a channel falls out of the signal. With reference to FIG. 4, the EDFAs of fixed gain 490, which have compensation channels, in addition to having an optical amplifier circuit 499, also include the circuitry 498 which is used to insert a compensation channel in the fiber. The compensation channel is mainly used to maintain a constant total power level at the points in the network where a single or multiple channels fall from the fiber, that is, in an Addition-Drop Multiplexer of the Wavelength Division Multiplexer (WADM). The compensation channel is usually transmitted at a wavelength that has a gain almost equal to the gain of the channels. The compensation channel may be between two channels or may be a spectral region just inside the flat gain portion of the EDFA passband. The power level in the compensation channel increases each time a channel falls in a WADM and vice versa, to maintain a constant output power. Returning now to FIG. 6A, there is represented an illustrative modality of marker wavelength or compensation channel 620 and a spectral band 630 used in accordance with the invention to detect a fiber cut or a signal loss when the fiber is 1 intact They are also represented in FIG. 6A the wavelengths or channels 640 that are used to transport the information in a WDM system. In accordance with the invention, the marker wavelength 620 could be generated by any of the methods discussed below. The region of the spectral band 630 is chosen sufficiently close to the marker wavelength region 620 so that an accurate comparison can be made. The spectral band is also chosen far enough away from the work signal channels 640 so that they can be separated using a filter. When using fixed gain amplifiers, it was found that a fiber cut or signal loss can be detected without the addition of some equipment or without the modulation or increase in the power of the compensation channel of a fixed gain EDFA. According to the invention, if the proportion or difference in power levels between the compensation channel and a spectral band without signal is monitored, then the fiber slices can be easily detected. As shown in FIG. 6A, when the fiber is intact, the score or compensation signal 620 and the work signals 640 - - are present in the spectrum. The spectral band 630 consists of power provided by the amplified stimulated emission and is linked within a region formed by lines 631 and 632. In accordance with this aspect of the invention, the spectral band 630 is chosen so that the amplified stimulated emission both of the marker wavelength 620 and the spectral band 630 are approximately equal. As such, when the fiber is intact, as in FIG. 6A, the ratio of the power at the marker 620 and the spectral band 630 is much larger than the unit. On the other hand, as shown schematically in FIG. 6B, when there is a fiber cut or some other event causing a signal loss, the power ratio of the marker wavelength 620 and the near spectral band 630 is approximately equal to unity. Using the compensation channel as the marker wavelength 620 avoids the use of additional equipment, eliminates the gain changes of the potential amplifier, and eliminates the possibility of cross-modulation of the signal channels, as can occur with other approaches or methodologies to generate the marker wavelength 620.

The marker wavelength 620 can be easily generated for the working fiber if the optical gain setting is used for the EDFA within the network element and if the optical power used to set the gain of the EDFA is allowed to propagate to the next element of net. Therefore, going back to FIG. 4, if the amplifier which is located after the multiplexer 430 in the nodes 400 WDM is an amplifier 490? of fixed gain, then the amplifiers 491? and 4912 that are inter-positioned in the fiber links would not need to be EDFAs with fixed gain. As such, the invention requires fixed gain EDFAs only at the WDM nodes 400 if the optical power used to set the gain of the EDFA is allowed to propagate to the next node 400 in the closed circuit; this advantage of the invention is equally applicable to point-to-point architectures as indicated in FIG. 5. In FIG. 4, it is also noted that instead of requiring the amplifier 490? whether it is a fixed gain amplifier, it may be more convenient to insert a marker wavelength into the output of element 400? of network. In accordance with this, the network architecture would be independent of the WDM node design, that is, it would be a decision of the network glider - - insert the marker wavelength by appropriate placement of fixed gain EDFAs. According to the analysis, on the other hand, it has been found that the generation of 620 marker wavelengths using the compensation channel of a fixed gain amplifier is not possible in the protection fiber for a WDM ring where the protection fiber carries the signals during the fault conditions. Although it is possible to generate a marker channel, not all faults requiring switching protection will be detected by this method, and in some cases protection switching will be initiated when it is not necessary. It has been found, as indicated, FIG. 4, that an additional laser 460 placed as shown between the input and output protection switches may be required at each network element for the protection fiber depending on the architecture of the ring. In accordance with another aspect of the invention and as schematically depicted in FIG. 6C, wherein a compensation channel is not available, ie, where the fixed gain EDFAs are not used, the proportion of the signal wavelength region 650, denoted as the region between s 649 and 651, as a near spectral band 635, denoted as the region between lines 634 and 636, can be used to detect fiber cuts. Again, as discussed above, when the fiber is intact, the ratio of the power in the wavelength region 650 and the band 635 is much greater than unity. On the other hand, when there is a fiber cut, the spectrum of FIG. 6B and the proportion of region 650 and band 635 is about unity, or more accurate proportions of wavelength ranges. It has been noted that this detection method is limited to spaces where the actual signal wavelengths are in service in the transportation traffic. That is, in a space without any power in the 650 wavelength region, the power ratio of the 635 region and the 650 band would be about unity even though the fiber was intact. This disadvantage for this aspect of the invention promotes the case for the propagation of the compensation channel. Of course, in optical networks that do not have fixed gain EDFAs and therefore do not have compensation channels, the network operator can generate a signal wavelength as a live live signal that can be used to monitor the link for a fiber cut until the link is active with customer service. The modality of the approach in FIG. 6C can also generate false negatives. A network operator can, however, generate a lifespan signal to protect against the generation of false negatives. FIG. 7 illustrates the fiber cut simulations in a sublink 413 of FIG. 4 according to the invention. As FIG. 7 illustrates, when the fiber is intact, the ratio of the power level of the marker channel 720 with respect to the near spectral band 730 is much larger than unity. Thus, the power difference 710, in units of decibels, between the marker channel 720 and the spectral band 730 is much less than zero, approximately -35 dB. On the other hand, when there is a fiber cut in the 413 sub-link, the power ratio is close to unity and the difference 710 in power is close to zero dB. Similar results are obtained for the fiber cuts in other sublinks in FIG. 4. According to the invention, the fiber cut is detectable within 100 μs, or within 10 ms the period of time allowed for the detection of catastrophic faults in the Switched Telecommunication Network Public. It is also noted that the time between cutting fiber and establishing a new power is less than about 250 μs required as shown in FIG. 3. Thus, by means of the present method not only is the contrast greater, but the switching takes place more rapidly. The results in FIG. 7 assume that the marker 730 and the band 720 have a nearly equal gain. The power ratios may be different if the marker 730 and the band 720 have different gains, but a large enough difference in the power ratios will still be detectable for a large number of EDFAs. During the course of the work it has been noted that some considerations must be given to increase the power in the compensation channel or marker 620 to achieve a better contrast or simply to use the compensation channel to detect a fiber cut. It has been found that while increasing the power of the compensation channel, it can increase the contrast, this reduces the gain in fixed gain amplifiers and the power per channel in the chain. Alternatively, the EDFA can be redesigned for the point of operation so that when the power in the compensation channel - - increase, the gain remains constant. In addition, care must be taken to avoid tilt of the gain because the fixed gain EDFA has an equivalent input that is greater than for which the amplifier is designed. It has been found that the gain inclination can be avoided by attenuating the work signals 640 before the EDFA enters. Concerning the fixed profit EDFA, the attenuation is equivalent to the fall of some channels, thus increasing the power in the compensation channel. The above description is exemplary of the invention. Numerous modifications and variations can be made by those skilled in the art without departing from the scope and spirit of the invention. It is noted that in relation to this date, the best method known by the applicant to carry out the aforementioned invention, is that which is clear from the present description of the invention.

Claims (19)

1. CLAIMS Having described the invention as above, property is claimed as contained in the following 1. In a wavelength division multiplex system having at least two nodes connected by a fiber and the fiber having at least one amplifier fiber coupled thereto, a method for detecting in one of said nodes a fiber cut, said method characterized in that it comprises: generating a marker wavelength; and detecting, in said node, a power ratio between the generated signal wavelength and a near spectral band.
2. 2. - The method according to claim 1, characterized in that the method further comprises the steps of: determining that there is no fiber cut when said detection step indicates a much larger proportion than the unit; and determining that there is a fiber cut when said detection step indicates a ratio approximately equal to unity.
3. 3. - The method according to claim 1, characterized in that the generated signal wavelength is the amplifier compensation channel.
4. 4. - A system for detecting a fiber cut in a fiber containing at least one fiber amplifier between a monitoring point and the fiber cut, said system is characterized in that it comprises: the circuitry for generating a compensation channel, and the circuitry to detect the power ratio between the compensation channel generated and a spectral band having approximately the same gain as the compensation channel; said compensation channel circuitry and detection circuitry are coupled to the fiber so that a cut at any location or location along the fiber is detected.
5. 5. - The system according to claim 4, characterized in that the detection circuitry further comprises the circuitry to indicate that there is a fiber cut if the detected ratio is approximately equal to the unit and to indicate that the fiber is intact if the ratio detected is much greater than unity.
6. 6. - A wavelength division multiplex system characterized in that it comprises: at least two node elements interconnected by a working fiber and a standby fiber; a plurality of amplifiers in said fibers; and means for determining a fiber cut coupled to one of said fibers, said means comprise means for comparing the power ratio between a marker wavelength in one of said fibers and a near spectral band.
7. 1 . - The multiplex wavelength division system according to claim 6, characterized in that the amplifiers are amplifiers that contain erbium impurities.
8. 8. - The wavelength division multiplex system according to claim 7, characterized in that the marker wavelength is provided by the erbium-containing amplifiers as a substance in the form of traces.
9. 9. - The wavelength division multiplex system according to claim 7, characterized in that the node elements are arranged in a ring configuration.
10. 10. - The wavelength division multiplex system according to claim 9, characterized in that it further comprises a laser connected to the optical fiber waiting for the ring configuration to provide the marker wavelength.
11. 11. - The wavelength division multiplex system according to claim 7, characterized in that the node elements are arranged in the point-to-point configuration and also comprise means that respond to the determination means for switching the working fibers to the waiting fibers.
12. 12. - The wavelength division multiplex system according to claim 7, characterized in that at least one of said amplifiers includes means for inserting a compensating channel in the fiber in which an amplifier is located.
13. 13. - A method for detecting a fiber cut in a wavelength division multiplex system having at least one amplifier between two node elements, said method characterized in that it comprises comparing the power ratio at a monitoring point in the fiber between the power in a first spectral band and the power in an adjacent spectral band.
14. 14. - The method according to claim 13, characterized in that the detection of a power ratio approximately equal to the unit indicates a fiber cut in said fiber.
15. 15. - The method according to claim 14, characterized in that the first spectral region comprises a marker channel.
16. 16. - The method according to claim 15, characterized in that the amplifier is a fiber amplifier containing erbium as a substance in the form of traces, the marker channel comprises a compensation channel of said amplifier and the second spectral region comprises the work signals of said fibers.
17. 17. - The method according to claim 13, characterized in that the amplifier is a fiber amplifier containing erbium as a substance in the form of traces and the first spectral region comprises the stimulated, spontaneous, amplified emission of said amplifier and the second spectral region comprises the working signals of said fiber.
18. 18. - The method according to claim 14, characterized in that it also comprises the step of inserting a marker channel in said fiber.
19. 19. - The method according to claim 18, characterized in that the marker channel is inserted by a laser. The method according to claim 18, characterized in that said at least one amplifier is a fiber amplifier containing erbium as a trace substance, compensated and the marker channel is inserted into the fiber as the compensation channel of said amplifier. .