JPH0579134B2 - - Google Patents

Description

[Detailed description of the invention]

(1) Field of Industrial Application The present invention relates to a method for non-destructively detecting the position of a defect such as a break point or a connection point of an optical fiber or the position of an open end from one end of an optical fiber to be measured. (2) Prior Art Techniques for non-destructively detecting defect positions in optical fibers or optical fiber lines are very important in the production of optical fibers or the construction and maintenance of optical fiber lines. A light pulse with a constant period is sent into an optical fiber, and among the scattered light generated in the fiber, the backscattered light power and reflected light power that return to the input end side are received in time series at the input end, and the light is The power is electrically converted, logarithmically converted, A/D converted, and averaged, and the temporal change in these optical powers is calculated as z=vt/2 (z is the distance or position from the input end, and v is the distance in the optical fiber. propagation speed of light pulse,
OTDR (Optical
It is a very useful method called Time Domain Reflectometry. A measuring instrument using this method is an optical pulse tester (OTDR).
Domain Reflectometer) has been put into practical use and is used for inspecting manufactured optical fibers and for construction and maintenance of optical fiber lines. FIG. 9 is an example of the backscattered waveform of the optical fiber under test observed by OTDR. The horizontal axis represents the distance z, and the vertical axis represents the logarithmically transformed backscattered light level P _B . The optical fiber to be measured is a fusion spliced optical fiber A with a length of 575 m, an optical fiber B with a length of 470 m, and an optical fiber C with a length of 1030 m, and has a total length of 2075 m. The width of the transmitted optical pulse was 1 μsec. 9th
The backscattered waveform in the figure continuously displays the power of backscattered light digitized for every 2m fiber length by A/D conversion. The backscattered waveform includes, from the shortest distance, Fresnel reflection at the input end, step 1 at the connection point between fibers A and B, step 2 at the connection point between fibers B and C, and Fresnel reflection at the far end. can be seen. The backscattered light power from an optical fiber varies depending on the fiber's scattering coefficient and relative refractive index difference, so if fibers with different characteristics are connected or the core axis of the optical fiber is misaligned,
Also, if the optical fiber is not manufactured uniformly, a step as shown in FIG. 9 will occur at the connection point or non-uniform point. Since the optical power of Fresnel reflection is much larger than the backscattered optical power, there is a steep difference in level. The distance measurement function of the OTDR is used to detect the location of the difference in level.
This is because when the OTDR operator sets a cursor on a stepped portion of the backscattered waveform, as shown in FIG. 9, the OTDR uses a function that automatically displays the position where the cursor is set. The true connection point or break point on the OTDR waveform corresponds to the position where the step starts to occur, so set the cursor at the position where the step starts to occur. In addition, since the slope of the backscattered power of each optical fiber in the backscattered waveform corresponds to the loss of the optical fiber, the loss of the optical fiber can be measured from the slope of the waveform in the section between cursors a and cursor b. can. Furthermore, by appropriately setting the cursor at five locations, the size of the step difference (corresponding to splice loss when optical fibers with the same characteristics are connected) can be measured. However, these features typically
This is achieved by the OTDR operator setting the cursor, and since the cursor setting position is likely to differ depending on the operator, it is difficult to reproduce the measurement of distance, optical loss, and step size. It had the disadvantage of being bad. For example, if the width of the optical pulse to be sent is 1 μsec, the distance between the steps is approximately 100
m, so depending on how you set the cursor, there will be a difference of about 100 m in distance measurement. In addition, in the case of an optical fiber to be measured that includes a large number of characteristic points, the cursor can be moved to one position in order to detect each step point and measure the optical loss and the size of the step.
Since it is necessary to set the cursor each time, it has the disadvantage that it takes time to set the cursor. On the other hand, in order to find the connection loss at the connection point and judge the quality of the connection, from the backscattered waveform data P _B (l ₁ ) and P _B (l ₂ ) of two points l ₁ and l 2 separated by _Δl , the following A method has been proposed in which the difference is calculated using the formula γ (l ₁ , l ₂ ) = P _B (l ₁ ) − P _B (l ₂ ) (1) and the connection loss is calculated from the magnitude of this difference. (Literature;
RDJEFFERY, JLHULLETT “Joint loss
measurement using two-point processing”,
Optical and Quantum Electronics, 16, pp.1
-7, 1984). For backscattered waveform data in which a step is seen due to a fusion splice point, etc., when a difference is calculated at a distance of Δl, the difference value in the section including the step will be larger than that in other sections. Therefore, it can be determined that a connection point or the like exists in a position section where the difference value is a predetermined value, and the size of the step can be determined from the size of the connection point. In this literature,
The connection loss was calculated assuming Δl = 160m, but the calculation of the connection point position was not considered. Therefore, the connection point is within a range of 160 m, and it has been impossible to detect the position of the connection point with higher accuracy using the conventional differential method. (3) Problems to be solved by the invention The conventional OTDR method has the problems mentioned above.
The present invention eliminates the need for the OTDR operator to manually set a cursor when detecting defect positions such as connection points and open end positions from the shape of the backscattered waveform.
The purpose is to provide a method to automatically and accurately set cursors in OTDR. (4) Means for Solving the Problems The present invention aims at injecting a light pulse into the optical fiber to be measured with a half-width τ, and calculating the optical power of the scattered light and reflected light from the light pulse that returns to the input end side. is received in time series, the optical power is electrically converted, logarithmically converted, A/D converted, and averaged, and the temporal change is calculated by z=vt/2 (where z is the distance from the incident end or position, v is the propagation velocity of the optical pulse in the optical fiber, and t is the round-trip propagation time).The backscattered waveform is determined by converting it into a distance change, and the defect in the optical fiber is determined from the shape of the backscattered waveform. In the method of detecting the position or the open end position, when two positions of the optical fiber to be measured whose section width including the position z is vτ/2 are z−z ₁ and z+z ₂ (z ₁ ≧
0, z ₂ ≧ 0, and z ₁ and z ₂ are constants that satisfy the relationship z ₁ + z ₂ = vτ/2), backscattered power P _B (z - z ₁ ) at position (z - z ₁ ) and position Find the difference between the backscattered power P _B (z+z ₂ ) at (z+z ₂ ) and
In a position range in which the difference value changes stepwise to positive or negative, the position where the difference value increases in a stepwise manner is defined as _zS , and the position further away by _z2 from _zS is defined as the defect position or open end position, and the difference value changes stepwise. In a position range that does not change in the same way, find the point z _P where the difference value is greater than a predetermined value and is the maximum or minimum, and set the position closer to the incident end by z ₁ than z _P as the defect position. It is characterized by detecting. The conventional technology depends on the sending pulse width τ.
The difference is that the difference in backscattered power between positions separated by Vτ/2 is determined, and the position where this difference value changes stepwise or the position where it is maximum or minimum is determined. (5) Effect When a light pulse with a half width τ is incident on the optical fiber to be measured, the light pulse is backscattered at the connection point of the optical fiber or at a position where Fresnel reflection occurs, and returns to the input end of the optical fiber. When calculating the difference in backscattered power between two positions z−z ₁ and z+z ₂ with interval width vτ/2, if the difference value does not change in a stepwise manner as in the case of backscattering at a connection point of an optical fiber, then Where the difference value is larger than a predetermined value, the position z _P where the difference value is maximum or minimum
exists, and it can be theoretically shown that this position z _P is farther from the input end by z ₁ than the true connection point. Therefore, the true connection point exists on the incident end side by z ₁ from the position z _P. In addition, if Fresnel reflection occurs, such as when the connection of the optical fiber by the connector is incomplete, when there is a break point, or when the far end of the optical fiber is open, the interval of vτ/2 If we calculate the difference in backscattered power with the width, we can see that the position where Fresnel reflection truly occurs is a position farther from the incident end by z ₂ than the position where the difference function changes stepwise. Therefore, find the position z _S where the difference function changes stepwise, and calculate the position z _S
The position further away from the incident end by z ₂ is the position where Fresnel reflection occurs. (6) Embodiment FIG. 1 shows a processing flow when detecting a defect position or an open end position of an optical fiber using the present invention. In the present invention, first, the difference between positions located ±vτ/4 apart from an arbitrary position z on the backscattered waveform is determined. That is, in this embodiment, z ₁ =z ₂ =
This is an example of vτ/4, and satisfies the relationship z ₁ +z ₂ =vτ/2. Here, v is the speed of light in vacuum divided by the refractive index of the optical fiber core, and is a constant. τ is a set value of OTDR and is a constant. Therefore, z ₁ and z ₂ are both constants. Next, it is determined how this difference value changes with respect to distance. If the change occurs in a step-like manner, the position farther from the change point _zS by vτ/4 is defined as the defect position or the open end position. The step-like change in the difference value is due to the sharp change in the backscattered waveform due to Fresnel reflection or fiber breakage, and the position where the change occurs due to the difference is vτ/4. Since there is a deviation, the defect position or open end position is determined by correcting the deviation. If the difference value does not change in a stepwise manner, find the position _zP where the difference value is maximum or minimum within the distance section where the difference value is larger than a predetermined value. Further, a position closer to the incident end than _zP by vτ/4 is defined as a defect position. These processes can be easily realized using a personal computer or the like. Next, we will explain the reason why the present invention calculates the difference between positions separated by ±vτ/4 around an arbitrary position z on the backscattered waveform. First, we will explain the backscattered waveform observed by OTDR. To analyze. Figure 2 shows the position z=vT in the optical fiber (where,
(v is the propagation time of a light pulse, T is an arbitrary time), that is, the time at which the backscattered light generated at time t=T reaches the input end. The horizontal axis indicates time t. Here, the backscattered light is generated in a pulsed manner corresponding to the transmitted optical pulse. Figure a
shows a situation in which backscattered light is generated in pulses at intervals of Δt near time t=T (T=z/v) when a light pulse with a pulse width τ is sent out. b
shows the status of the backscattered light pulse train when the backscattered light pulse train generated near time t=T reaches the input end near time t=2T. Since each pulse goes back and forth, each The pulse interval is 2Δt. From Figure 2, the scattered pulses of width τ that occur at t<T all return to the input end at t<2T, and the scattered pulses that occur at t>(T+τ/2) all return to the input end at t>2T. I know I'll be back. Therefore, when observing the backscattered pulse at the input end, at time t=2T, T<t<(T+
This means that parts of the backscattered light pulses generated during τ/2) are added. Also, time t
= (2T+τ), (T+τ/2)<t<
Parts of the backscattered light pulses generated at (T+τ) are added. Next, the scattered light from which position in the optical fiber is t
= Explain whether it is observed at 2T. Time t=T+
The time t' at which the light scattered at the position z = vT + z' reaches the input end at T' (where T' is 0<T'<τ/2 as described above) is t, taking into account the round trip. ′=T+T′+(vT+z′)/v =2T+T′+z′/v (2). Since we are now focusing on time t = 2T,
When t′=2T in equation (2), z′=−vT′ (3) holds true. In equation (3), since T' has the condition 0<T'<τ/2, the following equation holds true. −vτ/2<z′<0 (4) That is, at time t=2T, backscattered light from the range vT−vτ/2<z<vT (5) is added and observed. Similarly, at t=(2T+τ), backscattered light from the range vT<z<vT+vτ/2 is added and observed. This situation is shown in FIG. In the figure, a shows the position range where backscattered light is generated in the optical fiber, b shows the time when the averaged backscattered light power reaches the input end, and c shows the position displayed on the OTDR. In c, since the optical pulse goes back and forth, the time axis in b is halved. Third
From the figure, it can be seen that the backscattered power on the OTDR is the average backscattered power of the section width vτ/2 in the actual optical fiber. This indicates that the waveform observed by the OTDR matches the moving average of the backscattered light power obtained by a virtual optical pulse with an extremely narrow pulse width with an interval width vτ/2. Next, the section width for calculating the difference will be explained. FIG. 4 shows an ideal backscattered waveform on which no noise is superimposed in the optical fiber connection. Here, z′ is the position coordinate of the scattering point with the origin at the connection point, a ₁ and a ₂ are the optical fiber losses in the range of z′<0 and z′>0, and b is the loss of the optical fiber in the range of z′=0. This is the size of the step at the connection point. FIG. 4 shows the case where a ₁ , a ₂ and b are negative. The backscattered waveform has an interval width vτ/2
The waveform Y(z') is a moving average of ≡Δz (hereinafter also referred to as a moving average function), and can be expressed by the following equation. Y(z')=1/Δz∫ ^z ′ _z ′ _- 〓 _z ydz′ (6) a ₁ z′−a ₁ Δz/2 [z′<0] = a ₂ −a ₁ /2Δzz′ ² +( a ₁ +b/Δz)z′ −a ₁ Δz/2[0≦z′<Δz] (7) a ₂ z′+b−a ₂ Δz/2[Δz≦z′] Here, y is the original function Yes, for z′<0, y=
When a ₁ z' and z'>0, y=a ₂ z'+b. From formula (7),
Backscattered waveform Y(z') of the stepped portion caused by the connection
is a downward convex parabola when (a ₂ − a ₁ )>0, (a ₂ −
When a ₁ ) < 0, it becomes a convex parabola. Regarding the moving average function in equation (7), the function when calculating the difference with the interval width D is the difference function ΔY (z'),
ΔY(z′) is expressed by the following formula.

【table】
(9)
FIG. 5 shows the difference function ΔY(z') corresponding to the backscattered waveform (=moving average function) in FIG. 4. Here, the level difference at the connection point is negative as described above. Figure a shows the case where D<Δz. ΔY(z') consists of three straight lines and two parabolas connecting them, and the difference function has a minimum value at z'=Δz-D/2. Figure b shows the case where D=Δz. It can be seen that ΔY(z') consists of two straight lines and two parabolas connecting them, and takes its minimum value at a point shifted by Δz/2 from the true connection point z'=0. On the other hand, when the level difference at the connection point is positive, when Δz=D, ΔY(z') reaches its maximum value at a point shifted by Δz/2 from the true connection point z'=0. In other words, if the difference between two points is calculated from the OTDR backscattered waveform with interval width D = Δz = vτ/2, if the difference value does not change stepwise, the position where the difference value is larger than the predetermined value is determined. The position z _P where the difference value is the minimum or maximum within the range is found.
Since the true connection point is located at a position closer to the input end by Δz/2 than this position _zP , the connection point is a position closer to the input end by vτ/4 than the position _zP . Figure 6 shows cases when the optical fiber is incompletely connected by the connector, there is a break point, or there is Fresnel reflection, etc., such as when the far end of the optical fiber is an open end. This is a backscattered waveform, and the waveform changes sharply at z'=0. FIG. 7 shows this difference function. Since we are calculating the difference at an interval of ±D/2 from an arbitrary point,
The difference function changes stepwise from the position z'=-D/2. Therefore, it can be confirmed that the true defect position or open end position is a position D/2 further away from the position where the difference function changes stepwise. Therefore, in the case of D=Δz=vτ/2, it is necessary to correct the position to be vτ/4 further away. Next, for the backscattered waveform shown in Figure 9,
The results of applying the method of the present invention will be described. FIG. 8 shows a difference function obtained by changing the interval width D for calculating the difference for the backscattered waveform data shown in FIG. 9. In the same figure a, when D/Δz=0.4,
The figure b shows D/Δz=1.0, and the figure c shows D/Δz=1.0.
1.6, d in the figure shows the case when D/Δz=3.0.
In the example of FIG. 9, there is backscattered waveform data every 2 m, pulse width γ=1 μsec, and vτ/2=100 m, so 50 points of data correspond to vτ/2.
The horizontal axis in Figure 8 represents distance, and the vertical axis represents difference. Normally, in the backscattered waveform of an OTDR, it is assumed that there is a positive connection loss when there is a negative step, so the difference is shown in Figures 5 and 8. The sign has been changed from that shown in Figure 7. The difference value is negatively large near 600m, and is 1100
The positive increase near m is due to connection point 1 and connection point 2. Also, the step-like change near 2100 m is due to the influence of Fresnel reflection at the far end (open end). When the interval width D for determining the difference is small (D/Δz=0.4), the difference function contains many noise components. Section width D
As becomes larger, the noise decreases, but if the interval width becomes too large than Δz (D/Δz=3.0
), it can be seen that it becomes impossible to distinguish between changes in the difference value at the connection point and changes in the difference value due to fluctuations in the backscattered waveform. When D/Δz=1.0,
The difference function near the connection point is expressed by two parabolas, and the maximum and minimum points can be seen very clearly. It can also be confirmed that it is easier to find a connection point within a distance range where the difference exceeds a predetermined level than in other cases. Table 1 summarizes the results of detecting the positions of connection points 1 and 2 using the method of the present invention for the example shown in FIG. The predetermined level for connection point detection was the same regardless of D/Δz.
In addition, the difference from the actual optical fiber connection point position is also shown. When D/Δz=1.0, the difference from the actual connection point position is the smallest, confirming the validity of the method of the present invention. In this example, if the connection point could be detected, the calculation error of the connection point position was within 35m regardless of D/Δz, but in the case of D/Δz=0.4, the error was within 35m at the 1600m point where there was originally no connection point. the connection point and the error,
When D/Δz=3.0, the connection point at 575 m could not be detected. When D/Δz = 1.6, the connection point could be detected with an error within 22 m, but as can be seen from the difference function in Figure 6, the maximum and minimum values of the difference are smaller than when it is 1.0, so Depending on how it is set, there is a greater possibility of misjudging the presence or absence of a connection point compared to when it is 1.0. Therefore, the difference is determined by setting the division interval equal to Δz, that is, vτ/2, and the position where the difference value is maximum or minimum within the range where the difference value exceeds a predetermined level is determined.
According to the method of shifting the position by vτ/4,
It was confirmed based on actual data that connection points can be detected best. In addition, the position where Fresnel reflection occurs at the far end is 2070 m as a result of applying the method of the present invention, and the difference from the actual fiber length is 5
m, which confirmed the validity of the method of the present invention.
This means that the actual length of the optical fiber can be determined not only by measuring the location of defects in the optical fiber, but also by measuring the location where Fresnel reflections occur.

[Table] As described above, according to the present invention, it is possible to automatically calculate the defect position such as a connection point in the optical fiber to be measured. By automatically setting the cursor at vτ/2), it becomes possible to calculate the amount of step difference between defective positions and the optical fiber loss between defective positions in a short time. For an optical fiber sample to be measured containing three connection points, it took 4 seconds to automatically detect the connection points using the method of the present invention and further determine the amount of step and loss of each optical fiber.
On the other hand, when the OTDR operator set the cursor one location at a time and calculated the amount of step difference and optical fiber loss,
It took an average of about 7 minutes for 16 repeated experiments. From the results of this experiment, it was confirmed that the automatic detection of connection points using the method of the present invention is very effective in terms of time reduction. Further, according to the present invention, by measuring Fresnel reflection at the far end (open end), the actual length of the optical fiber can be measured with high accuracy. In addition, in the present invention, as shown in equation (8), the difference between positions separated by ±vτ/4=D/2 around an arbitrary point z is calculated, but as shown in the following equation, It is also possible to find the difference between point z and a point before or after point z, which is a distance of vτ/2=D. ΔY(z)≡{Y(z)−Y(z−D)}/D (10) ΔY(z)≡{Y(z+D)−Y(z)}/D (11) In this case, a true defect The position or open end position differs slightly from that previously described. In the case of equation (10), D/2= for the true position explained so far.
It is necessary to shift it toward the incident end by about vτ/4. Formula (11)
In this case, it is necessary to shift the position by D/2=vτ/4 from the true position described above.
Table 2 shows the positions where the difference function changes stepwise.
The correction amount for finding the true defect position or open end position from z _S or the position z _P where the maximum or minimum value is obtained is summarized. In Table 2, + corresponds to shifting farther than z _S and z _P , and - corresponds to shifting toward the incident end side. In general, when two positions including position z and having an interval width of vτ/2 are defined as z−z ₁ and z+z ₂ , the position z _S at which the difference function of the position changes in a step-like manner, or the maximum and minimum values are The true defect position or open end position can be determined from the position _zP taken as shown in the bottom column of Table 2.

[Table] (7) Effect of the invention As explained above, the position (z−
Find the difference between the backscattered power P _B (z - z ₁ ) at the position (z 1 ) and the backscattered power P _B (z + z ₂ ) at the position ₍ z + z ₂ ), and find the point where the difference value changes stepwise, or the difference The point where the value is greater than a predetermined value and is the maximum or minimum can be found, and the defect position or open end position can be detected based on this point. There is no need to detect the position or open end position,
There is an advantage that the defect position or the open end position can be detected automatically and accurately in a short time. Further, there is an advantage that the step amount at the defect position or the open end position and the optical fiber loss between the defect positions or the open end positions can be automatically calculated from the backscattered waveform data within a short time.

[Brief explanation of the drawing]

Fig. 1 is a flowchart when implementing the method of the present invention, and Fig. 2 shows a backscattered pulse train generated around time t=T and a backscattered pulse train generated around time t=2T when a pulse of pulse width τ is sent out. Figure 3 is a diagram explaining the relationship between the backscattered pulse train etc. arriving at the input end, Figure 3 is a diagram explaining the relationship between the position range where backscattered light is generated in the fiber and the position displayed on the OTDR, Figure 4 is Figure 5 is a diagram explaining the ideal backscattered waveform (=moving average function) Y(z') at the connection point, Figure 5 is a diagram explaining the difference function ΔY(z'), and Figure 6 is the diagram at the optical fiber break point. A diagram explaining the ideal backscattered waveform (=moving average function) Y(z'), Figure 7 is a difference function corresponding to Figure 6, Figure 8 is a diagram showing the actual difference function, Figure 9 is a diagram showing a backscattered waveform of an optical fiber observed with a normal OTDR.

Claims (1)

[Scope of Claims] 1. A light pulse having a half-width τ is incident on the optical fiber to be measured, and among the scattered light and reflected light from the light pulse, the light power returning to the input end side is received in time series, The optical power is electrically converted, logarithmically converted, and A/D
Transform, average, and calculate the temporal change as z=
vt/2 (where z is the distance or position from the input end, v is the propagation speed of the optical pulse in the optical fiber,
t is the round-trip propagation time) to obtain a backscattered waveform by converting it into a distance change, and detect the defect position or open end position of the optical fiber from the shape of the backscattered waveform. The width of the section including the position z of the fiber is
When the two positions where vτ/2 are z−z ₁ and z + z ₂ (z ₁ ≧0, z ₂ ≧0, and z ₁ and z ₂ are z ₁ + z ₂ =
constant that satisfies the relationship vτ/2), position (z−z ₁ )
Backscattered power P _B (z−z ₁ ) and position (z
+z ₂ ) and the backscattered power P _B (z + z ₂ ), and in a position range where the difference value changes in a positive or negative step-like manner, the position where the difference value increases in a step-like manner is set as z _S and the z _S The defect position or the open end position is defined as a position about z ₂ away from the position, and in the position range where the difference value does not change in a stepwise manner, the point z _P at which the difference value is greater than a predetermined value and is maximum or minimum is determined. 1. A method for detecting a defect position or an open end position of an optical fiber, characterized in that a position closer to the incident end by z ₁ than z _P is detected as a defect position.