STATISTICAL NORMALIZATION OF THE ATTENUATION OF OPTICAL FIBERS IN THEREGENERATION SECTIONS OF OPTICAL-FIBER TRANSMISSION-SYSTEMS

The main primary communications networks in the future will be transmi ssion lines based on optical-fiber cables. The signal transmission qua lity in optical-fiber channels of transmission systems is governed to a considerable extent by the parameters of the regeneration sections, standardized during construction and operation. When standardizing the most important parameter, namely, the attenuation of the optical powe r, a fundamental difficulty arises due to the fact that the attenuatio n factor of an optical fiber alpha and the losses at the joints betwee n optical fibers alpha(j) are random. The transmission parameters of c ables with copper strands also have random production spreads, but the y do not exceed 1-2% of the nominal value and, as a rule, are ignored. It is not possible to ignore the random nature of the attenuation fac tor of an optical fiber since here the maximum acceptable values excee d the minimum values by a factor of two or more. The losses at the joi nts of optical fibers are even more random; their values are scattered over a range from zero to the value of the attenuation of half a cons tructional length of the cable. Nevertheless, the problem of taking th e random nature of the losses of optical power into account has not be en solved, and in the CCITT Recommendations [1-3], which reflect prese nt-day technical experience, it is merely acknowledged and recommended for further investigation. In the present paper we show that it is be st to carry out a statistical normalization of the attenuation of an o ptical fiber in the regeneration sections of an optical-fiber transmis sion system. A constructional solution of the problem is obtained usin g the central limit theorem of probability theory [6], which opens up the possibility of determining the distribution of the sum of random q uantities [7, 8]. In addition, as is also acceptable when considering random quantities, an appropriate value of probability is assigned to each range of values of attenuation.