【正文】 weight into account. The train schedule or the positioning rules are also defined. Modelling the normal mode Let us first consider modelling single track railway sections. Note the main parameters that should be determined by the working model: * Maximum traction work currents * Maximum voltage losses The current in traction work should not exceed the bound permissible by heating norms. Voltage losses should not be too large because train speed depends on the voltage. According to the Russian standards, if the voltage in traction work is kV, minimal possible voltage on a lootive should be 21 kV (which means the losses should never exceed kV) On design stage the type and diameter of a traction work is determined according to these three criteria: * Costs * Maximum traction work current 6 * Maximum voltage losses It is mon to start the research from the least possible traction work diameter. Then it is increased until all three choice criteria are met. On this stage a special train positioning is used: a pack of cargo trains is formed with a minimal possible gap in between and sent to one direction. Most cargo trains in this pack have average weight. but some of them (usually one or two) have maximum weight allowed for the inter substation zone. It should be noted that the number of trains in a pack, as well as the number of heavy trains is determined by a special standard document that fixes their dependency from the overall number of trains on that particular direction. This document is always a very important part of research on Russian railway power supply systems design (see Table 1). Number of trains in a day Time interval (minutes) Number of cargo trains in a pack Cargo Passenger Up to 24 Up to 20 20 2 More than 20 15 3 2436 Up to20 12 4 More than 20 10 6 3748 Up to 20 9 5 More than 20 8 7 4972 Up to 20 8 7 More than 20 6 10 More than 72 Up to 20 7 8 More than 20 5 12 Table 1: Determining the number of cargo trains and the time interval 7 between subsequent trains in a pack. Additionally, the model takes care of: * different ordering of average cargo trains and heavy cargo trains * displacement interval (the time difference between arrival on the inter substation zone of trains from one direction and the other) * feeder length (the length of the feeding wire which connects railway substation with traction work) It is worth mentioning that some railway sections place substations on some distance from the railway. If the length of the feeders is long enough to influence the voltage on the lootive, this should be taken into account. It is also possible to have one railway substation placed at a distance with a feeder length of several kilometres, while the other one is placed directly on a railway. This leads to the currents of the substation with a shorter feeder length to be always greater than the currents of the substation with a longer feeder. In order to pensate the currents it is possible to artificially increase the length of a feeder on the substation which is too close to the railway (Figure 3). Figure 3: Artificial increase in feeder length. Taking all section peculiarities into account is a distinct feature of our model: it also allows to change the length and resistance of a feeder. In order to pletely simulate trains moving on a real life railway section, the model utilises data about points where the trains ing from 8 different directions can pass by each other. During simulation railway substation currents and voltage losses are calculated every minute. Later their maximums are determined and produced. Modelling the breakdown mode If one of the railway substations goes down, it plicates the situation for power supply system significantly. The main goal is then to not let breakdown mode bee emergency mode. The biggest plication is that in AC systems the inter substation zone can only be fed from one side (Figure 4)adjacent inter substation zones are fed from different phases. Figure 4: