Research on Improving Quality of Electricity Energy in Train’s Traction

To serve as one of the main transportations and carriers of the world nowadays, electrified railway is developing rapidly in its speed and length of line. In order to improve the power supply capability and power quality of traction system, scientific researchers have put forward many methods. Adding shunt capacitor compensator or series capacitor compensator in traction system are both effective methods. They can filter, regulate voltage, raise the utilization rate of the power supplying equipment capacity, and improve power factor.


Introduction
To serve as one of the main transportations and carriers of the world nowadays, electrified railway is developing rapidly in its speed and length of line. In order to improve the power supply capability and power quality of traction system, scientific researchers have put forward many methods. Adding shunt capacitor compensator or series capacitor compensator in traction system are both effective methods. They can filter, regulate voltage, raise the utilization rate of the power supplying equipment capacity, and improve power factor. This section will analyze the performance of power supplying capability increased equipment and the influence of each other, besides, study the strategy of coordinated control.

Shunt capacitor compensation of traction power supply system
Shunt capacitor compensation of traction power supply system means to connect a capacitor group and control device in parallel on feeder line or supply arm so as to improve power factor and power quality. Now dynamic capacitor compensation is used primarily.

Principle of shunt capacitor compensator
The connection and equivalent circuit diagram of shunt capacitor compensator is shown as figure 1, where 1 U  is supply voltage, and  1 +jW 1 is each phase impedance of power that converting into low-voltage side by internal impedance, line impedance and traction transformer. 2 U  is traction bus bar voltage of traction substation, X c is capacitor of shunt compensative capacitor banks, X L is inductance of reactor connected with capacitor group in series, Z is traction load impedance, I c is static var compensator current of shunt condenser, and I q is current of traction load. Before installing traction shunt capacitor compensator, current flow through the traction transformer is,  As two equations above indicates, after installing, the current flow through the traction transformer I q reduces to I, active power remains unchanged, and reactive power decreases.
It can be seen from 3 equations above, after installing, traction power factor can increase from cos 1 to cos 2 .
As figure 2 shows, 12  In Traction Power Supply system, when calculating and defining the capability of shunt capacitor, we should be based on the maximum average of traction load to choose shunt capacitor.

Dynamic shunt capacitor compensator
Nowadays in order to track changing traction loading timely, provide with rational reactive power compensation, and save the energy cost, dynamic compensator with shunt capacitor is widely used in electric railway. There are 3 commonly used devices.

Static var compensator (SVC)
The immediate purpose of parallel var compensator is to lower reactive current and negative sequence current. Three-phase SVC is an effective mature technology of improving harmonics, negative sequence and voltage fluctuation caused by the load of electric railway. SVC is by means of thyristor controlled reactance (TCR) and the strategy of three-phase balanced control based on "Steinmetz" to make dynamic reactive power compensation, control voltage, and improve negative sequence. This method can not only balance three-phase reactive power, but also balance three-phase active power. Therefore, it is an effective method.

Static synchronous compensator (STATCOM)
In contrast with SVC, STATCOM has its advantages of fast speed, great loading rate adaptation, high work efficiency, and small output harmonic content. Especially, adopting two-phase structure can achieve four-phase control of active and reactive power, provide two supply arms of power substation with dynamic reactive compensation, besides, regulate active flow of two supply arms, so as to dynamically balance the loading.
Since 1980s, the researches on the technology of STATCOM dynamic voltage compensation have become one of the hot topics in the field. Because STATCOM is usually supported by DC voltage provided by capacitor on DC side, it can't provide continues active power. But if change a power supply on DC side, STATCOM, served as voltage source inverter, can exchange energy with system. Connect a single-phase converter on both supply arms of Scott transformer, which are interconnected through intermediate links. The work mode of controlling two transformers makes them be DC power supply of the other one so as to achieve active power facility between both supply arms. When adopting impedance matching balance transformer or Scott wiring transformer, active load of both supply are balanced. Therefore it can eliminate the negative sequence completely. Accordingly, in order to achieve power factor compensation and negative sequence improvement of traction substations, we can install two single-phase STATCOM on traction transformer secondary side to carry on reactive compensation and phase transformations of active power.

Static var generator/ absorber (SVG)
Compared with traditional SVC, the regulation speed of SVG is faster, and the range of operation is wider. Besides, adopting multiplex, multi-level or PWM can significantly reduce the content of harmonics in compensation current. More importantly, because of the smaller size of capacity cell in SVG, the size and cost will reduce greatly. www.intechopen.com The basic principle of SVG is connecting self-communicated bridge circuit in parallel on grid or through a reactor, then regulating the phase and amplitude of output voltage of bridge circuit on AC side or controlling the AC current directly to make the absorption and output of reactive current meet the demands. Finally achieve dynamic compensation. Because SVG converts the DC side voltage to AC side output voltage which has the same frequency with the grid through the switching of power semiconductor, just like a voltage-type inverter, when considering fundamental frequency, SVG can be seen as an AC voltage source whose phase and amplitude could both be controlled. It is connected on power grid through AC reactor. The working principle is shown as figure 3. , and the voltage of reactance It is the phasor difference of U S and U I . The current I, flowing through reactance X, absorbed from power grid by SVG, is controlled by voltage of reactance. Hence converting the amplitude of U I on AC side and the relative phase with U S can convert the voltage of reactance so as to control the phase and amplitude of current absorbed by SVG, and the nature and size as well. Using appropriate control methods can control the currents running through each branch in three-phase bridge converter circuit individually, and make energy storage capacitor absorb energy from light load, and release energy to heavy load, so as to compensate for reactive power and negative sequence of traction load at the same time. If adopting PWM with appropriate testing method, it still can active filter.

Series compensator of traction power supply system
Series compensation of Traction Power Supply system is a capacitor group and control and protection device connected on feeder line or supply arm. Its function is to raise terminal voltage of supply arm and the power factor, then improve power quality.

The principle of series compensator of traction power supply system
Connect a capacitor in series on traction substation or power-supply section. The capacitor could offset part of inductance of Traction Power Supply system, and reduce voltage loss. The voltage loss is: Where  is the angle between load-terminal current vector and voltage vector.
The voltage loss is negative when load current running through series capacitor. It means output voltage of capacitor is greater than entrance voltage. The effect of voltage improvement has some connection with load current and power factor. The higher the load current, and the lower the power factor, the better the effect. The absolute value of voltage loss of capacitor is proportional to load current. It has automatic regulation to the voltage of supply arms, which is compatible with the drastic changes of the load of electric traction. The installed capacity of the series compensation is decided by its rated current and rated voltage. Series compensation is different from other electric devices for its working voltage is proportional to its current. Therefore, when choosing the content of the series compensation, the higher the rated current, the lower the working voltage. It means the lower the compensative voltage, the worse the compensative effect. That is the compensative effect is in inverse proportion to the installed capability The working voltage of series capacitor is proportional to its current. When choosing the series capacitor, the higher the rated current, the lower the working voltage. It means the lower the compensation voltage, the worse the compensation effect. The compensation capability is in inverse proportion to the installed capability. So the key point to meet the demands of compensation effect is to choose the rated current of series capacitor.
In order to meet the demands of compensation effect, the key point is choosing the rated current of series capacitor.

Thyristor controlled series compensation capacitor
Thyristor controlled series compensation capacitor includes Thyristor Switched Series Capacitor (TSSC) and Thyristor Controlled Series Capacitor (TCSC) commonly.
TSSC is consisted of a series of series capacitor. Each capacitor is connected with a transistor valve, which has a pair of anti-shunt transistors. TSSC adopts discrete step to increase or decrease the connected capacitor so as to control the compensation capacitor.
TCSC is a good way of series compensation, which is consisted of a series capacitor and a thyristor controlled reactor L. Actually, TCSC is also consisted of a protective device. It is shown as figure 2.
When the thyristor is turned off completely, the reactor is at non-conducting state, and TCSC manifests itself as a series capacitor compensator. When the conduction angle of thyristor is increasing gradually, the capability of the reactor of circuit branch and capacitor connected in parallel is increasing as well. When the reactor reaches the size of capacitor, shunt parameter will resonate, and the impedance of TCSC will be infinite. When the conduction angle of thyristor is increasing further, TCSC is manifesting itself as an inductive impedance, and the inductive impedance is decreasing gradually. When the thyristor is conducting completely, the impedance is the least. Therefore, TCSC can provide capacitive or inductive equivalent impedance within a certain range by controlling the conduction angle of thyristor.

Unified power flow controller (UPFC)
The UPFC consists of two voltage-sourced AC-DC converters connected back-to-back with some means (e.g., a chopper or specially designed1 converters) to permit interchange of power between the two converters.
Research work directed to analyze and build a scaled model of the UPFC has been sponsored by the Western Area Power Administration, the Electric Power Research Institute, and the Westinghouse Science and Technology Center. Refer to figure 7 for details.
Among the unique capabilities of the UPFC is the ability to control the flow of both real and reactive power on its series transmission line simultaneously and independently, as well as the ability to control bus voltage by operation of its shunt (STATCOM) element.

The interaction between devices
Take the interaction between Thyristor Switched Capacitor (TSC) and series compensation of Traction Power Supply system for example to explain the interaction between ordinary equipments of the system. The circuit of power-supplying capacity increased system of electric traction which has dynamic shunt compensation equipment and series compensation equipment is shown as figure 8. q I is the current of traction loading.   figure 8. Before installing the dynamic shunt compensation equipment, system current I=I q , and the system can be described as： 13 123 After installing TSC, following changes occur： a. The power factor changes from cos 1 to cos 2 .
The power factor increases from cos 1 to cos 2 after paralleling a capacitor X c .
Since TCS generates a compensative current contrary to the reactive component, the current flows through  1 +jX 1 is decreasing, and a voltage rise appears in 1 c. The reactive power absorbed from system by transformer increases.
It is shown as figure 9.
It can be seen from the equation above: The reactive power absorbed from system by transformer is proportional to its voltage. When the voltage increases from U 1 to U 1 , the reactive power absorbed from system by transformer increases as well.
The current flowing through the series compensation decreases, and it leads to decrease of the terminal voltage of supply arms.
The series capacitor X C can be calculated by the rated voltage of series capacitor U CN , the rated current I CN , and the power factor cos 1 . If the current flowing through the series compensation is I before installing the parallel devices, the compensative voltage is: After installing TSC, if the current flowing through the series compensation is I＋I C , because of the contrary direction of the reactive current I C and I we can get | I＋I C |＜| I|, so It means the compensative capacity of the series compensation will decrease after installing TSC, and can't meet the expectations of compensation effect.
e. The active power loss of the system increases.
From the respect of TSC only, the increased active power loss is:  is the dielectric loss angle of the capacitor.
From the above analysis, if seeking a coordinate control strategy without proceeding from the entire system, each control system will disturb others mutually, and then lead to the difficulty of reaching ideal operating condition.
Adopting TCSC and thyristor controlled parallel capacitor in Traction Power Supply system, and optimally distributing the capacity of TCSC and thyristor controlled parallel capacitor according to the real-time condition of electric traction can not only save power, improve the power quality of Traction Power Supply system, but also protect the equipments, increase the safety of operation of trains. Solving the problem of coordinately control the electric traction reasonably is meaningful for the development of Traction Power Supply system. www.intechopen.com

Common analytical methods of interaction
Traction Power Supply is a kind of flexible AC transmission mode. The technology of Flexible AC Transmission System (FACTS) is one of the most attracting directions of nowadays new technology of power system. It becomes an effective method of solving the problems of economic operation, security and stability in power system.
Nowadays in the field of interaction analysis of FACTS, mainly there are two kinds of methods: Simulation of nonlinear time-domain. This method can only observe some sever interaction through simulation waveforms, while can't obtain accurate quantitative analysis. Traditional modal analysis of characteristic root. This method can study the influence on system from FACTS. Its objects mainly focus on the changes of one or some oscillation modals. However, it can't obtain quantitative analysis of the degree of interaction among controllers.
In the last 30 years, research on interaction has attracted widely focus. There are many documents in the field of analysis of steady-state and dynamic interactive control.
Introductions of analytical methods of interaction are shown below.

Analytical method of modal
As a traditional analytical method, analytical method of modal has been widely used in controlled field. It also gets a lot of achievements in the research on the analysis of interaction Recent years, with the development of non-linear system theory and application of modern mathematic methods, as a successful mathematical analysis tool and linear modal analysis, Normal Form (NF) is widely used in the analysis of mode interaction, study on stability region of system, and the designation of controller. Reference [4] put forward an analytical technology to reduce the order of state equation of system to second-order equation near working point and revealed the interaction among modals. NF provides new way for site selection and designation of controller. Reference [5] adopted NF to predict the separating phenomenon of oscillation modals in regional system. It revealed the importance of the influence on the stability of system from non-linear factors. Reference [6] and [7] put forward a research method of using NF mathematical tools to analyze the interactions among controllers www.intechopen.com of FACTS. Linear analysis ignores the influence on system response from non-linear factors, and considers based on NF by preserving second-order equation. Then put forward a nonlinear interactive index to evaluate the intensity of interaction among SVC, SVC and TCSC, and UPFC. In the cases of study of multi-machine power system, the effectiveness of NF analytical method was verified through theoretical analysis and time-domain simulation.

Analytical method of relative gain array
Relative Gain Array (RGA), put forward by Bristol in 1996, is an effective method of analyzing the interaction of multi-variable control system. Also, it is a widely used designing tool of controlling system [8]. As for multi-variable control system, RGA is a method of providing optimal combination of control variable and controlled variable by observing the interaction among each input and output variables. It can also provide interaction messages among different control processes. Therefore, RGA is widely used now. Reference [9] studied the possibility of the application of RGA in static analysis of power system. It compared traditional residue, characteristic root and factor analysis method, indentified and studied the problem of site selection of PSS. Reference [10] studied the possibility of designing damping FACTS controller by improving RGA.

Singular value decomposition method
Singular Value Decomposition (SVD) Method is an important method of analyzing the interaction of input and output variables of system [11]. When apply SVD in a frequency range, the main point is, as opposed to original vector, singular vector decreases little, and the interaction of control circuit is small. The advantage of SVD is capable of dealing with time-lag system and non-positive-order system. But when the singular values are too close, SVD will be very sensitive or even unreliable.

NI method
NI method, put forward by Niederlin ski in 1971, is a controller matching method [12]. NI method is widely used in the choice of variable matching of control system just like RGA. Both methods are easy, and they can indicate the degree of coupling of control circuit only depend on object model. The standard of NI method based on variable matching was introduced in reference [13]. In reference [13], it said that NI value is an indicator of global interaction. And it can solve the problem of the choice of variable matching. Reference [14] and [15] put forward the combination of RGA and NI method. But commonly NI indicator is served as a tool of RGA, when analyze the interaction and pairing pattern among controllers. Use RGA to pair, then check if the system stable through NI method.

Analytical method of interaction based on gramian
Gramian is built on dynamic model of state space of system. It can describe the controllability and observability of one stable system, and it suits for continuous or discrete system. Take Gramian of system as quantitative analytical way to describe the signal of input and output as well as the controllability and observability of system [16]. Reference [17] took Gramian as representation of information content, combined with Homology group and information www.intechopen.com content, put forward an optimum distribution of rotor angle measurement device. Gramian is model based on state space of system, but when a system is represented by four matrixes (A, B, C, D), if feedback matrix of D exists, this method is not applicable.

The standard of Eigen value of Jacobian determinant
The standard of Eigen value of Jacobian Determinant, put forward based on difficulty analysis of calculating inverse matrix of steady state gain matrix G, can be used to guide variables pairing. Best match of variables can make single circuit independent. They are matches that off-diagonal elements of G have least inverse matrix effect on G. For a system that doesn't have interaction, inverse matrix of G is equal to the inverse of a matrix consisted of diagonal elements.
From Reference [18] and [19], we knew that when the inverse of G G 21 is represented by Jacobian and diagonal matrix, the necessary and sufficient condition of converging G 21 is that the value of max eigen value of Jacobian matrix must be less than 1. Therefore, the value of eigen value is the least among all possible pairings get from Jacobian matrix. But the difficulty of applying this method is that there are many pairings needed detecting, and the off-diagonal elements of G have significant influence on the process of inversing. So this method doesn't apply to predict pairing.

RHP zero method
RHP zero method can be used to guide variables pairing. Considering pairing different outputs and inputs, system would have different zeros. On some conditions, RHP zero exists. These zeros would make the control performance of closed-loop worse, which we should try to avoid when pairing controller.
From reference [12], we can get that zero would not change with feedback control, but the pole would change with feedback control. As the feedback control gain decreases, the pole of closed-loop would move to the position of the pole of open-loop. And it is going to make closed-loop system unstable possibly. Therefore, the principle of choosing pairing output and input is let closed-loop have least RHP zeros. Especially, we should avoid there are RHP zeros in the frequency region considered.

Coordinated control of interaction
Theoretically, the most effective method of solving the problem of interaction among FACTS controllers is to apply multi-variable coordinated control. Because a MIMO system is consisted of a flexible AC transmission system that has many control function (may be many FACTS devices), so a MIMO is the easiest method of solving the problem of interaction among controllers and the stability of closed-loop system. However, because of the complexity of multi-variable control, actually it's difficult to apply it in power system. Usually, we use coordinated control to solve the problem of the interaction among controllers.

Coordinated control of multi-objective of single FACTS
FACTS has many functions. Generally, controllers are designated aiming for different functions respectively. It makes every control function isolated or even contradicted, so it's www.intechopen.com time to consider the coordinated control among multiple targets. Reference [21] pointed out that STATCOM which has static parameters couldn't have both satisfying voltage control accuracy and damping control effect. In order to achieve the coordination between two targets, controller based on rules was designated. It means defining the structure and parameters of controller according to the operating condition.
Reference [22] introduced a new control method of multi-objective of FACTS. That is an intelligent control method of FACTS device. This method combined with the advantages of predictive control and inverse system control. The process of the signal of optimal control of FACTS has two sub-processes. First, filter the candidate outputs of FACTS device, and get the best output. Second, inversed calculate the actual control signal of FACTS device according to the best output. In every sub-process, Artificial Nervous Network Technique (ANNT) and fuzzy reasoning are used to solve different problems. Reference [22] took Advenced Static Var Generator (ASVG) as an example to introduce designation steps of intelligent predictive controller based on the method of intelligent predictive. Electromechanical simulation example proved the effectiveness of this new control method.

The theory of multi-objective evolution
In real world, most optimum problems are relative to multiple targets. Those targets are not alone. They are usually coupling and competing. Each target has different meanings and dimensions. The complexity and competitiveness would make optimization difficult [23].
The optimal solution of single target has clear definition, however, the definition couldn't promote to multiple targets. It's different from the definition of single target that multiobjective problem doesn't have global optimal solution, but it has a collection of global optimal solution. The optimal solution of multi-objective problem is called collection of global optimal solution, and the elements of it incomparable for global target.
Commonly multi-objective problem can be described as The essence of multi-objective is finding a group of decision vector which is meeting the demand of restraint condition so as to make target vector be the maximum or minimum at feasible region. It's different from optimal problem of single target that it's almost impossible that all target functions are maximum or minimum. Therefore, people put forward the concept of optimal solution of Pareto.

ZMOEA method
Revolution algorithm based on population can implicit parallel search for multiple solutions in solution space, and it can also improve the efficiency of calculating through similarity of different solution. The combination of revolution algorithm and the concept of optimal www.intechopen.com Pareto can produce real revolution algorithm based on the concept of optimal Pareto so as to search for the non-inferiority optimal solutions.
Using revolution algorithm to solve multi-objective problem is like traditional algorithms. Objective function must be scalarized. Fitness, served as the assessed value of individual of next generation, must be scalarized. The process of scalarization should monotonic transformation of coordinate of objective function so as to let the individual endow the best fitness of Pareto. This transformation isn't the only one. It concludes of the preference information of designer when he is assessing individual. Generally, if one scalared fitness is achieving, revolution algorithms use common selection method to progress [24]. There are 3 kinds of revolution algorithms to solve the multi-objective problem [25].
Quantitative method: Multiple targets usually convert to single one so as to optimize them. This kind of method includes weighted method, minimax method, and objective vector method and so on. These methods are usually the same as single target optimum method. Sorting method based on Pareto: Sorting population directly according to Pareto.

Steps of multi-objective evolution
Steps of Multi-Objective evolution are as follows.
a. Decompress a group which has N chromosomes initially. Coding chromosomes in floating mode. b. Decoding chromosomes and transferring it to real value of parameter region of controller optimized. Then substitute it into equation unknown. Sorting target values in Pareto way. Then calculating the initial fitness value. c. Choosing and copying, then producing new population. According to the difference among fitness values, choosing in the way of roulette, and generating new population. d. Crossover the individuals of new population, and mutate them. Increase the diversity of population. Interleaved mode adopts single-point crossover. And mutation can prevent algorithm from being optimum partially to some extent. e. Operating Elitism [26], and combining the population produced above with old population. Sorting them, and generating a new population. f. After evolution of one generation and producing next generation, go back to step B until it meet the termination conditions. Find out some optimum point of Pareto or the some close to it.
Termination conditions can be maxgenterm, or we can choose other indicators to judge restraint condition.

Multi-objective optimal design of TCSC and SVC
Among evolutionary algorithms, fitness function decides the evolutionary direction of population. In order to transfer coordination problem to multi-objective optimal design www.intechopen.com problem, we should choose a function consisted a series of performance index of system response. Choose control objectives of TCSC and SVC, it means active power of the line, signal of node voltage supported by SVC, and variance integral of reference value set by controller to be performance index. As mentioned above, the problem of multi-objective optimal design of TCSC and SVC can be described as: Where K P is scale factor of TCSC PI controller, K I is double integral TCSC PI controller, and K svc is the gain factor of SVC voltage regulator.

The coordinate control case of TCSC and SVC
As figure 7 shows, install TCSC and SVC in system. TCSC is installed on supply arms. The purpose of installing TCSC is to control active power of line by changing the reactance of line. SVC is installed on bus bar in parallel. The purpose of installing SVC is to make the voltage of installation point stable. Adopt multi-objective evolution algorithm to optimize 3 parameters of both controllers of FACTS Let the population size be 50, let maximum optimum algebra be 50, let crossover frequency be 0.8, and let mutation rate be 0.07. Figure 10 is convergence rate curve of MOEA, and figure 11 is non-optimal Pareto solution. It can be seen from figure 10. The progress ratio of MOEA is relatively high initially. After evolution, the progress ratio gradually tends to zero. It means population is close to Pareto optimal solution. As for Pareto optimal solution described by figure 11, according to the definition of Multi-Objective problem, first, the nonoptimum solution is a solution in feasible region. Second, when comparing any solution with other feasible solutions of non-optimal solution, at least one object is better than other feasible solution. Table 1 shows partial Pareto solution.  As you can see in figure 12, the coordinate design of TCSC and SVC is successful, and the difference between different groups of parameters of Pareto solution set is: one group of optimal solution has better control effect on TCSC and worse effect on the stability of voltage of SVC. The purpose of Multi-Objective coordinate designation is to analyze which control object is more important according to actual operation when the different objects of SVC and TCSC interact each other, and lead to the worst situation that both objects couldn't be optimum. Then choose a group of optimal solution among Pareto solution to solve the problem of coordinate operation after installing SVC and TCSC.
Now analyze the problem of coordinate control considering the least energy consumption. Active power losses of traction transformer are iron loss P 0 and copper loss. The iron loss is no-load loss, which is fixed, and has nothing to do with current flows through load. P 0 can be calculated by copper loss, which is called load loss or changeable loss, and changing with current flows through load. Active power loss of single-phase is: According to the principle of reasonably utilize the place and the analysis above, assume installing TSC and series compensation which are on 30-meter supply arm on the place at the distance of 8km from transformer, and use l and s to represent the distances of two sections. The high-voltage side and low-voltage side of transformer are 100kV and 27.5kV. Assume the capacitors of SVC and TCSC are X C1 , and X C2 . X C1 is a continuous variable, and X C2 is a undetermined variable. The energy consumption function of entire system is: From equation (13) and (14), we can get the objective function of system is: W=min{P 2 (X C1 , X C2 )+Q 2 (X C1 , X C2 )} 0.5

www.intechopen.com
The functions of restrain condition are： IUX  and I C =I q sinΦ 1 -IsinΦ 2 , we can get, U 2 =( I q sinΦ 1 -IsinΦ 2 )×X C =I q X C (sinΦ 1 -tgΦ 2 cosΦ 1 ) U 1 =U S -I q (R T +X T tgΦ 2 ) cosΦ 1 then, U S -I q [(R T + 1 ) +(X T + X 1 )tgΦ 2 ] cosΦ 1 = I q X C (sinΦ 1 -tgΦ 2 cosΦ 1 ) We can get from the equation above, 11  In conclusion, the capacity of compensator with parallel capacitor XC1 is decided by U 2 , and the capacity of TCSC is decided by X C2 . In return, after installing TCSC, the voltage drop caused by it subtracts the voltage drop caused by inductance of contact system. It could be considered that capacitor subtracts inductance. Because the voltage drop of supply arms caused by TCSC and the voltage drop caused by inductance cancel out, the voltage loss of system reduces obviously, and the power factor improves a lot.
Adopting TCSC and optimally distributing the capacity of SVC and TCSC according to the real-time condition of electric traction can not only save power, improve the power quality of Traction Power Supply system, but also protect the equipments, increase the safety of operation of trains. Solving the problem of coordinately control the electric traction reasonably is meaningful for the development of Traction Power Supply system.