Enhancing Power Flow with Heuristic Techniques

CAREER EPISODE 1: The Optimization of Power Flow Systems

A. INTRODUCTION

CE 1.1

I have elaborated the nature of work required to ensure the completion of the “Optimal Power Flow design for Power System using Heuristic Technique” through this part presented here. I did this project while I was studying at A.U College of Engineering, Andhra University, pursuing my degree in this. I initiated this project on …… and it took me about …. to push it to the completion.

B. BACKGROUND

CE 1.2

In the process of transmission of electric power where much power is lost, it’s very crucial to come up with a solution to facilitate optimal power flow to obtain minimization of fuel cost and minimization of loss of active power. This very concern and solution were addressed in this project, where I used Genetic Algorithm (GA) as heuristic technique for the OPF solution. In order to obtain the power system’s optimal settings, the fuel cost minimizing approach was taken as objective domain. Similarly, loss of active power was considered as objective function for optimization of reactive power. This technique was implemented using TCSC (Thyristor Controlled Series Compensator) device which was used for relieving of lines overload in case of heavy loading scenarios.

CE 1.3

To make the project successful, I worked on below-mentioned objectives:

• To increase the transmission power to a particular place with the use of TCSC.
• To improve in system stability, power balance, reduction in transmission losses using FACTS device.
• To increase the safety, controllability, and flexibility in the transmission of electricity with the use of the TCSC.
• To reduce the cost of fuel using the TCSC circuit which minimizes the loss of electric power.

CE 1.4

This was my academic project which I did on an individual basis, however under the guidance of my project guide, so the entire associated works of project solely rested on my shoulder of responsibility. I started with preparation of project proposal after brainstorming and literature review during the conceptualization phase of the project and after the approval, worked on the design of the power system with OPF solution, carried out modeling of GA-OPF and TCSC with objective functions as minimization of fuel cost and loss of active power. I documented the study results, prepared the report and submitted it to the department.

CE 1.5

I have depicted my position in the given figure.

Fig: Hierarchical Structure

CE 1.6

The list of activities I put to execution in the process of achieving the end result of the project are mentioned below in brief points:

• To study the transmission of electric power from one place to another place.
• To implement the Genetic Algorithm for collecting information about the cost of fuel and loss of active power
• To use the 75-bus and 30-bus system in the GA method to read about the fuel cost and active power loss.
• To implement the TCSC in the power transmission system connecting inductor and thyristor in series and then again connected in the parallel with the conductor.
• To use 75-bus and 30-bus system in TCSC method for the optimization of cost and minimization of power loss

C. PERSONAL ENGINEERING ACTIVITIES

CE 1.7

I conducted a study on OPF (optimal power flow) solutions, conventional optimization techniques for addressing OPF problems, the shortcomings of these approaches and the need for efficient techniques like heuristic algorithms with objective functions of different types. I implemented the GA method to collect the information about cost of fuel and the loss in active power. I used the 30-bus system and 75-bus system to collect information about power loss and cost of fuel. I also implemented the TCSC method which was the combination of inductor and thyristor in series and both were connected in parallel to form the TCSC circuit. Then, I studied about the objective functions for both 30 and 75-bus system using TCSC in the transmission of the electric power. Then, I compared both systems and found that the TCSC method was effective for power transmission.

CE 1.8

I made use of theoretical as well as practical knowledge imparted to me from the electrical engineering subjects to execute this project. I utilized my knowledge of power design, power transmission system, power control system, electrical devices, and circuitry, etc. for the designing of power system with optimal power flow with the use of heuristic techniques. In this transmission path, the impedance can be altered and affects the power flow. With the change in the impedance, the losses of the electric power can be minimized to zero. I referred to the IEC 60700-1:2008 standard specification for the high voltage direct current power transmission in the Thyristor. I also referred to CFR 1910.269 standard for the electrical power generation, distribution, transmission and addressing control of hazardous energy sources of power plant site.

CE 1.9

CE 1.9.1

At first, I carried out the research and literature review on the transmission of electric power from one place to another. For the formulation of OPF solution, I considered the objective functions as total fuel cost as , where NG is the count of generators and f_i is the generator’s cost of fuel, given by quadratic equation. Likewise, for minimization of loss, I took objective function as , where x and y are continuous and discrete variables respectively and nline gave total branches’ number. I defined the real power and reactive power balance equation associated with each bus for defining equality constraints and for inequality constraints, I defined branch flow limits, limits of load buses’ voltage, slack power and reactive power of generator limits. Then, I moved on with OPF solutions with minimization of objective functions that were non-linear. I applied the Genetic Algorithm for study of the Optimal Power Flow which gives information about the power transmission to distant places.

CE 1.9.2

I implemented the random generation method as optimization technique to carry out the search randomly where the data obtained from the search could be used for the next search. I used the random generation method to obtain search points in large numbers as from the overall search space. All the search point was used as the parameters of the problem in the search space. The set representing the search points obtained and utilized for the functioning is termed as population. I evaluated the population implementing GA for formation of new population and carried out the evaluation unless global point was obtained. I carried out implementation in three phases to obtain the global optimum point in the genetic algorithm.

CE 1.9.3

In the generation phase, the number of chromosomes (total genes – coded bits of each parameter of search space, of entire parameters) was generated same as population size. The population size had effect on efficiency and performance of GA. I formed the higher size of the population and string length which results high accuracy and better resolution of GA method study. After calculation of Psp(i) as well as Qsp(i), I formed Y-bus using sparsity and generated chromosome’s initial population. I set the count for chromosome and generation and calculated the active power and magnitude of voltage outputs after the chromosomes were decoded. Then, using these new values for active power, I calculated Psp(i) and ran NR power flow.

I calculated the bus powers and line flows and checked the above-mentioned limits i.e. voltage limit for load bus, reactive power, etc. After that, I calculated the fitness function for the evaluation of initial population and determination of the optimization problem. I calculated fitness for all chromosomes using 100/minf. Then, once the chromosome counts exceeded population size, I arranged them in descending order with fitness value as reference. For the selection of parents, I applied roulette wheel reproduction process after copying chromosomal probability to next generation. The next generation’s children were obtained after which the mutation was carried out and replacement of old with new population was done. After the converged iteration count, the optimized values were obtained.

Fig: Block diagram of the GA method

CE 1.9.4

After that, the GA applied the genetic operation in the fitness function if the resulted value was not satisfied. Then, I implemented the reproduction operator for transferring the number into the next generation which was fit for the accurate result. Next, I used the crossover operation for selecting the exact value of the population. If the value didn’t meet the population value of the particular place, then the operation could be repeated to obtain a higher number of the population in the search space. Then, I used the IEEE 30 bus type system for minimization of loss of active power of OPF using the parameter of population size was assumed= 40 and the maximum number of generations was assumed to be 100. I also considered elitism probability as 0.15, cross-over probability as 0.95 and mutation probability as 0.001.

Generator bus count	Active power obtained (in MW)	Voltage magnitude (p. u)	Cost of fuel ($/hr)
1	108.60	1.081	260.20
2	49.10	1.070	125.10
5	35.00	1.019	110.05
8	29.10	1.062	98.20
11	15.30	1.079	53.12
13	53.10	1.025	224.2

I obtained the active power loss and fuel cost using GA are 5.70MW and 870.87$/hr respectively.

CE 1.9.5

I also implemented a 75-bus Indian study system for the objective functions by using the parameters of as 40 for population-size, generated maximum number was assumed to be 100, elitism probability as 0.15, cross over probability was assumed as 0.001. The resultant table is given below.

Generator bus count	Active power (p. u)	Voltage magnitude(p. u)	Cost of fuel ($/hr)
2	1.890	0.960	140.055
3	1.600	0.945	225.50
4	0.823	1.088	269.45
5	0.875	1.0030	709.10
6	0.940	1.0040	2305.0
7	0.880	1.019	384.20
8	5.145	1.072	517.20
9	2.300	1.044	54.40
10	1.920	1.025	259.90
11	2.140	1.056	290.00
12	8.500	1.066	345.10
13	1.380	1.0090	32.20
14	3.250	1.070	365.25
15	7.99	1.0499	141.00

I performed the calculation of the active power loss and fuel cost using GA are 6038.355 $/hr and 175.05MW respectively.

CE 1.9.6

Next, I made the study of Optimal Power Flow using FACTS devices like TCSC (Thyristor Control Series Capacitor) for the increment of dynamic stability in transmission of power and improvement in the regulation of voltage. The FACTS controller was used to transfer power closer to the thermal rating. I also observed that the TCSC circuit is the combination of inductor (L) with the series combination with Thyristors. I linked the series combination of inductor and thyristor with the capacitor in the parallel series. The change of the impedance value of TCSC was achieved with the change of thyristor controlled inductive reactance of inductor linked parallel to the capacitor. I made the combination of different components as stated above for the improvement in angular and voltage stability.

CE 1.9.7

After that, I made case study using IEEE 30 bus system in TCSC using all the parameters in the GA method. I obtained the OPF result of IEEE 30 bus system for optimized constraints as:

Generator bus count	Active power (MW)	Voltage Magnitude(p. u)	Cost of Fuel ($/hr)
1	92.20	1.0140	216.30
2	45.90	1.030	121.20
5	34.84	1.0165	110.70
8	26.45	1.045	90.75
11	17.50	1.0755	56.30
13	50.10	1.075	214.20

I performed the calculation for the loss of active power and cost of fuel using TCSC using 75-bus, which were 5.40MW and 809.95$/hr respectively.

I also made the study of OPF system using a 75-bus Indian system for the optimization parameters in TCSC by using the same parameters used in the GA method.

Generator bus count	Active power (p. u)	Voltage magnitude(p. u)	Cost of fuel ($/hr)
2	0.799	1.0060	256.50
3	1.450	1.0140	209.00
4	1.980	1.0430	270.10
5	0.978	1.0680	708.10
6	0.70	0.960	305.10
7	0.980	1.079	384.10
8	5.454	1.0129	600.13
9	1.75	1.038	294.25
10	1.052	1.035	260.15
11	1.920	0.98	254.01
12	8.80	0.899	1120.25
13	1.080	1.070	213.25
14	3.50	0.98	503.95
15	8.00	0.950	566.10

The values for loss in active power and the cost of fuel were 154.52MW and 5944.99$/hr respectively. I found that objective functions using TCSC in 30-bus system were reduced from 870.87$/hr to 809.95$/hr for cost of fuel, and loss in active power decreased from 5.70MW to 5.40MW. And, for TCSC in 75-bus system, the cost of fuel was reduced from 6038.355$/hr to 5944.99$/hr, whereas the loss in active power was found to decrease from 175.05MW to 154.52MW.

CE 1.10

TECHNICAL PROBLEMS AND SOLUTIONS

• In case of the implementation of OPF using a genetic algorithm without use of TCSC for the case study of IEEE 30 bus system, I was unable to obtain the desired optimum value for the voltage magnitude, generator’s active power, etc. with reference to the result of case study. To know where I went wrong, I checked the algorithm from the beginning but also couldn’t trace the error in the implementation. I sought the help from my guide and was told that there was a missing step in the algorithm and when I rechecked the process, I found that I missed to implement a conditional function for carrying the uniform cross over in case the r (a random number in between (0,1)) was less than the cross over probability for obtaining next generation’s children item.
• I encountered the rise of the fuel cost in the GA method for power transmission. To overcome the drawback, I implemented the use TCSC method which controlled the power regulation and minimized the power loss and fuel cost. In TCSC circuit, I first connected the inductor parallel to the Thyristor due to which the power regulation was not obtained. To overcome this problem, I connected the inductor and thyristor in the series form which helped in power regulation and maintained stability in the power flow.

CE 1.11

CREATIVE WORK

I used a random search technique instead of going with other optimization techniques like Calculus-based method and Enumerate method for obtaining the optimized power flow solution. The genetic algorithm falls under random search technique, which I chose because it works with parameter coding, population point instead of single point. It could be also expressed in objective function and probabilistic form tradition rather than the deterministic rule.

CE 1.12

To make the project successful, I coordinated with the project guide, Head of the Department and faculty members to share my opinion about the project work to be carried out. I prepared the working schedule of the project and performed according to it, for executing the project at the right time. I informed about the progress I made in the project to my guide weekly and took his help in solving the issues. I made detailed report on the project mentioning the procedure of the project and submitted to the project guide.

D. SUMMARY

CE 1.13

I applied the genetic algorithm (GA) to optimize cost of the fuel and give solution for active power loss. The information obtained from the GA method and TCSC method using a 30-bus and 75-bus system were compared. The effects of FACTS devices like TCSC helped in minimization of the cost of fuel and loss in active power in the power flow system.

CE 1.14

First, I implemented the GA method to study fuel cost and the loss of active power during the transmission of power. I implemented the parametric constraints for the search points in the search space. Then, I implemented the use of the FACTS device like TCSC and make comparison using the 30-bus and 75-bus system. I found that the TCSC system was effective to control the voltage regulation due to which fuel cost and active power loss were minimized.

CE 1.15

I took part in performing the project with great determination. I applied my theoretical engineering knowledge in the project work and made my knowledge regarding power design effectiveness better. The project helped me to develop my skills and confidence to do works in effective ways. It helped me to establish my managerial skills which could be utilized in further works in daily life.