Artificial Intelligence: An Advanced Approach in Power Systems

This article elucidates the application of Artificial Intelligence methods in power system expansion. - Rajesh Chourishi

New Products Information, Latest Technology, free resources, white papers, Planning & Design| energy & power industry | Artificial Intelligence: An Advanced Approach in Power Systems - Electrical India Magazine on Power & Electrical products, Renewable Energy, Transformers, Switchgear & Cables
Artificial Intelligence: An Advanced Approach in Power Systems

Early to mid-1980s, providing a solution to complex problems in many areas of power system engineering was tough and tedious. Presently with Artificial Intelligence (AI), many constraints can be handled easily such as economic load dispatch, load forecasting, optimisation of generation and scheduling, transmission capacity and optimal power flow, real and reactive power limits of generators, bus voltages and transformer taps, load demand in interconnected large power system and their protections etc. Now, most of the efforts in power system analysis have been successfully reduced by AI techniques.

Power Systems

Power system engineering is an important branch of electrical engineering that deals with the generation, transmission, distribution, and utilisation of electric power.

Artificial Intelligence

AI is the science of automating intelligent behaviours presently accomplishable by a computer interfaced with machines like robots. Artificial General Intelligence (AGI) is the intelligence of a hypothetical machine or computer which can accomplish any intellectual assignment successfully which a human being can accomplish.

Necessity of AI in Power Systems

For industrial development with power system expansion; stability, strengthening, reliability, technical advancements, selection and dynamic response of the power system are essential. With the growth of the power system, complexity in the networks is increased tremendously. As a consequence of this power system analysis by conventional techniques and conclusions from the acquired data, the process for the information, management of remote devices and utility became more complicated and time-consuming.

As necessity is the mother of invention, AI is developed with the help of sophisticated computer tools and applied to resolve all aforesaid problems for large power systems.

AI Techniques

Modern AI technologies include the following techniques:

  • Artificial Neural Networks (ANNs)
  • Expert System Techniques (XPS)
  • Fuzzy Logic systems (FL)
  • Genetic algorithm (GA)

These are the major families of AI techniques which are considered in the field of modern power system.

Artificial Neural Networks

Artificial Neural Networks (ANN) are biologically inspired systems. ANN mathematical models simulate the human biological neural network for processing information where each neuron produces one output as a function of inputs. Each type of neural network is capable of some specific work after being trained and is able to conclude a function from observations faced in real life such as function approximation, classification, data processing, etc. Its primary advantage is the capability to learn algorithms, an online adaption of dynamic systems, quick parallel computation, and intelligent interpolation of data. They are classified by their architecture, a number of layers, topology, connectivity pattern, feed forward, back propagation and radial basis function or recurrent, etc.

A neural network consists of some layers of artificial neurons linked by weight connections.

  • Input Layer: The input units do not process the data and information but distribute other units.
  • Hidden Layer: The hidden units provide the ability to map or classify the nonlinear
  • problems.
  • Output Layer: The output units encode possible values to be allocated to the case under consideration.

Figure 1: Architecture of a feedforward ANN

ANNs Characteristics

ANNs are fast and robust and do not need any appropriate knowledge of the system model. Since they are fault tolerant, they can handle situations of incomplete or corrupt data and information. They have learning and data adaptation ability. On the other hand, ANNs cannot perform a task other than the one for which they are trained. For any other task, they have to be retrained. ANNs always generate the result although the inputs data is unreasonable.

Applications

ANNs can be particularly useful for problems which require quick results, like those in real time operation. ANN techniques can be applied to power system protection.

Methodology

Real world problems in generation, transmission, and distribution of electricity can be fed to the ANNs to obtain a solution.

Figure 2: Fuzzification

Fuzzy logic

Fuzzy systems were developed in 1965 and had become popular in technical problem-solving. They are considered as mathematical means of describing ambiguity in linguistic terms instead of exact mathematical description.

Since it performs and can take a decision like a human brain, it can be standardised and systematised approximate reasoning. Therefore, with certain or even approximate information and data, it produces accurate solutions. Hence, this technology is used in machines so that they can perform like a human.

Fuzzy Logic Characteristics

Fuzzification provides oversimplification, superior expressive power, and an improved capability to model a complex problem at low cost. It allows a particular level of uncertainty throughout an analysis, as a consequence of allowed uncertainty it minimises problem complexity and available specifies information.

As most of the power system analysis is performed either with an approximation or with assumption-based data, Fuzzy logic can be of great use to derive a stable and exact output free from uncertainty.

Applications

Fuzzy logic has suitable applications in power system, like reactive power and voltage control, system stability analysis and control, fault analysis, security assessment, load forecasting, power system protection, etc. It can be used to increase the efficiency and for designing physical components of power systems from small circuits to large mainframes.

Methodology

Fuzzy logic provides the conversions from numerical to symbolic inputs, and back.

Expert systems

Expert systems were developed during the 1960s and 1970s and commercially applied throughout the 1980s. It is also called knowledge-based systems or rule-based systems. It is a computer program that incorporates knowledge derived from experts in a specific subject to provide problem analysis to users. This knowledge is generally stored in one of the many forms, like rules, decision trees, models, and frames. It uses this knowledge and interface mechanism to solve problems which cannot be or difficult to be solved by human skill and intellect. The common form of an expert system is a computer program containing the rules for analysis and recommendations for users.

Characteristics

Since expert systems are basically computer programs, it is based on the process of writing codes which is simpler than actually calculating and estimating the value of parameters. Therefore, any modifications even after design can be easily done. These systems are incapable of accepting new problems or situations other than programmed.

Applications

Expert systems are especially useful for problems when a large amount of data and information have to be processed in a short time. Many applications in power systems related to Power system designing and analysis match the abilities of expert systems.

Methodology

The methodologies of expert systems can be classified into the categories of rule-based systems, knowledge-based systems, neural networks, object-oriented methodology, case-based reasoning, system architecture, intelligent agent systems, database methodology, modelling, and ontology.

Expert systems are also combined with fuzzy systems to fuzzy-expert systems or combined with neural networks to neuron-expert systems. Recently, with the development of computer techniques, expert systems are also applicable to online applications of the power system. The structure of the expert system is shown in figure 3.

Figure 3: Structure of the Expert system

Genetic Algorithms (GA)

The Genetic algorithm gives a global technique based on biological metaphors. It is an optimisation technique based on the study of “Natural selection and natural Genetics.” Several methods for increasing the efficiency and analysis of power system to increase power output can be proposed, but out of these methods, Genetic Algorithms withstands all selected constraints. It is the best method for solving complex and nonlinear problems. it is used for planning of power generation, transmission and distribution. It adjusts the parameters of excitation to solve the voltage control problem and reactive power compensation.

AI – Exposure in Power System

Several problems in power systems cannot be solved by conventional techniques. Therefore, AI techniques in power system applications are being focused widely. Particular emphasis has been put on Artificial Neural Networks (ANN), Fuzzy Logic (FL) and Expert system (XPS). Some of the areas of the power system applications are highlighted here.

  • Economic load dispatch, generation and operational planning based on load forecasting, optimisation of hydrothermal generation scheduling.
  • Power transmission capacity and optimal power flow, real and reactive power limits of generators, and system reliability.
  • Control of voltage and frequency for system stability, sizing and control of FACTS devices.
  • Analysis of electricity markets and strategies for bidding.
  • Automation for power restoration and management, fault diagnosis, and security margins.
  • Planning and operation of distribution, network reconfiguration, demand-side response and management, operation and control of smart grids.

Some typical application of power system protection, ANN application, have been introduced to CT and VCT transient correction. For digital relays, fuzzy criteria signals, fuzzy settings, and multi-criteria decision making have been applied. FL and ANN application is applied for Differential protection for power transformers.

AI – Techniques for Transmission Line Performance Improvement

A practical application to improve the performance of transmission line is described with the help of a combination of AI techniques.

To improve the performance of a transmission line, the following functions are allotted to various AI techniques as:

  • Fuzzy systems: To diagnose the fault.
  • ANNs: Trained to change the values of line parameters based on environmental conditions.
  • Expert systems: To deploy outputs as a value of line parameters.
  • Environmental sensors: To sense the environmental and atmospheric conditions and provide input to the expert systems.

If any fault occurs in the transmission line, the angular difference between phasors of fault and pre-fault current is detected and fed to the fuzzy system for diagnosis.

The environmental sensors sense the environmental and atmospheric conditions as inputs to the expert systems. The expert systems provide the value of line parameters to be deployed as the output. ANNs improves and check the performance corresponding to the parameters provided by environmental sensors, if needed it changes the line parameters within the specified range to achieve the desired performance of the line.

Since the processing speed is directly proportional to the number of neurons, therefore, to improve the performance up to the desired level, a number of hidden layers and a number of neurons in each layer can be varied.

To acquire desired output, networks take different activation functions between input and hidden layer and hidden and output layer. Similarly, different neurons can also be taken for different layers.

AI – New Applications in Power Systems

Many problems in power systems are based on several non-feasible requirements. Therefore, AI techniques are the only option to solve them. Current approach of AI in power system applications are:

  • Planning for Generation expansion, power system reliability, transmission expansion, and reactive power.
  • Control of voltage, frequency and stability, and power flow.
  • Control of a Fuel Cell and thermal power plant.
  • Automation for restoration management, fault analysis and network security
  • Planning and operation of the distribution system, demand-side response and management, smart grids operation and control, and network reconfiguration.
  • Forecasting for electricity market, solar power, and wind power.

Saving Potential in AI

To avoid an impact on the environment, reliable and efficient power supply has become an important need of the world. This is being achieved by close monitoring of the power system equipment and consumption. It needs AI-based techniques which are highly reliable, accurate and automated systems such as EMS, Intelligent sub-station ornamented by high-speed protection, monitoring, and communication systems. With the promotion of these developments by AI techniques, savings can be achieved in the field of remote monitoring of equipment, operation, maintenance, and production. Plenty of research has been performed, and a lot of research is yet to be performed to derive full advantages of AI technology for cost reduction by improving the efficiency of the power system, distributed control and monitoring system, renewable energy resources system, and electricity market and investment system also.


7 COMMENTS

  1. It was a very nice idea! Just want say thank you for the data you have distributed. Just continue posting this kind of post. I will be your loyal reader. Gives Thank you again.

  2. Id wish to thank you for the efforts you have made in writing this article. I am hoping the same ideal function from you inside the long term also. In reality your creative writing abilities has inspired me to start my very own BlogEngine blog now.

  3. I was researching on the net for some information since yesterday night and I at last found what i was looking for! This is a impressive web site by the way, however it appears a little difficult to read on my verison phone.

Leave a Reply