DEVELOPMENT OF A STATE ESTIMATION BASED IMPROVED DETECTION AND LOCALIZATION OF NON-TECHNICAL LOSSES USING SMART METER MEASUREMENTS

ABSTRACT

This research work presents the development of branch current based state estimation for NonTechnical Losses (NTLs) Detection and Localization. The use of weighted least square (WLS) state estimation for the evaluation of branch current of a network during theft is considered. In order to confirm the presence of theft in a network, current measurement value obtained from Distribution Transformer Controller (DTC) installed at substation was compared with that of all customers’ smart meters readings, a difference above an estimated threshold signifies the presence of theft. For the case of locating the point of theft, the concept of weighted least square state estimation was used for the evaluation of the actual branch current of each branch of the network despite theft, the estimated branch current is compared with the calculated branch current based on meter reading, and the difference is exploited in order to locate the point of location. The developed method was implemented on a 415V Low Voltage network used in this literature. The results obtained were validated by comparing it with the work of Marques et al., 2016. All modelling and analysis were carried out using OPENDSS and MATLAB R2015a. From the results obtained, when the total theft in the network is 30%, 40% or 50% the maximum variation of the estimated branch current are 0.62%, 0.83%, 1.02% respectively, these are taken to be the threshold for decision of theft in the network. It was also observed that the True Positive Rate (TPR) and The False Positive Rate (FPR) irrespective of the percentage of theft in the networkshow an improvement of 27.5% and 11.11% respectively. The method was further compared with other works where the used of machine learning was exploited. This method shows 7.5% improvement in terms of TPR than the use of Artificial Intelligent method.

 

 

 

 

TABLE OF CONTENT

 

DECLARATION ………………………………………………………………………………………………………………………………….. ii

CERTIFICATION ……………………………………………………………………………………………………………………………….. iii

DEDICATION …………………………………………………………………………………………………………………………………… iv

ACKNOWLEDGEMENT ………………………………………………………………………………………………………………………. v

ABSTRACT ………………………………………………………………………………………………………………………………………. vi

TABLE OF CONTENT ………………………………………………………………………………………………………………………… vii

LIST OF TABLES ………………………………………………………………………………………………………………………………… x

LIST OF FIGURES ……………………………………………………………………………………………………………………………… xi

LIST OF APPENDICES ……………………………………………………………………………………………………………………….. xii

LIST OF ABBREVIATIONS …………………………………………………………………………………………………………………. xiii

CHAPTER ONE ………………………………………………………………………………………………………………………………….. 1 Background to the Study …………………………………………………………………………………………………………………… 1

1.2 Motivation …………………………………………………………………………………………………………………………………. 2

1.3 Significance of Research ………………………………………………………………………………………………………………. 3

1.4 Statement of Problem …………………………………………………………………………………………………………………. 3

1.5 Aim and Objectives …………………………………………………………………………………………………………………….. 3

1.6 Methodology ……………………………………………………………………………………………………………………………… 4

CHAPTER TWO …………………………………………………………………………………………………………………………………. 5 LITERATURE REVIEW ………………………………………………………………………………………………………………………… 5

2.1 Introduction……………………………………………………………………………………………………………………………….. 5

2.2 Review of Fundamental Concept ………………………………………………………………………………………………….. 5

2.2.1 Electric power system …………………………………………………………………………………………………………… 5

2.2.2 Technical characteristic of low voltage distribution network ……………………………………………………… 5

2.2.3 Power flow in low voltage networks ……………………………………………………………………………………….. 7

2.2.4 Distribution power system losses …………………………………………………………………………………………. 10

2.2.4.2 Non-Technical losses …………………………………………………………………………………………………….. 11

2.2.5 Analysis of non-technical losses ……………………………………………………………………………………………. 11

2.2.6 Description of methods for detection and location of NTLs ……………………………………………………… 14

2.2.6.1 The Artificial Intelligent Methods (AIM) ………………………………………………………………………….. 15

2.2.6.2 The Smart Metering Based Methods ………………………………………………………………………………. 16

2.3 Smart Grid ……………………………………………………………………………………………………………………………….. 17

2.3.1 Advanced Metering Infrastructure (AMI) ………………………………………………………………………………. 18

2.3.2 Smart Meters (SMs) ……………………………………………………………………………………………………………. 19

2.3.3 Distribution Transformer Controller (DTC) …………………………………………………………………………….. 21

2.3.4 Data Concentrator ………………………………………………………………………………………………………………. 22

2.4 State Estimation (SE) …………………………………………………………………………………………………………………. 23

2.4.1 Challenges of state estimation in distribution systems ……………………………………………………………. 23

2.4.2 State estimation in low voltage network ……………………………………………………………………………….. 25

2.4.3 WLS Estimator ……………………………………………………………………………………………………………………. 25

2.4.3.1 WLS State Estimation Algorithm …………………………………………………………………………………….. 27

2.5 Monte Carlo Simulation …………………………………………………………………………………………………………….. 28

2.5.1 Characteristics of Monte Carlo ……………………………………………………………………………………………… 28

2.5.2 Steps Involve in the Monte Carlo Simulation. …………………………………………………………………………. 29

2.5.3 Monte Carlo simulation procedure in Branch Current State Estimation (BCSE) Method. ……………… 29

2.6 Description of the Reference Algorithm and the Proposed Algorithm. …………………………………………….. 30

2.7 Case Study ……………………………………………………………………………………………………………………………….. 32

2.7.1 Description of Case Study ……………………………………………………………………………………………………. 32

2.8 Review of Similar Works …………………………………………………………………………………………………………….. 34

CHAPTER THREE …………………………………………………………………………………………………………………………….. 39 MATERIALS AND METHODS …………………………………………………………………………………………………………….. 39

3.1 Introduction……………………………………………………………………………………………………………………………… 39

3.2 Materials …………………………………………………………………………………………………………………………………. 39

3.2.2 Softwares …………………………………………………………………………………………………………………………… 39

3.2.2.1 Matlab 2015a software …………………………………………………………………………………………………. 39

3.2.2.2 OPENDSS software ……………………………………………………………………………………………………….. 39

3.2.3 The test case feeder. …………………………………………………………………………………………………………… 40

3.3 Methodology ……………………………………………………………………………………………………………………………. 40

3.3.1 Detection Algorithm ……………………………………………………………………………………………………………. 40

3.3.2 Development of power flow algorithm based on thevenin’s and norton’s equivalent circuit

approach in distribution network. ………………………………………………………………………………………………… 41

3.4 Development of State Estimation Algorithm ………………………………………………………………………………… 42

3.4.1 Formulation for branch current based state estimation ………………………………………………………….. 42

3.4.1.1 Basic WLS Formulas ………………………………………………………………………………………………………. 42

3.4.1.2 Measurement equations and jacobian matrices ………………………………………………………………. 43

3.4.1.3. The Jacobian Matrix. (H(x)) …………………………………………………………………………………………… 45

3.4.1.5 State Estimator Accuracy ………………………………………………………………………………………………. 48

3.4.2 Calculation of Branch Current ………………………………………………………………………………………………. 48

3.5 Developed Method for the NTLs Location ……………………………………………………………………………………. 49

3.6 Scenario Considered ………………………………………………………………………………………………………………….. 51

3.7 Modification of the Test case Feeder. ………………………………………………………………………………………….. 52

3.8 Performance Evaluation. ……………………………………………………………………………………………………………. 52

CHAPTER FOUR ………………………………………………………………………………………………………………………………. 54 RESULTS AND DISCUSSION ………………………………………………………………………………………………………………. 54

4.1 Introduction……………………………………………………………………………………………………………………………… 54

4.2 Assumption Made …………………………………………………………………………………………………………………….. 54

4.3 Calculation of the errors for the Detection and Localization of NTLs ……………………………………………….. 54

4.3.1 Calculation of the Per Phase Error (PPE). ……………………………………………………………………………….. 54

4.3.2 Result of the localization error for location of NTLs ………………………………………………………………… 55

4.4 Simulation and Result Analysis for Detection methodology. …………………………………………………………… 59

4.4.1 Simulation and result analysis for localization methodology. ………………………………………………………. 60

4.5 Validation of the Improved Method ……………………………………………………………………………………………. 64

CHAPTER FIVE ………………………………………………………………………………………………………………………………… 66 CONCLUSION AND RECOMMENDATIONS ………………………………………………………………………………………….. 66

5.1 Summary …………………………………………………………………………………………………………………………………. 66

5.2 Conclusion ……………………………………………………………………………………………………………………………….. 66

5.3 Significant Contribution …………………………………………………………………………………………………………….. 67

5.4 Limitations ……………………………………………………………………………………………………………………………….. 67

Recommendations …………………………………………………………………………………………………………………………. 67

REFERENCES ………………………………………………………………………………………………………………………………….. 68

CHAPTER ONE

INTRODUCTION

1.1 Background to the Study

Non-Technical Losses (electrical energy theft) has been a major concern in traditional power systems worldwide. In the United States (U.S.) alone energy theft was reported to cost utility companies around $6Billion/year (McDaniel & McLaughlin, 2009)this Figure appears relatively low when compared to the losses faced by utilities in developing countries such as Nigeria, Bangladesh, India and Pakistan (Eskom Annual Report 2009).Implementation of Advanced Metering Infrastructure (AMI) as one of the key technologies in smart grids promises to mitigate the risk of energy theft through its monitoring capabilities and the fine grained usage measurements. However, application of digital smart meters and addition of a cyber-layer to the metering system introduce numerous new vectors for energy theft.While traditional mechanical meters can only be compromised through physical tampering, in AMI the metering data can be tampered with, both locally and remotely before being sent to the smart meters or inside the smart meters or over the communication links. Penetration tests have already revealed several vulnerabilities in smart meters (Wright, 2009). In 2009, an organized energy theft attempt against AMI was reported by U.S. Federal Bureau of Investigation, which potentially could cost a utility company up to $400Million annually (Krebs B. 2012).Therefore, an Energy Theft Detection and Localization System (ETDLS) that can effectively and efficiently detect and localize energy theft attacks against AMI is urgently required.

Technical losses are inherent losses in power system network due to the inefficiency of power systems devices or iron core losses which occur during the transmission and distribution of electric power. While nontechnical losses on the other hand, are caused by actions external to the power system. These may include electricity theft, partial or non-payment of energy used by the customers.

In 1970s, Schweppe first proposed the idea of state estimation in power systems. Power system state estimation constitutes the core of the on-line system monitoring, analysis and control functions. State estimation acts like a filter between the raw measurements received from the system and all the application functions that require the most reliable data base for the current system operation state, and it typically includes bad data processing, state estimation solutions, parameter and topology error processing. Energy theft comprises of meter by pass, data attacks among other. In this work, the ideaof weighted least square state estimation will be explored to detect theft in power network.

Percentage of Non-Technical Losses (NTLs) and their monetary equivalent in some of the

European countries is shown in Table 1.1

Table 1.1: Non-Technical Losses in European Countries (Marques et al., 2016)

Country	NTLs (M€/year)	Total Losses (%)
Germany Italy Spain	504 408 426	4.7 6.3 7.8

 

1.2 Motivation

The power sector economic fatality caused by NTL cannot be over emphasized, the benefits of researches done on the reduction of technical losses will only be completely harnessed if the menace caused by NTL is completely or reasonably reduced. In Nigeria, According to a World Bank report seventy five percent (75%) of the total cost of energy injected in the various distribution feeder to serve customers are loss due to the combine effects of both technical and nontechnical loss of which only less than twenty percent are as a result of technical loss (Antmann, 2009). This implies that large portion of the energy loss are attributed to NTL which are mostly due to deliberate energy theft by customers, improper billing of customers, high level of indiscipline and unawareness of the customers to pay bills of electricity consumed and high level display of unprofessionalism on the side of the Electricity marketers. Despite the advent ofSmart Meters(SMs) which are introduced purposely to reduce and eliminate the menace of estimated billing and reduce energy theft, unfortunately rate of energy theft is increasing.

1.3 Significance of Research

The significance of this research is the development of branch current of LV Power network under theft. The accurate estimation of branch current makes it more easier to locate the point of theft which increases the true positive rate. To the best of my knowledge previous researchers have not considered the use of this method for improving the Localization of energy theft as at this time of compilling this thesis.

1.4 Statement of Problem

When there is an energy theft on a network supplying consumers’ load, there tend to be imbalance between the energy sold by utility companies and the energy purchase by customers which lead to economic fatality of utility companies and high energy charge to the customers. Therefore to create a profitable market for the utility company, it is quite important to locate the points of theft in order to finethe consumers and recover the monetory value of the energy loss. So many attempts have been made using the Artificial Intelligent method but accurate prediction of customers’ consumptions have always been very difficult which affect the accuracy of the result obtained(Marques et al., 2016). In this work the localization of theft is done considering the network operating condition through the data obtained from smart meters. Thus, this research present Detection and Localization of energy theft using branch current based weighted least square state estimation.

1.5Aim and Objectives

The aim of this research is developmentof a state estimation based improved detection and localization of Non-Technical Losses (NTLs) Using Smart Meter (SM) Measurements

The objectives of the research are as follows:

1. To replicate thedetection of NTLs algorithm from the work of(Marques et al., 2016)
2. To Develop a Weighted Least Square State Estimator for the estimation of state variables andmodification of the localization of NTLs algorithms iii. Performance comparison of the result obtained in ii above with the work of (Marques et al., 2016)in terms of True Positive Rate (TPR) and False Positive Rate (FPR).

1.6 Methodology

The following steps described the methodology used to carry out this research work:

1. Replication of the Detection of NTLs algorithm from the work of marques et al., (2016)

• Data acquisition (SMs’ current, DTC currents and power factors)
• Acquisition of Network Topology
• Calculation of the Per Phase Error (PPE)
• Comparison of the SMs reading and DTC Reading.

2. Development of Weighted Least Square State estimator for the estimation of the network’s states
   • Development of the network model equations
   • Collection of measurement and initial states.
3. Development of Localization algorithm
   • Acquisition of network model and topology (type of network configuration)
   • Acquisition of network parameters
   • Branch current Calculation and Estimation
   • Searching for the suspicious branches starting from the last level of the network.
4. Validation of the developed model using the work of marques et al., (2016).

 

DEVELOPMENT OF A STATE ESTIMATION BASED IMPROVED DETECTION AND LOCALIZATION OF NON-TECHNICAL LOSSES USING SMART METER MEASUREMENTS

Sharing is caring!