VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITY HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY NGUYEN DUC PHUONG HUY APPLICATION OF SUPPORT VECTOR MACHINE (SVM) TO PREDICT POTENTIAL LEAKS ON WATER SUPPLY NETWORK Major: Construction Management Major code: 8580302 MASTER’S THESIS HO CHI MINH CITY, June 2024 THIS RESEARCH IS COMPLETED AT UNIVERSITY OF TECHNOLOGY - VNU - HCM CITY Advisor: Assoc. Tran Duc Hoc The examiner 1: PhD. Nguyen Anh Thu The examiner 2: PhD. Chu Viet Cuong Master’s thesis was defended at HCM city University of Technology, VNU- HCM City on June 20th, 2024.
The board of the master’s thesis defense council includes: 1. Luong Duc Long 2. Do Tien Sy 3. Counter-argument member 1: PhD.
Nguyen Anh Thu 4. Counter-argument member 2: PhD. Chu Viet Cuong 5. Council members: PhD.
Nguyen Thanh Viet Verification of the Chairman of the master’s thesis defense council and the Dean of faculty of Civil Engineering after the thesis being corrected (If any). CHAIRMAN OF THE COUNCIL DEAN OF FACULTY OF CIVIL ENGINEERING NGUYEN DUC PHUONG HUY - 2270185 i VIETNAM NATIONAL UNIVERSITY - HO CHI MINH CITY SOCIALIST REPUBLIC OF VIETNAM HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY Independence – Freedom - Happiness THE TASK SHEET OF MASTER’S THESIS Full name: NGUYEN DUC PHUONG HUY Student ID: 2270185 Date of birth: November 19th, 1996 Place of birth: Ho Chi Minh City Major: Construction Management Major ID: 8580302 I. THESIS TITLE (In Vietnames): Ứng dụng công cụ máy vectơ hỗ trợ (SVM) để dự đoán các điểm xì bể tiềm ẩn trên mạng lưới cấp nước. THESIS TITLE (In English): Application of support vector machine (SVM) to predict potential leaks on water supply network.
TASKS AND CONTENTS: Assessment of delay risks in project of repairing, renovating, and upgrading water supply pipelines in Ho Chi Minh City. Besides, applying support vector machine in the mentioned field is expected to help identify risk factors to organise better planning in water pipelines construction project management. THESIS START DAY: January 15th, 2024. THESIS COMPLETION DAY: May 20th, 2024.
Tran Duc Hoc. Ho Chi Minh City, date ……… SUPERVISOR HEAD OF DEPARTMENT (Full name and signature) (Full name and signature) Assoc. Tran Duc Hoc DEAN OF FACULTY OF CIVIL ENGINEERING (Full name and signature) NGUYEN DUC PHUONG HUY - 2270185 ii ACKNOWLEDGEMENT I would like to acknowledge and give my warmest thanks to the lecturers, administrators, and classmates of International Master Program - Ho Chi Minh City University of Technology, who have made this program a huge success and continuously supported me during the course. The completion of this research could not have been possible without kind support and guidance of Assoc.
Tran Duc Hoc, my admirable thesis advisor. Your brilliant comments, suggestions and connections have been extremely helpful for me, not only on this journey of undertaking my master’s thesis research, but also in my future career and life. I would like to send my special thanks to the interview participants who have facilitated their precious time and efforts to review the research outcomes and shared with me their best advice. Finally, I am forever grateful to my family members and friends, in Ho Chi Minh city, in my hometown and overseas, who have always been understanding and giving me enormous emotional support during difficult times.
Ho Chi Minh City, June 22nd, 2024 NGUYEN DUC PHUONG HUY - 2270185 iii ABSTRACT Water is a very key element for the evolution of all organisms. During the advancement phase, population growth sometimes becomes a national burden, requiring governments to devise policies to best meet social utility issues such as housing, healthcare, and infrastructure. Among issues related to infrastructure such as electricity, communication, and education, water resource makes remarkably decisive contributions to ensuring livelihood needs, socio-economic and health development. In fact, the water supply project is being promoted very vigorously, receiving a lot of support from the government and the public.
Notwithstanding, considering various objective reasons, impacts on the progress making the projects behind schedule occur often, and consequently, this phenomenon leads to difficulty in meeting the water desire. In addition, water leakage and damage problems in the water supply network deliver several drawbacks, cause waste of resources, and sometimes are more dangerous than water supply interruptions. Scientifically, a lot of conducted research around the world show different analysis and assessment methods that can help minimise the impact of delays on water projects; simultaneously, the studies also provide techniques for leakage detection and anticipation on water distribution system. By contrast, in Vietnam, an unsignificant number of relevant works urges the necessity to conduct a survey to analyse factors affecting the causes of delays in repairing, renovating, upgrading water pipelines project and to anticipate hidden leak points in the water supply network by applying support vector machine.
Key words: water supply, delay factors, support vector machine, linear regression, leakage prediction. NGUYEN DUC PHUONG HUY - 2270185 iv TÓM TẮT LUẬN VĂN Nước là nhân tố rất quan trọng đối với quá trình tiến hóa của mọi sinh vật. Trong giai đoạn phát triển, sự gia tăng dân số đôi khi trở thành gánh nặng của một quốc gia, đòi hỏi chính phủ phải đưa ra các chính sách để đáp ứng tốt nhất các vấn đề tiện ích xã hội như nhà ở, chăm sóc sức khỏe và cơ sở hạ tầng. Trong số các vấn đề liên quan đến cơ sở hạ tầng như điện, truyền thông và giáo dục, tài nguyên nước đóng góp đáng kể vào việc đảm bảo nhu cầu sinh hoạt, phát triển kinh tế xã hội và sức khỏe; bên cạnh đó, nhiều dự án cấp nước đang được thúc đẩy rất mạnh mẽ, nhận được rất nhiều sự ủng hộ từ chính phủ và người dân.
Mặc dù khi xét đến nhiều lý do khách quan, có nhiều yếu tố tác động đến tiến độ khiến các dự án chậm tiến độ thường xuyên xảy ra và do đó, sự chậm trễ tiến độ thường dẫn đến khó khăn trong việc đáp ứng nhu cầu về nước sạch ngày càng tăng. Ngoài ra, các vấn đề rò rỉ và hư hỏng nước trong mạng lưới cấp nước mang lại nhiều tiêu cực, gây lãng phí tài nguyên và đôi khi nguy hiểm hơn cả việc gián đoạn cung cấp nước. Về mặt khoa học, nhiều nghiên cứu được tiến hành trên khắp thế giới cho thấy các phương pháp phân tích và đánh giá khác nhau có thể giúp giảm thiểu tác động của sự chậm trễ đối với các dự án cấp nước; đồng thời, các nghiên cứu cũng cung cấp các kỹ thuật phát hiện và dự đoán rò rỉ trên hệ thống phân phối nước. Tại Việt Nam hiện nay chưa có nhiều công trình nghiên cứu liên quan, điều này thúc đẩy sự cần thiết phải tiến hành một cuộc khảo sát để phân tích các yếu tố ảnh hưởng đến nguyên nhân gây chậm trễ trong dự án sửa chữa, cải tạo, nâng cấp đường ống nước và dự đoán các điểm rò rỉ tiềm ẩn trong mạng lưới cấp nước bằng cách áp dụng máy vectơ hỗ trợ.
Từ khóa: cấp nước, các yếu tố chậm trễ, máy vectơ hỗ trợ, hồi quy tuyến tính, dự đoán rò rỉ. NGUYEN DUC PHUONG HUY - 2270185 v THE COMMITMENT OF AUTHOR I, Nguyen Duc Phuong Huy, the author of master’s dissertation entitled “Application of support vector machine (SVM) to predict potential leaks on water supply network” hereby confirm that the thesis has been researched, written, and completed by the author under the supervision of the advisor Assoc. Tran Duc Hoc. All works, ideas and materials acknowledged from other references have been fully cited in the correct way.
Ho Chi Minh City, June 22nd, 2024 NGUYEN DUC PHUONG HUY - 2270185 vi TABLE OF CONTENTS OUTLINES THE TASK SHEET OF MASTER’S THESIS. iv TÓM TẮT LUẬN VĂN .v THE COMMITMENT OF AUTHOR. vi TABLE OF CONTENTS. vii LIST OF ABBREVIATION.
xi LIST OF FIGURES. xii LIST OF TABLES. xiii CHAPTER 1 INTRODUCTION .1 HCMC water supply’s status quo .2 The use of machine learning in water supply industry .3 Constructions projects’ status quo .4 Scope of Study.2 Methodology to apply SVM.6 Contribution to academic and practical fields .9 CHAPTER 2 LITERATURE REVIEW .10 NGUYEN DUC PHUONG HUY - 2270185 vii 2.2 Capital construction investment (CCI).3 Features of capital construction investment .4 The role of capital construction investment .2 Investment project (IP) explanation .1 Definition of an investment project.2 Obligations of an investment project .3 Delay risks of previous research .1 Group of external environmental factors .2 Group of elements of management information system .3 Policy related factors .4 Group of factors on decentralisation of authority to the investor .5 Group of factors on capital sources for project managing .6 Group of factors related to the capacity of project participants.7 Group of factors on the capacity of investors .8 Group of project characteristics factors .4 Introduction to WDS .2 Water supply network in Ho Chi Minh City .5 Support vector machine.1 Introduction to support vector machine .2 How support vector machine works .3 Kernel function in support vector machine .4 Margin in support vector machine .5 Solution to optimisation in support vector machine .7 Previous studies on leakage pinpointing .1 Usage of variational autoencoder and support vector machine to establish a framework in water distribution system .33 NGUYEN DUC PHUONG HUY - 2270185 viii 2.2 Application of artificial neural network to strategically detect water leakage 34 2.3 Integration of hydraulic software in machine learning .4 Conventional methods besides machine learning techniques .5 The revolution of water leak detection and localisation .35 CHAPTER 3 FRAMEWORK STRUCTURING .1 Architecture of proposed framework .2 Survey data compilation.2 Variables size and methodology of determination .3 Data investigation technique .2 Evaluation of the reliability .3 Exploratory factor analysis .4 Linear regression analysis in statistics .5 Test for difference .1 Support vector regression (SVR) .2 Python programming language integration .1 Root mean squared error .2 Mean absolute error .3 Coefficient of determination .52 CHAPTER 4 DATA PROCESSING .5 Linear regression analysis .70 NGUYEN DUC PHUONG HUY - 2270185 ix 4.2 Inspection for Multicollinearity .1 Data taking technique .1 Calculation results of Support Vector Regression model .2 Analysis between two datasets in kernel function .2 Values of research .97 APPENDIX 2: SPSS STATISTICAL RESULT .114 NGUYEN DUC PHUONG HUY - 2270185 x LIST OF ABBREVIATION SVM Support vector machine RRUWPP Repairing, renovating, upgrading water pipeline projects WDS Water distribution system WSN Water supply network HCMC Ho Chi Minh City SAWACO Saigon Water Corporation Abbr. Number AI Artificial intelligence ML Machine learning CCI Capital construction investment IP Investment project WASUCO Water Supply Joint stock company TPI Time performance index MSE Mean squared error RMSE Root mean squared error MAE Mean absolute error R2 Coefficient of determination CFA Confirmatory factor analysis EFA Exploratory factor analysis KMO Kaiser-Meyer-Olkin EC Emitter coefficient α Alpha AVE Average Variance Extracted VIF Variance Inflation Factor NGUYEN DUC PHUONG HUY - 2270185 xi LIST OF FIGURES Figure 1-1 The logo on the SAWACO’ 17th anniversary of establishment.
2 Figure 2-1 Diagram of a project lifecycle. 15 Figure 2-2 Distribution of water supply. 23 Figure 2-3 Ho Chi Minh City’s water distribution system. 24 Figure 2-4 Hyperplane and support vectors.
25 Figure 2-5 Possible hyperplanes and one with the maximum margin. 26 Figure 2-6 Hyperplane in Rule 1. 27 Figure 2-7 Hyperplane with the largest margin. 27 Figure 2-8 The appropriate hyperplane even though it does not have the largest margin.
28 Figure 2-9 Outlier in SVM. 28 Figure 2-10 Description of the Kernel method. 29 Figure 2-11 A kernel added to create z axis. 29 Figure 2-12 Data points in two-dimensional space.
30 Figure 2-13 Linear regression model. 32 Figure 3-1 Research process. 40 Figure 4-1 The percentage of male and female interviewee. 56 Figure 4-2 Age group of participants.