Mở Bài
Trí tuệ nhân tạo (AI) trong bảo trì dự đoán cơ sở hạ tầng đang trở thành một chủ đề nóng hổi trong kỳ thi IELTS Reading, đặc biệt xuất hiện với tần suất ngày càng cao trong các đề thi từ năm 2020 trở lại đây. Chủ đề này kết hợp giữa công nghệ hiện đại, kỹ thuật xây dựng và quản lý đô thị thông minh – những lĩnh vực được Cambridge IELTS và British Council đặc biệt quan tâm khi biên soạn đề thi.
Trong bài viết này, bạn sẽ được trải nghiệm một bộ đề thi IELTS Reading hoàn chỉnh với 3 passages có độ khó tăng dần từ Easy (Band 5.0-6.5), Medium (Band 6.0-7.5) đến Hard (Band 7.0-9.0). Mỗi passage được thiết kế giống 100% với đề thi thật, bao gồm đầy đủ 40 câu hỏi với 7 dạng bài khác nhau thường gặp nhất. Bạn sẽ nhận được đáp án chi tiết kèm giải thích vị trí, paraphrase và chiến lược làm bài, cùng với bộ từ vựng chuyên ngành được phân loại cẩn thận theo từng passage.
Bộ đề này phù hợp cho học viên từ band 5.0 trở lên muốn làm quen với chủ đề công nghệ-cơ sở hạ tầng, đồng thời rèn luyện kỹ năng skimming, scanning và phân tích thông tin phức tạp – những kỹ năng cốt lõi để đạt band điểm cao trong IELTS Reading.
Hướng Dẫn Làm Bài IELTS Reading
Tổng Quan Về IELTS Reading Test
IELTS Reading Test kéo dài 60 phút với 3 passages và tổng cộng 40 câu hỏi. Đây là bài thi yêu cầu khả năng quản lý thời gian chặt chẽ và chiến lược làm bài thông minh.
Phân bổ thời gian khuyến nghị:
- Passage 1: 15-17 phút (độ khó thấp nhất, nên làm nhanh để dành thời gian)
- Passage 2: 18-20 phút (độ khó trung bình, cần đọc kỹ hơn)
- Passage 3: 23-25 phút (độ khó cao nhất, yêu cầu phân tích sâu)
Lưu ý quan trọng: Không có thời gian chuyển đáp án riêng, bạn cần ghi đáp án trực tiếp vào answer sheet trong 60 phút.
Các Dạng Câu Hỏi Trong Đề Này
Bộ đề thi này bao gồm 7 dạng câu hỏi phổ biến nhất trong IELTS Reading:
- Multiple Choice – Chọn đáp án đúng nhất từ A, B, C, D
- True/False/Not Given – Xác định thông tin đúng, sai hay không được đề cập
- Matching Information – Ghép thông tin với đoạn văn phù hợp
- Sentence Completion – Hoàn thành câu với từ trong bài
- Matching Headings – Ghép tiêu đề phù hợp với từng đoạn
- Summary Completion – Hoàn thành đoạn tóm tắt
- Short-answer Questions – Trả lời ngắn gọn dựa trên thông tin trong bài
Mỗi dạng câu hỏi đòi hỏi kỹ thuật làm bài riêng biệt, và bạn sẽ được hướng dẫn chi tiết trong phần giải thích đáp án.
IELTS Reading Practice Test
PASSAGE 1 – The Rise of Smart Infrastructure Monitoring
Độ khó: Easy (Band 5.0-6.5)
Thời gian đề xuất: 15-17 phút
Infrastructure forms the backbone of modern society. Roads, bridges, railways, water supply systems, and power grids enable daily life and economic activity. However, maintaining these critical assets has traditionally been reactive rather than proactive. Engineers typically inspect infrastructure at scheduled intervals or after problems become visible. This approach often leads to unexpected failures, costly emergency repairs, and potential safety hazards.
Artificial Intelligence (AI) is now revolutionising how we monitor and maintain infrastructure. By analysing vast amounts of data from sensors, cameras, and inspection reports, AI systems can predict when and where failures might occur before they happen. This shift from reactive to predictive maintenance represents a fundamental transformation in infrastructure management.
The concept is relatively straightforward. Sensors embedded in bridges, for example, continuously measure stress levels, vibrations, temperature changes, and material deterioration. These sensors generate enormous quantities of data – far more than human engineers could possibly analyse manually. AI algorithms process this data in real-time, identifying patterns and anomalies that indicate potential problems. When the system detects unusual readings, it alerts maintenance teams to investigate before a minor issue becomes a major crisis.
Transportation infrastructure has been among the first to adopt these technologies. In the United Kingdom, Network Rail uses AI-powered systems to monitor thousands of kilometres of railway tracks. Sensors detect track defects, unusual vibration patterns, and changes in rail alignment. The AI system analyses this information alongside historical maintenance records and weather data to predict where problems are most likely to develop. This approach has significantly reduced track failures and improved railway safety and reliability.
Bridges represent particularly critical infrastructure because their failure can have catastrophic consequences. Traditional bridge inspections require engineers to conduct visual assessments and manual measurements – a time-consuming process that provides only periodic snapshots of bridge condition. AI-enhanced monitoring offers continuous oversight. Strain gauges measure how bridges respond to traffic loads and environmental conditions. Computer vision systems analyse images to detect cracks, corrosion, and other deterioration signs. Machine learning algorithms compare current conditions with historical data to forecast maintenance needs.
The city of Barcelona provides an excellent example of AI in urban infrastructure management. The city has deployed thousands of sensors across its water distribution network. These sensors monitor water pressure, flow rates, and quality indicators. AI algorithms analyse this data to detect leaks, predict pipe failures, and optimise water distribution. Since implementing this system, Barcelona has reduced water losses by 25% and cut maintenance costs by 30%. The system pays particular attention to vulnerable sections of the network, where pipes are older or have experienced previous problems.
Power grids also benefit substantially from AI-powered predictive maintenance. Electricity distribution networks include thousands of components – transformers, cables, substations, and switching equipment – all of which require regular maintenance. Traditional approaches involve routine inspections and component replacement based on age rather than actual condition. AI systems monitor electrical parameters, equipment temperature, and loading patterns to identify components at risk of failure. Utility companies using these systems report 40-50% reductions in unexpected outages and significant cost savings.
The implementation of AI in infrastructure maintenance does face challenges. Initial costs for sensors and AI systems can be substantial, though these investments typically generate returns within a few years through reduced emergency repairs and extended infrastructure lifespan. Data quality and integration present another challenge – AI systems require accurate, consistent data from multiple sources. Cybersecurity concerns also arise, as connected infrastructure creates potential vulnerabilities to hacking or sabotage.
Despite these challenges, the trajectory is clear. As AI technology becomes more sophisticated and affordable, its application in infrastructure maintenance will expand. The shift from reactive to predictive maintenance represents not merely a technological upgrade but a new paradigm in how society manages its critical assets. By identifying problems before they cause failures, AI helps ensure infrastructure remains safe, reliable, and cost-effective for future generations.
Questions 1-13
Questions 1-5: Multiple Choice
Choose the correct letter, A, B, C or D.
1. According to the passage, traditional infrastructure maintenance has been characterised by:
A. Advanced warning systems
B. Responses to visible problems
C. Continuous monitoring
D. Predictive analysis
2. What is the primary advantage of AI-powered infrastructure monitoring?
A. It reduces the number of sensors needed
B. It eliminates the need for human engineers
C. It identifies potential failures before they occur
D. It makes infrastructure last forever
3. In the railway system mentioned, AI analyses data from all of the following EXCEPT:
A. Sensor readings
B. Past maintenance information
C. Weather conditions
D. Passenger complaints
4. The Barcelona water management system achieved:
A. Complete elimination of water losses
B. A 25% reduction in water wastage
C. A 50% decrease in maintenance costs
D. Replacement of all old pipes
5. According to the passage, one challenge in implementing AI infrastructure monitoring is:
A. Lack of available technology
B. Resistance from engineers
C. High initial investment costs
D. Shortage of electricity
Questions 6-9: True/False/Not Given
Do the following statements agree with the information in the passage?
Write:
- TRUE if the statement agrees with the information
- FALSE if the statement contradicts the information
- NOT GIVEN if there is no information on this
6. Sensors in bridges measure only temperature and vibration.
7. Visual bridge inspections provide continuous monitoring of bridge conditions.
8. Utility companies using AI systems experience 40-50% fewer unexpected power outages.
9. All countries have now adopted AI for infrastructure maintenance.
Questions 10-13: Sentence Completion
Complete the sentences below.
Choose NO MORE THAN TWO WORDS from the passage for each answer.
10. Infrastructure such as roads and bridges forms the __ of modern society.
11. AI algorithms can process sensor data in __, identifying patterns that suggest problems.
12. Traditional bridge inspections only provide __ of bridge condition rather than continuous data.
13. Connected infrastructure systems may face __ risks from hacking or sabotage.
PASSAGE 2 – Machine Learning Algorithms in Infrastructure Assessment
Độ khó: Medium (Band 6.0-7.5)
Thời gian đề xuất: 18-20 phút
The application of machine learning to infrastructure maintenance represents a sophisticated evolution in asset management. Unlike traditional rule-based systems that follow predetermined protocols, machine learning algorithms can identify complex relationships within data that human analysts might overlook. These algorithms improve their predictive accuracy over time, learning from both successful predictions and errors to refine their models continuously.
Deep learning, a subset of machine learning, has proven particularly effective in infrastructure diagnostics. These algorithms employ artificial neural networks inspired by the human brain’s structure, containing multiple layers of interconnected nodes that process information. When analysing infrastructure data, deep learning systems can detect subtle patterns indicating incipient failures – problems in their earliest stages before visible symptoms appear. This capability extends the window of opportunity for preventive action, potentially adding years to infrastructure service life.
Computer vision, powered by convolutional neural networks (CNNs), has transformed infrastructure inspection procedures. Traditional visual inspections require trained engineers to physically examine infrastructure components, a process that is labour-intensive, time-consuming, and potentially dangerous, especially for structures like bridges, towers, and high-voltage transmission lines. Computer vision systems process images from cameras or drones, automatically identifying defects such as cracks, spalling, corrosion, and deformation. Recent studies demonstrate that well-trained CNN models can match or exceed human expert accuracy in defect detection while processing images far more rapidly.
The city of Singapore has pioneered comprehensive infrastructure monitoring through its Smart Nation initiative. The Land Transport Authority employs AI systems monitoring the city-state’s extensive road network. Sensors embedded in road surfaces detect structural weaknesses and subsurface deterioration invisible to conventional inspection methods. The system integrates multiple data streams: traffic loads, rainfall patterns, subsurface water levels, and historical maintenance records. Machine learning algorithms analyse these diverse inputs to predict which road sections will require intervention and when. This data-driven approach enables authorities to schedule maintenance during periods of low traffic, minimising disruption while maximising efficiency.
Natural Language Processing (NLP), another branch of AI, contributes unexpectedly to infrastructure maintenance. Historical maintenance reports, inspection logs, and incident records contain valuable information often trapped in unstructured text format. NLP algorithms can extract insights from thousands of documents, identifying recurring problems, common failure modes, and effective repair strategies. This knowledge extraction helps organisations learn from past experiences systematically rather than relying on individual engineers’ memories.
Railway infrastructure presents unique challenges for predictive maintenance. Rail networks operate continuously with minimal downtime for inspections. Japan’s Shinkansen (bullet train) system represents a benchmark in reliability, achieving this through rigorous AI-assisted monitoring. Acoustic sensors detect unusual sounds from tracks, wheels, and overhead electrical systems. Vibration sensors identify bearing wear in wheels and irregularities in track geometry. Thermal imaging cameras spot overheating components before they fail. The AI system processes this multisensory data in real-time, distinguishing between normal operational variations and genuine problems requiring attention. The system’s sophistication allows trains to operate at speeds exceeding 300 kilometres per hour with exceptional safety records.
Water infrastructure benefits significantly from AI-powered leak detection. Traditional methods for finding leaks in underground pipes involve waiting until water surfaces or using acoustic equipment to listen for leak sounds – both inefficient approaches. AI systems analyse pressure fluctuations and flow patterns throughout water networks, using algorithms that can pinpoint leak locations within metres. The city of London, with its extensive but ageing water infrastructure, has implemented such systems. The AI identifies unusual consumption patterns that might indicate leaks, distinguishes between legitimate demand variations and actual losses, and prioritises response based on leak severity and location. This targeted approach reduces water losses while optimising maintenance crew deployment.
Digital twins represent the cutting edge in AI-enhanced infrastructure management. A digital twin is a virtual replica of physical infrastructure, continuously updated with real-time sensor data. AI algorithms run simulations on these digital twins, testing how infrastructure might respond to various conditions: extreme weather, increased loads, or component failures. Engineers can experiment with different maintenance strategies virtually, determining optimal approaches before implementing them physically. The Port of Rotterdam, Europe’s largest port, employs digital twins of its maritime infrastructure. The system models how dock structures, cranes, and navigation channels will behave under different conditions, enabling proactive maintenance and capacity optimisation.
However, the sophistication of these AI systems introduces implementation complexities. Developing accurate machine learning models requires extensive training data representing various conditions and failure modes. For infrastructure with long service lives and infrequent failures, accumulating sufficient training data can take years. Transfer learning – applying models trained on one type of infrastructure to another – offers a partial solution, though with limitations. Algorithm transparency presents another concern; deep learning models sometimes function as “black boxes,” making predictions without explicable reasoning. For critical infrastructure decisions, stakeholders may hesitate to rely on systems they cannot fully understand.
The integration of AI into existing asset management frameworks requires not just technology but organisational change. Maintenance teams must develop new skills, transitioning from reactive repair work to data interpretation and predictive planning. Organisations need governance structures ensuring AI recommendations receive appropriate human oversight rather than blind acceptance. Regulatory frameworks must evolve to address AI-based maintenance decisions, particularly regarding liability when AI predictions prove incorrect.
Hệ thống AI giám sát cơ sở hạ tầng với cảm biến thông minh và phân tích dữ liệu thời gian thực
Despite these challenges, the trajectory towards AI-driven predictive maintenance appears irreversible. As technology matures and costs decline, even smaller municipalities and organisations will access these capabilities. The fundamental value proposition – identifying problems before they cause failures, optimising maintenance resources, and extending infrastructure longevity – proves compelling across diverse contexts. The question is no longer whether AI will transform infrastructure maintenance, but how quickly this transformation will occur and which organisations will lead the transition.
Questions 14-26
Questions 14-18: Yes/No/Not Given
Do the following statements agree with the claims of the writer in the passage?
Write:
- YES if the statement agrees with the claims of the writer
- NO if the statement contradicts the claims of the writer
- NOT GIVEN if it is impossible to say what the writer thinks about this
14. Machine learning algorithms can identify relationships in data that humans might miss.
15. Computer vision systems are always more accurate than human experts in detecting defects.
16. Singapore’s Smart Nation initiative was the first infrastructure monitoring system in the world.
17. Natural Language Processing helps analyse historical maintenance documents.
18. All railway systems worldwide now use AI-assisted monitoring.
Questions 19-23: Matching Headings
The passage has ten paragraphs (A-J). Choose the correct heading for paragraphs B, D, F, H, and J from the list of headings below.
List of Headings:
i. The challenge of limited training data
ii. Processing visual information automatically
iii. Water leak detection through pattern analysis
iv. Understanding how algorithms learn and improve
v. Singapore’s integrated monitoring approach
vi. Digital replicas for testing maintenance strategies
vii. Railway monitoring in high-speed systems
viii. Organisational changes required for AI adoption
ix. Extracting knowledge from text documents
Paragraph B: __
Paragraph D: __
Paragraph F: __
Paragraph H: __
Paragraph J: __
Questions 24-26: Summary Completion
Complete the summary below.
Choose NO MORE THAN TWO WORDS from the passage for each answer.
Digital twins are 24. __ of physical infrastructure that receive continuous updates from sensors. AI can run 25. __ on these virtual models to test different scenarios without affecting real structures. The Port of Rotterdam uses this technology to manage its **26. __, allowing proactive maintenance planning and better resource utilisation.
PASSAGE 3 – The Socio-Technical Dimensions of AI-Driven Infrastructure Resilience
Độ khó: Hard (Band 7.0-9.0)
Thời gian đề xuất: 23-25 phút
The integration of artificial intelligence into infrastructure maintenance transcends purely technical considerations, embodying a paradigmatic shift in how societies conceptualise and operationalise the stewardship of critical assets. This transformation necessitates examination through multiple lenses – technological, economic, institutional, and ethical – to apprehend its full implications for urban resilience and sustainability.
Contemporary discourse on infrastructure resilience emphasises the capacity of systems to withstand, adapt to, and recover from disruptions. Traditional resilience frameworks have focused on physical redundancy and robustness – designing infrastructure to survive extreme stresses through over-engineering. However, this approach proves increasingly untenable in an era of resource constraints and accelerating climate change. AI-driven predictive maintenance offers an alternative paradigm: adaptive resilience through intelligent monitoring and anticipatory intervention. Rather than building ever-stronger structures, this approach emphasises dynamic responsiveness – detecting vulnerability accumulation and implementing corrective measures before threshold failures occur.
The epistemological foundations of AI-based infrastructure assessment merit scrutiny. Machine learning models generate predictions by identifying statistical correlations within training data. These correlations may or may not reflect causal relationships. When an algorithm predicts bridge failure based on sensor readings, the model identifies patterns statistically associated with past failures without necessarily understanding the underlying physical mechanisms. This distinction becomes consequential when infrastructure encounters unprecedented conditions outside the training data distribution. An algorithm trained on historical data may fail to recognise novel failure modes resulting from climate change-induced stresses or new materials degradation processes. The epistemic humility required in deploying these systems – acknowledging their limitations and maintaining human expertise in parallel – often receives insufficient attention in implementation discussions.
Economic analyses of AI infrastructure monitoring typically emphasise cost-benefit ratios and return on investment. These metrics capture important dimensions but potentially overlook broader economic implications. The transition to predictive maintenance reconfigures labour markets within infrastructure sectors. Traditional craft skills in manual inspection and reactive repair become less valued relative to data science competencies and predictive analytics. This transition creates distributional consequences: some workers gain opportunities in emerging roles while others face skill obsolescence. Equitable transitions require proactive policies including retraining programmes and social safety nets, yet such considerations rarely feature prominently in AI implementation strategies.
Furthermore, the economics of AI infrastructure monitoring exhibit significant scale dependencies. Large metropolitan authorities and national infrastructure operators can amortise system development costs across extensive asset portfolios, achieving rapid returns. Smaller municipalities and developing nations face disproportionately high per-asset costs, potentially exacerbating existing inequalities in infrastructure quality. This disparity raises questions about universal access to advanced maintenance technologies and the risk of a two-tiered infrastructure landscape where affluent areas benefit from AI-enhanced reliability while resource-constrained communities rely on outdated approaches.
The algorithmic governance of infrastructure introduces novel accountability challenges. When AI systems recommend deferring maintenance on a bridge that subsequently fails, liability becomes ambiguous. Is responsibility attributable to the AI developers, the infrastructure operators who relied on the system, or the public authorities overseeing operations? Legal frameworks developed for human decision-making prove inadequate for algorithmically-mediated choices. The distributed nature of AI systems – involving sensor manufacturers, algorithm developers, data integrators, and infrastructure operators – creates fragmented responsibility that complicates accountability.
Privacy and surveillance dimensions warrant consideration, particularly for infrastructure monitoring in urban environments. Sensor networks generating data for AI analysis often incidentally capture information about individuals’ movements and behaviours. A comprehensive road monitoring system, for instance, might track vehicle movements throughout a city. While this data serves legitimate infrastructure management purposes, it simultaneously creates surveillance capabilities that could be repurposed for mass monitoring. Governance frameworks must balance infrastructure maintenance needs against civil liberties concerns, establishing robust data protection and use limitations.
Cybersecurity constitutes a paramount concern for AI-integrated infrastructure. Networked systems create attack surfaces that malicious actors might exploit. Sophisticated adversaries could manipulate sensor data to disguise infrastructure deterioration or create false alerts that divert maintenance resources. More insidiously, attackers might subtly corrupt training data, introducing biases that degrade algorithmic performance gradually. The critical nature of infrastructure makes it an attractive target for state-sponsored cyberattacks and terrorism. Resilient AI infrastructure monitoring requires defensive architectures that maintain functionality despite compromised components and detection mechanisms identifying anomalous data patterns suggesting manipulation.
The temporal dynamics of AI system deployment deserve attention. Current machine learning models excel at pattern recognition within stationary data distributions but struggle with non-stationary environments where underlying patterns shift over time. Climate change fundamentally alters the stress regimes infrastructure experiences – temperature extremes, precipitation patterns, sea levels, and storm intensities all deviate from historical norms. AI models trained on past data may systematically underestimate future risks if they cannot extrapolate beyond historical experience. Addressing this limitation requires hybrid approaches combining data-driven machine learning with physics-based models that understand infrastructure behaviour from first principles.
Trung tâm kiểm soát AI dự báo bảo trì cơ sở hạ tầng với công nghệ học máy và phân tích dữ liệu
Institutional capacity for AI adoption varies considerably across jurisdictions. Successful implementation demands not merely technology acquisition but organisational transformation. Infrastructure agencies must cultivate interdisciplinary teams combining domain expertise in engineering with competencies in data science, statistics, and computer systems. Bureaucratic structures developed for traditional maintenance paradigms may impede the agile, data-responsive approaches AI enables. Procurement regulations designed for physical infrastructure prove ill-suited for software systems requiring continuous updates and iterative development. Regulatory frameworks specifying maintenance intervals and procedures may conflict with AI-generated recommendations. These institutional frictions can substantially delay or undermine AI adoption regardless of technical merits.
The geopolitical dimensions of AI infrastructure monitoring emerge as strategic considerations. Dominant technology providers – predominantly based in the United States and China – exert influence over global infrastructure through AI systems and platforms. Dependencies on foreign AI technologies for critical infrastructure create vulnerability to supply chain disruptions, technology denial, and geopolitical leverage. Some nations pursue technological sovereignty in AI infrastructure monitoring, developing indigenous capabilities despite higher costs and longer timelines. These strategic calculations balance efficiency gains from commercially available technologies against autonomy and security considerations.
Ethical frameworks for AI infrastructure governance remain nascent. Questions of algorithmic fairness arise when systems prioritise maintenance across infrastructure networks. Should algorithms optimise for economic efficiency, minimising total costs, or prioritise equity, ensuring comparable service quality across socioeconomic groups? When resources constrain maintenance, AI systems effectively make value judgments about which communities receive timely interventions. Explicating these embedded values and subjecting them to democratic deliberation represents an unfulfilled imperative in current practice.
The trajectory of AI in infrastructure maintenance portends profound transformation in how humanity manages its built environment. Realising this technology’s potential while mitigating associated risks demands holistic governance addressing technical, economic, social, and ethical dimensions conjointly. The challenge lies not in the technology itself but in developing institutional arrangements and normative frameworks ensuring AI serves collective interests in infrastructure resilience, sustainability, and equitable access. Success requires transcending narrow technocratic perspectives to embrace socio-technical complexity, recognising that infrastructure systems are ultimately social constructs embodying society’s values, priorities, and power relations.
Questions 27-40
Questions 27-31: Multiple Choice
Choose the correct letter, A, B, C or D.
27. According to the passage, adaptive resilience differs from traditional resilience by:
A. Building stronger physical structures
B. Using redundancy and over-engineering
C. Responding dynamically to detected vulnerabilities
D. Eliminating all infrastructure failures
28. The epistemological limitation of machine learning models is that they:
A. Cannot process sensor data
B. Identify correlations that may not indicate causation
C. Require too much training data
D. Are too expensive to implement
29. The passage suggests that the transition to AI-based maintenance:
A. Benefits all workers equally
B. Has no impact on employment
C. Creates distributional consequences for different skill sets
D. Eliminates the need for human expertise
30. According to the passage, accountability becomes ambiguous when:
A. AI systems are too expensive
B. Engineers make mistakes
C. AI-recommended decisions lead to failures
D. Infrastructure is too old
31. The author’s view on geopolitical considerations is that:
A. All countries should use the same AI systems
B. Technology dependence creates strategic vulnerabilities
C. Only wealthy nations can afford AI systems
D. Geopolitics has no impact on infrastructure
Questions 32-36: Matching Features
Match each challenge (32-36) with the correct area of concern (A-H).
Areas of Concern:
A. Economic inequality
B. Privacy rights
C. Cybersecurity threats
D. Climate change adaptation
E. Legal frameworks
F. Labour market changes
G. Technical accuracy
H. Cultural differences
32. Sensors may incidentally capture data about individuals’ movements
33. Small municipalities face disproportionately high implementation costs
34. Models trained on historical data may fail under unprecedented climate conditions
35. Traditional craft skills become less valued than data science competencies
36. Malicious actors could manipulate sensor data to disguise deterioration
Questions 37-40: Short-answer Questions
Answer the questions below.
Choose NO MORE THAN THREE WORDS from the passage for each answer.
37. What type of resilience do traditional frameworks emphasise through over-engineering?
38. What kind of approach combines data-driven learning with physics-based understanding?
39. What must infrastructure agencies cultivate that combines engineering with data science?
40. According to the passage, infrastructure systems ultimately embody society’s values and what else?
Answer Keys – Đáp Án
PASSAGE 1: Questions 1-13
- B
- C
- D
- B
- C
- FALSE
- FALSE
- TRUE
- NOT GIVEN
- backbone
- real-time
- periodic snapshots
- cybersecurity
PASSAGE 2: Questions 14-26
- YES
- NOT GIVEN
- NOT GIVEN
- YES
- NOT GIVEN
- iv
- v
- vii
- vi
- viii
- virtual replicas
- simulations
- maritime infrastructure
PASSAGE 3: Questions 27-40
- C
- B
- C
- C
- B
- B
- A
- D
- F
- C
- physical redundancy / redundancy and robustness
- hybrid approaches
- interdisciplinary teams
- power relations
Giải Thích Đáp Án Chi Tiết
Passage 1 – Giải Thích
Câu 1: B
- Dạng câu hỏi: Multiple Choice
- Từ khóa: traditional infrastructure maintenance, characterised
- Vị trí trong bài: Đoạn 1, dòng 3-5
- Giải thích: Bài viết nói rõ “Engineers typically inspect infrastructure at scheduled intervals or after problems become visible” – nghĩa là họ phản ứng với những vấn đề có thể nhìn thấy được, tức là “Responses to visible problems”. Đáp án A, C, D đều mô tả các phương pháp hiện đại mà bài viết đối lập với phương pháp truyền thống.
Câu 2: C
- Dạng câu hỏi: Multiple Choice
- Từ khóa: primary advantage, AI-powered monitoring
- Vị trí trong bài: Đoạn 2, dòng 2-4
- Giải thích: “AI systems can predict when and where failures might occur before they happen” được paraphrase thành “identifies potential failures before they occur”. Đây là lợi ích chính được nhấn mạnh trong toàn bài.
Câu 6: FALSE
- Dạng câu hỏi: True/False/Not Given
- Từ khóa: sensors in bridges, measure only, temperature and vibration
- Vị trí trong bài: Đoạn 3, dòng 2-3
- Giải thích: Bài viết liệt kê “stress levels, vibrations, temperature changes, and material deterioration” – có nhiều yếu tố hơn chỉ temperature và vibration, nên câu này là FALSE.
Câu 8: TRUE
- Dạng câu hỏi: True/False/Not Given
- Từ khóa: Utility companies, 40-50%, unexpected power outages
- Vị trí trong bài: Đoạn 7, dòng cuối
- Giải thích: Bài viết nói chính xác “Utility companies using these systems report 40-50% reductions in unexpected outages” – trùng khớp hoàn toàn với câu hỏi.
Câu 10: backbone
- Dạng câu hỏi: Sentence Completion
- Từ khóa: Infrastructure, forms, modern society
- Vị trí trong bài: Đoạn 1, câu đầu tiên
- Giải thích: Câu mở đầu có cụm “Infrastructure forms the backbone of modern society” – từ “backbone” phù hợp về ngữ pháp và ngữ nghĩa.
Câu 13: cybersecurity
- Dạng câu hỏi: Sentence Completion
- Từ khóa: Connected infrastructure, risks, hacking, sabotage
- Vị trí trong bài: Đoạn 8, dòng cuối
- Giải thích: “Cybersecurity concerns also arise, as connected infrastructure creates potential vulnerabilities to hacking or sabotage” – từ “cybersecurity” là đáp án chính xác.
Passage 2 – Giải Thích
Câu 14: YES
- Dạng câu hỏi: Yes/No/Not Given
- Từ khóa: Machine learning algorithms, identify relationships, humans might miss
- Vị trí trong bài: Đoạn A, dòng 2-3
- Giải thích: “machine learning algorithms can identify complex relationships within data that human analysts might overlook” – đồng ý hoàn toàn với quan điểm tác giả.
Câu 17: YES
- Dạng câu hỏi: Yes/No/Not Given
- Từ khóa: Natural Language Processing, analyse, historical maintenance documents
- Vị trí trong bài: Đoạn E, toàn đoạn
- Giải thích: “NLP algorithms can extract insights from thousands of documents, identifying recurring problems…” – rõ ràng khẳng định NLP giúp phân tích tài liệu lịch sử.
Câu 19: iv (Understanding how algorithms learn and improve)
- Dạng câu hỏi: Matching Headings
- Đoạn văn: Paragraph B
- Giải thích: Đoạn B tập trung vào việc giải thích cách deep learning algorithms học và cải thiện độ chính xác theo thời gian, phù hợp với heading về “how algorithms learn and improve”.
Câu 24: virtual replicas
- Dạng câu hỏi: Summary Completion
- Từ khóa: Digital twins
- Vị trí trong bài: Đoạn H, câu đầu
- Giải thích: “A digital twin is a virtual replica of physical infrastructure” – cụm “virtual replicas” chính xác mô tả digital twins.
Câu 26: maritime infrastructure
- Dạng câu hỏi: Summary Completion
- Từ khóa: Port of Rotterdam
- Vị trí trong bài: Đoạn H, dòng cuối
- Giải thích: “The Port of Rotterdam…employs digital twins of its maritime infrastructure” – đáp án trực tiếp từ bài.
Công nghệ AI phân tích kết cấu cầu với cảm biến thông minh và dự đoán hư hỏng
Passage 3 – Giải Thích
Câu 27: C
- Dạng câu hỏi: Multiple Choice
- Từ khóa: adaptive resilience, differs from, traditional resilience
- Vị trí trong bài: Đoạn B, giữa đoạn
- Giải thích: Bài viết đối lập adaptive resilience (detecting vulnerability accumulation and implementing corrective measures) với traditional resilience (over-engineering). Đáp án C “Responding dynamically to detected vulnerabilities” paraphrase chính xác ý này.
Câu 28: B
- Dạng câu hỏi: Multiple Choice
- Từ khóa: epistemological limitation, machine learning models
- Vị trí trong bài: Đoạn C, dòng 2-5
- Giải thích: “Machine learning models generate predictions by identifying statistical correlations…These correlations may or may not reflect causal relationships” – đây là hạn chế nhận thức luận chính được nhấn mạnh.
Câu 29: C
- Dạng câu hỏi: Multiple Choice
- Từ khóa: transition to AI-based maintenance
- Vị trí trong bài: Đoạn D, giữa đoạn
- Giải thích: “This transition creates distributional consequences: some workers gain opportunities in emerging roles while others face skill obsolescence” – đúng với đáp án C.
Câu 32: B (Privacy rights)
- Dạng câu hỏi: Matching Features
- Từ khóa: Sensors, incidentally capture, individuals’ movements
- Vị trí trong bài: Đoạn G, dòng 2-3
- Giải thích: Đoạn này nằm trong phần bàn về “Privacy and surveillance dimensions”, nói về sensor networks “incidentally capture information about individuals’ movements”.
Câu 34: D (Climate change adaptation)
- Dạng câu hỏi: Matching Features
- Từ khóa: Models trained on historical data, fail, unprecedented climate conditions
- Vị trí trong bài: Đoạn I, giữa đoạn
- Giải thích: “Climate change fundamentally alters the stress regimes…AI models trained on past data may systematically underestimate future risks” – đây là thách thức về thích ứng biến đổi khí hậu.
Câu 37: physical redundancy / redundancy and robustness
- Dạng câu hỏi: Short-answer Questions
- Từ khóa: traditional frameworks emphasise, over-engineering
- Vị trí trong bài: Đoạn B, dòng 2
- Giải thích: “Traditional resilience frameworks have focused on physical redundancy and robustness” – cả hai cụm đều chấp nhận được.
Câu 40: power relations
- Dạng câu hỏi: Short-answer Questions
- Từ khóa: infrastructure systems, embody, society’s values
- Vị trí trong bài: Đoạn cuối cùng, câu cuối
- Giải thích: “infrastructure systems are ultimately social constructs embodying society’s values, priorities, and power relations” – “power relations” là đáp án chính xác.
Từ Vựng Quan Trọng Theo Passage
Passage 1 – Essential Vocabulary
| Từ vựng | Loại từ | Phiên âm | Nghĩa tiếng Việt | Ví dụ từ bài | Collocation |
|---|---|---|---|---|---|
| backbone | n | /ˈbækbəʊn/ | xương sống, nền tảng | Infrastructure forms the backbone of modern society | form the backbone of |
| reactive | adj | /riˈæktɪv/ | phản ứng lại, bị động | maintaining these assets has been reactive | reactive approach/maintenance |
| proactive | adj | /prəʊˈæktɪv/ | chủ động, tích cực | rather than proactive | proactive measures/strategy |
| revolutionising | v | /ˌrevəˈluːʃənaɪzɪŋ/ | cách mạng hóa | AI is revolutionising infrastructure monitoring | revolutionise the way |
| predictive maintenance | n | /prɪˈdɪktɪv ˈmeɪntənəns/ | bảo trì dự đoán | shift to predictive maintenance | implement predictive maintenance |
| embedded | adj | /ɪmˈbedɪd/ | được nhúng, gắn vào | Sensors embedded in bridges | embedded sensors/systems |
| real-time | adj | /ˈrɪəl taɪm/ | thời gian thực | process data in real-time | real-time analysis/monitoring |
| anomalies | n | /əˈnɒməliz/ | bất thường, dị thường | identifying patterns and anomalies | detect anomalies |
| catastrophic | adj | /ˌkætəˈstrɒfɪk/ | thảm khốc | catastrophic consequences | catastrophic failure/damage |
| deterioration | n | /dɪˌtɪəriəˈreɪʃn/ | sự xuống cấp | detect deterioration signs | material deterioration |
| vulnerable | adj | /ˈvʌlnərəbl/ | dễ bị tổn thương | vulnerable sections of network | vulnerable to attack |
| substantial | adj | /səbˈstænʃl/ | đáng kể, lớn | benefit substantially | substantial investment/cost |
Passage 2 – Essential Vocabulary
| Từ vựng | Loại từ | Phiên âm | Nghĩa tiếng Việt | Ví dụ từ bài | Collocation |
|---|---|---|---|---|---|
| sophisticated | adj | /səˈfɪstɪkeɪtɪd/ | tinh vi, phức tạp | sophisticated evolution | sophisticated system/technology |
| subset | n | /ˈsʌbset/ | tập hợp con | a subset of machine learning | subset of data |
| incipient | adj | /ɪnˈsɪpiənt/ | mới bắt đầu, sơ khai | incipient failures | incipient stage/problem |
| convolutional neural networks | n | /ˌkɒnvəˈluːʃənl ˈnjʊərəl ˈnetwɜːks/ | mạng nơ-ron tích chập | powered by CNNs | train/deploy CNNs |
| spalling | n | /ˈspɔːlɪŋ/ | vết bong tróc | identify spalling | concrete spalling |
| pioneered | v | /ˌpaɪəˈnɪəd/ | tiên phong | Singapore has pioneered | pioneer the development |
| data-driven | adj | /ˈdeɪtə drɪvn/ | dựa trên dữ liệu | data-driven approach | data-driven decision/strategy |
| unstructured | adj | /ʌnˈstrʌktʃəd/ | không có cấu trúc | unstructured text format | unstructured data |
| benchmark | n | /ˈbentʃmɑːk/ | chuẩn mực, điểm chuẩn | represents a benchmark | set a benchmark |
| rigorous | adj | /ˈrɪɡərəs/ | nghiêm ngặt, chặt chẽ | rigorous monitoring | rigorous testing/standards |
| acoustic | adj | /əˈkuːstɪk/ | âm thanh | acoustic sensors | acoustic equipment/method |
| cutting edge | n | /ˈkʌtɪŋ edʒ/ | tiên tiến nhất | cutting edge technology | at the cutting edge |
| digital twin | n | /ˈdɪdʒɪtl twɪn/ | bản sao số | digital twins of infrastructure | create digital twins |
| black box | n | /blæk bɒks/ | hộp đen (không rõ cơ chế) | function as black boxes | black box approach |
| transparency | n | /trænsˈpærənsi/ | tính minh bạch | algorithm transparency | lack of transparency |
Passage 3 – Essential Vocabulary
| Từ vựng | Loại từ | Phiên âm | Nghĩa tiếng Việt | Ví dụ từ bài | Collocation |
|---|---|---|---|---|---|
| transcends | v | /trænˈsendz/ | vượt qua, siêu việt | transcends technical considerations | transcend boundaries |
| paradigmatic | adj | /ˌpærədɪɡˈmætɪk/ | mang tính mô hình | paradigmatic shift | paradigmatic change |
| stewardship | n | /ˈstjuːədʃɪp/ | quản lý, trông coi | stewardship of critical assets | responsible stewardship |
| apprehend | v | /ˌæprɪˈhend/ | hiểu, nhận thức | to apprehend its implications | apprehend the complexity |
| resilience | n | /rɪˈzɪliəns/ | khả năng phục hồi | urban resilience | build resilience |
| untenable | adj | /ʌnˈtenəbl/ | không thể duy trì | proves increasingly untenable | untenable position |
| epistemological | adj | /ɪˌpɪstɪməˈlɒdʒɪkl/ | nhận thức luận | epistemological foundations | epistemological approach |
| causal relationships | n | /ˈkɔːzl rɪˈleɪʃnʃɪps/ | mối quan hệ nhân quả | may not reflect causal relationships | establish causal relationships |
| unprecedented | adj | /ʌnˈpresɪdentɪd/ | chưa từng có | unprecedented conditions | unprecedented scale/level |
| epistemic humility | n | /ɪˈstemɪk hjuːˈmɪləti/ | sự khiêm tốn nhận thức | epistemic humility required | demonstrate epistemic humility |
| amortise | v | /ˈæmətaɪz/ | phân bổ chi phí | amortise system costs | amortise investment |
| exacerbating | v | /ɪɡˈzæsəbeɪtɪŋ/ | làm trầm trọng hơn | exacerbating existing inequalities | exacerbate the problem |
| algorithmic governance | n | /ˌælɡəˈrɪðmɪk ˈɡʌvənəns/ | quản trị thuật toán | algorithmic governance | establish algorithmic governance |
| attributable | adj | /əˈtrɪbjʊtəbl/ | có thể quy cho | responsibility attributable | attributable to factors |
| incidentally | adv | /ˌɪnsɪˈdentəli/ | một cách ngẫu nhiên | incidentally capture information | incidentally discovered |
| paramount | adj | /ˈpærəmaʊnt/ | tối quan trọng | paramount concern | of paramount importance |
| malicious actors | n | /məˈlɪʃəs ˈæktəz/ | kẻ xấu, tác nhân độc hại | malicious actors exploit | protect from malicious actors |
| insidiously | adv | /ɪnˈsɪdiəsli/ | một cách ngấm ngầm | more insidiously corrupt | insidiously spread |
| stationary | adj | /ˈsteɪʃənri/ | tĩnh, không đổi | stationary data distributions | stationary conditions |
| non-stationary | adj | /nɒn ˈsteɪʃənri/ | không tĩnh, thay đổi | non-stationary environments | non-stationary process |
| extrapolate | v | /ɪkˈstræpəleɪt/ | ngoại suy, suy rộng | cannot extrapolate | extrapolate from data |
| indigenous | adj | /ɪnˈdɪdʒənəs/ | bản địa, nội địa | indigenous capabilities | indigenous technology |
| nascent | adj | /ˈnæsnt/ | mới nảy sinh | ethical frameworks remain nascent | nascent industry/technology |
| conjointly | adv | /kənˈdʒɔɪntli/ | cùng nhau, liên kết | addressing dimensions conjointly | work conjointly |
| technocratic | adj | /ˌteknəˈkrætɪk/ | theo chủ nghĩa kỹ trị | narrow technocratic perspectives | technocratic approach |
Kết Bài
Chủ đề “AI In Predictive Maintenance Of Infrastructure” không chỉ phản ánh xu hướng công nghệ hiện đại mà còn thể hiện cách IELTS Reading đánh giá khả năng đọc hiểu văn bản học thuật phức tạp của thí sinh. Qua ba passages với độ khó tăng dần từ Easy đến Hard, bạn đã được trải nghiệm một bài thi hoàn chỉnh với 40 câu hỏi thuộc 7 dạng bài khác nhau – hoàn toàn giống với đề thi thật.
Passage 1 giới thiệu khái niệm cơ bản về AI trong bảo trì cơ sở hạ tầng với ngôn ngữ dễ tiếp cận, giúp bạn làm quen với chủ đề và rèn luyện kỹ năng scanning thông tin cụ thể. Passage 2 đi sâu vào các công nghệ machine learning và ứng dụng thực tế tại nhiều quốc gia, yêu cầu khả năng hiểu paraphrase và suy luận cao hơn. Passage 3 đặt ra thách thức lớn nhất với phân tích đa chiều về các khía cạnh kinh tế, xã hội và đạo đức, đòi hỏi trình độ đọc hiểu học thuật ngang band 7.0-9.0.
Phần đáp án chi tiết không chỉ cung cấp đáp án đúng mà còn giải thích rõ ràng vị trí thông tin, cách paraphrase giữa câu hỏi và passage, cũng như chiến lược làm bài cho từng dạng câu hỏi. Bộ từ vựng được phân loại theo passage giúp bạn học một cách có hệ thống, tập trung vào những collocations và academic words thường xuất hiện trong IELTS.
Hãy sử dụng bộ đề này như một công cụ tự đánh giá thực lực, luyện tập quản lý thời gian và hoàn thiện kỹ thuật làm bài. Đừng quên rằng việc hiểu sâu từng câu trả lời quan trọng hơn nhiều so với việc chỉ biết đáp án đúng. Chúc bạn đạt band điểm mục tiêu trong kỳ thi IELTS sắp tới!
