From Technical Challenges to Successful Implementation: A Case Study of a Municipal Pipe Network Project

城市管网 2026-11-15 5 min read
From Technical Challenges to Successful Implementation: A Case Study of a Municipal Pipe Network Project
Drainage Pipe Network Liquid-Level and Flow Monitoring and Waterlogging Prevention Early Warning Project in the Old Urban District of a Certain City: ...

From Technical Challenges to Successful Implementation: A Case Study of a Municipal Pipe Network Project

Yicai Technology News: During the 2024 flood season, the drainage pipe network liquid-level and flow monitoring and waterlogging prevention early warning project in the old urban district of a certain city completed its first full flood-season operation.

The project covers approximately 36 square kilometers of the old urban district and deployed 186 front-end monitoring points, including 128 liquid-level monitoring points, 38 flow monitoring points, 12 rainfall monitoring points, and 8 video monitoring points. It connected 4 drainage pump stations, 6 gates, and 3 mobile pump trucks. Through coordinated liquid-level, flow, rainfall, video, and dispatch data, the project formed a closed loop of "monitoring-warning-dispatch-review". Flood-season operating data showed that the system average warning lead time was about 38 minutes, the duration of waterlogging events exceeding 0.5 m decreased from an average of 4.2 hours for historical rainfall events of comparable magnitude to 1.7 hours, and the data completeness rate reached 99.2%, providing quantifiable support for urban drainage and waterlogging prevention.

Drainage pipe network liquid-level and flow monitoring system overview with 186 monitoring points

Project Background: The Old Urban District Drainage System Faces the Pressure of "Cannot See, Cannot Dispatch Accurately, Cannot Respond Quickly"

The built-up area of the old urban district of a certain city is approximately 36 square kilometers, and the total length of the drainage pipe network is approximately 860 kilometers, including about 420 kilometers of stormwater pipes and channels, about 310 kilometers of combined sewer pipes and channels, and about 130 kilometers of sewage and interception pipes and channels. Roads, buildings, and underground pipelines are dense in the area. Most drainage facilities were built in the 1980s and 1990s, with pipe diameters mainly DN300 to DN1000, and some trunk pipes DN1200 to DN1500. Historically, the old urban district had 23 low-lying waterlogging-prone points. Road waterlogging during the flood season not only affected traffic but also caused underground garage flooding, damage to roadside shops, and disruption to residents' travel.

Before the project started, local drainage management mainly relied on manual patrols, rainfall stations, and pump station operating records. Statistics from the 2021-2023 flood seasons showed that the old urban district experienced a cumulative 212 waterlogging events of varying degrees, including 72 events with waterlogging depth exceeding 0.3 m and 18 events exceeding 0.5 m. Because real-time liquid-level and flow data inside the pipe network were lacking, management authorities found it difficult to determine whether waterlogging was caused by rainfall intensity, pipe network bottlenecks, pump station capacity, or downstream backwater. Dispatch often remained at the passive stage of "pumping only after water overflows".

With the increasing frequency of extreme rainfall, urban waterlogging management shifted from "post-event response" to "pre-event early warning, in-event dispatch, and post-event review". In 2023, the city launched a drainage pipe network monitoring project, aiming to establish a liquid-level and flow monitoring system covering trunk pipes, waterlogging-prone points, pump station forebays, and critical nodes, and to coordinate it with the drainage dispatch platform. The project required monitoring data to provide pipe network risk assessment at least 30 minutes before heavy rain, support coordinated dispatch of pump stations, gates, and mobile pump trucks during rainfall, and form correlation analysis between waterlogging events and pipe network operating parameters after the flood season.

Technical Challenges: Aging Pipe Network, Partially Full Pipe Flow, and Multi-Source Threshold Linkage Pose Engineering Challenges

Challenge 1: Incomplete Basic Data

Field reconnaissance found deviations between design drawings and actual pipe positions, diameters, and burial depths. About 12% of pipe segments had positional deviations exceeding 1.5 m, and about 8% of manhole numbers were inconsistent with the drawings. Installation of level sensors and flow meters required determining manhole chamber dimensions, manhole bottom sediment, pipe diameter, burial depth, upstream-downstream relationships, and maintenance access. Some manhole openings were only 500 mm wide, making conventional installation difficult.

Challenge 2: Partially Full Pipe Flow Measurement

Stormwater and combined sewer pipes were not full in most periods. Initial tests showed ordinary ultrasonic flow meters could have errors of 12% to 18% under low liquid-level conditions. The project adopted a "level sensor + cross-section flow meter" combination using Doppler, electromagnetic, and radar flow meters, targeting errors within 5% after calibration for siltation and geometry.

Challenge 3: Underground Environment & Communication

Monitoring points faced high humidity (80%-95%) and corrosive gases like hydrogen sulfide. The project used IP68-rated equipment, anti-corrosion brackets, and NB-IoT/4G/LoRa communication. For 26 points without mains power, solar panels and lithium batteries ensured data collection every 5 minutes even under 7 days of overcast weather.

Challenge 4: Dynamic Warning Thresholds

Nonlinear relationships exist among liquid level, road waterlogging, and rainfall. Initial false alarm rates reached 27%. Subsequent joint analysis of video, rainfall, liquid-level, flow, and pump data established tiered thresholds and dynamic adjustment rules, aligning warnings closer to actual road waterlogging risk.

Solution: Coordinated Liquid-Level, Flow, Rainfall, and Video Sensing Builds a Drainage and Waterlogging Prevention Monitoring Network

The project adopted a technical route of "front-end sensing, edge transmission, platform analysis, and dispatch coordination". The front-end sensing layer includes radar and ultrasonic level sensors, Doppler and partially full pipe flow meters, tipping-bucket rain gauges, and road waterlogging recognition cameras. Points were strategically arranged at trunk pipe nodes, upstream of waterlogging-prone areas, pump station forebays, gate wells, and key road sections.

"During the flood season, the overall online rate of the 186 front-end points remained at 98.7%, and the data completeness rate was 99.2%."

The data transmission layer adopted hybrid MQTT and HTTP protocols with breakpoint resume support. Liquid-level data uploaded every 5 minutes, flow data every 15 minutes, and rainfall data every 1 minute. The platform layer established a drainage and waterlogging prevention one-map, overlaying GIS pipe networks with real-time status. Three-level thresholds (Yellow, Orange, Red) were set based on liquid levels and adjusted with future 30-minute rainfall forecasts. For example, at a DN800 node, 0.8 m triggers yellow, 1.2 m triggers orange, and 1.6 m triggers red.

The dispatch layer coordinates 4 pump stations (26 m³/s capacity), 6 remotely controlled gates, and 3 mobile pump trucks (500-800 m³/h each). The system generates recommendations such as pre-lowering liquid levels, early pump startup, and flow diversion, recording execution status to form a closed loop.

Implementation Process: Three-Phase Construction from Survey Pilot to Flood-Season Operation

The project started in June 2023, main construction was completed in June 2024, and entered real flood-season operation from July to September 2024. Total investment was about RMB 15.60 million. The implementation was divided into three phases:

  • Phase 1 (June-August 2023): Data Survey & Reconnaissance. Reviewed 860 km of pipe network and surveyed 186 candidate points. Ground-penetrating radar was used for 38 points with significant drawing deviations to avoid installing equipment on wrong segments.
  • Phase 2 (September 2023): Pilot Construction & Commissioning. Installed equipment at 20 representative points. After addressing issues like low-flow errors and moisture-induced failures, standards for full construction were finalized, improving data completeness from 91.3% to 98.6%.
  • Phase 3 (October 2023 - March 2024): Full Construction & Platform Commissioning. Advanced in three work sections using nighttime operations. All 186 points were installed and connected to SCADA and control systems by March 2024.

From April to June 2024, model calibration reduced the system false alarm rate from 27% to 8% and the miss rate from 11% to 3% through three rounds of threshold adjustments based on historical rainfall events.

Data Governance and Model Calibration: Making Liquid-Level and Flow Data Truly Support Waterlogging Dispatch

The project established strict data governance rules. Zero-point calibration was performed based on manhole bottom elevation, and weekly comparisons generated drift alarms if exceeding 2 cm. Flow meters were calibrated with cross-section measurements accounting for sediment thickness. The 186 monitoring points were classified into four categories: trunk-pipe bottleneck points, upstream nodes of waterlogging-prone points, pump-station dispatch nodes, and outfall backwater nodes.

Case Study: Identifying Pipe Blockages via Data

For a DN1000 combined sewer, design flow was 1.6 m³/s, but monitoring showed only 0.9 m³/s at a 1.3 m liquid level. This discrepancy indicated siltation or deformation. On-site dredging found 20 cm sediment and localized collapse. After repair, flow increased to 1.1 m³/s under the same level, improving conveyance capacity by 22%. This demonstrates how liquid-level and flow data can reverse-identify pipe network defects.

Threshold management adopted "static thresholds + dynamic adjustments". Static thresholds were based on pipe geometry and history, while dynamic adjustments considered future rainfall forecasts, change rates, and pump status. The system monitored liquid-level change rates; a rise of >0.2 m in 10 minutes triggered high-attention warnings. During the 2024 flood season, this feature discovered 3 potential blockages in advance, allowing timely dredging to prevent larger-scale waterlogging.

Final Results: Quantified Improvements in Warning Lead Time, Waterlogging Duration, and O&M Efficiency

38 min Average Warning Lead Time
1.7 h Waterlogging Duration (>0.5m)
99.2% Data Completeness Rate

During the 2024 flood season, the system issued 146 yellow, 58 orange, and 17 red warnings. Dispatch response time shortened from an average of 45 minutes to 12 minutes. Waterlogging events exceeding 0.3 m occurred 19 times, down 20.8% from previous years. Events exceeding 0.5 m dropped significantly compared to historical averages.

On July 18, 2024, with 118 mm of rainfall, the system triggered an orange warning 62 minutes prior. Early pump startup and pre-deployment of mobile pumps resulted in only 1 road section with >0.5 m waterlogging (down from an average of 5), and duration reduced to 1.7 hours.

Operational efficiency improved as manual manhole inspections were replaced by continuous online monitoring. Manhole opening inspections were reduced by about 210 times per storm, lowering confined-space safety risks. Economically, precise data enabled targeted repairs instead of full replacements, saving about RMB 1.80 million in engineering costs. Traffic control and pre-positioning of pumps avoided prolonged closures, saving hundreds of thousands of RMB in social costs per event.

Case Insights and Outlook: From Single-Point Monitoring to Urban Drainage Resilience Governance

This project demonstrates that drainage monitoring is not just about installing instruments, but incorporating pipe network status into a perceivable, analyzable, and dispatchable system. Key insights include:

  • Strategic Point Deployment: Focus on waterlogging-prone points, bottlenecks, and pump nodes rather than uniform coverage.
  • Flow Accuracy: Combine partially full pipe conditions with siltation correction for reliable flow data.
  • Dynamic Thresholds: Use flood-season review and model calibration to avoid false/missed alarms.
  • Data Governance: Prioritize O&M and data quality (drift, connectivity) to ensure judgment accuracy.

In the next phase, the city plans to extend monitoring scope to 78 km², adding 80 liquid-level, 25 flow, and 15 video points. Integration of nowcasting rainfall forecasts will further improve warning lead times. Future explorations will focus on pipe network siltation identification, structural defect early warning, and intelligent dispatch, moving towards resilience governance of the urban drainage system.

🔧 Recommended Products

Based on this article, we recommend the following Ecolor products:


24h Fast Response

Need Help Choosing the Right Product?

Our engineering team is ready to provide free technical consultation and quotes

Share this article: