Model-Driven Sewage System Design and Intelligent Management of the Wuhan East Lake Deep Tunnel Drainage Project

Abstract

Rapid urbanization in China has overwhelmed traditional drainage systems, resulting in frequent flooding and water pollution in densely populated urban areas. This study focuses on the East Lake core area of Wuhan, proposing a deep tunnel drainage system to improve sewage storage and conveyance capacity. A pilot-scale pipe model was employed to determine the critical non-silting velocity for full-pipe sewage flow. Based on projected dry-season inflows and intercepted combined sewer discharges, the design capacities for pumping stations and pretreatment facilities were defined. A three-dimensional gas–liquid two-phase numerical model was used to simulate inflow shaft hydraulics at Erlangmiao, Luobuzui, and Wudong pretreatment stations. Simulation results confirm that all shafts meet energy dissipation and ventilation requirements, with uniform flow and velocity distributions that could be obtained by a vortex-type shaft. The system not only mitigates regional environmental challenges but also shows significant social, environmental, and economic benefits. Overall project design, applied methodology, simulation study, and outcomes could provide a valuable reference to deep tunnel drainage design and research.

1. Introduction

Wuhan, a megacity located in the Yangtze River Basin, is characterized by low-lying terrain and a dense water network [1]. During the rainy season, urban waterlogging occurs frequently, and traditional drainage systems are often unable to cope with extreme weather events [2]. With the acceleration of urbanization, these systems are facing increasing challenges [3]. In the urban core, wastewater treatment plants are constrained by multiple issues, including conflicts between land demand and urban development, the urgent need to improve ambient water quality, insufficient sanitary protection distances, the necessity of upgrading and expanding treatment facilities under limited land availability, and incomplete supporting sewer networks [4]. As the shallow underground space has already been saturated, it is necessary to extend development into deeper strata (below 30 m) to achieve combined sewer separation and enhance flood control and drainage capacity [5]. Deep tunnel drainage technology, as an innovative solution, offers significant advantages in optimizing sewage storage and conveyance, improving drainage efficiency, and mitigating combined sewer overflows [6].

At present, deep tunnel drainage technology in China is still at the stage of experimentation and exploration [7]. The East Lake deep tunnel project, as the first urban wastewater deep conveyance tunnel formally constructed and put into operation in the country, has filled the gap in practical application in this field. Prior to this, only the DongHaoChong deep tunnel project in Guangzhou had been carried out as a pilot project. This project was mainly used for stormwater storage and regulation, and although it provided some practical data, its reference value for the promotion and application of deep tunnel projects in other regions was limited [6].

Although a series of exploratory studies on deep tunnel technology have been conducted in China in recent years, most of them have focused on the introduction and analysis of foreign cases [8,9]. A systematic and theoretical research framework has not yet been established, and the adaptability of localized deep tunnel projects, key technological breakthroughs, and engineering experience remains relatively insufficient. Therefore, the East Lake deep tunnel project, developed in the absence of mature domestic experience, is a typical case of independent innovation, in which localized research and key technological domestic development were carried out to meet the needs of water environment management in China. This project, therefore, holds important research value. At present, systematic research on models of independent innovation in deep tunnel projects remains limited in both academia and industry, which provides an important space for further study.

Therefore, this study develops a model-driven approach for sewage system design and intelligent management in the East Lake deep tunnel project. Focusing on the East Lake core area of Wuhan, the research establishes a deep tunnel drainage model to enhance sewage storage and conveyance capacity. Pilot-scale experiments and three-dimensional gas–liquid two-phase simulations of inflow shafts at key pretreatment stations were conducted to support hydraulic optimization and intelligent control. Based on these experimental and simulation results, the optimized deep tunnel system demonstrates strong hydraulic stability and operational efficiency. Consequently, the proposed system effectively mitigates regional flooding and pollution, raising the drainage protection standard in the central area from a “1-in-1-year” to a “1-in-10-year” event and reducing waterlogging risk by up to 80%. To further validate and highlight its advantages, this study compares the deep tunnel system with traditional drainage approaches (

Table 1. Comparison between the Deep Tunnel System and Traditional Drainage Approaches.

Index	Deep Tunnel System	Traditional Drainage	Comparative Conclusions
Payback period	15 to 50 years	10 to 20 years	Short-term advantage of the traditional method
Land occupancy rate	Main tunnel: zero surface occupation; regulation tank: occupies 0.15 km2	2.6 km2 per plant (Shanghai case)	Deep tunnel saves 90%
Water Quality Compliance Rate	CSO control rate ≥ 92% (measured in 2023)	CSO control rate: 60–75%	Deep tunnel improves by 25%

).

2. Methodology

2.1. Project Background

The East Lake core area is served by twelve distinct sewage collection systems. Among them, the Wugang Industrial, Chemical New Town Industrial, and Ethylene Industrial sewage systems primarily collect and treat industrial wastewater, with each system and its associated treatment plant operating independently. The remaining nine systems mainly collect and treat municipal wastewater, with six treatment plants already constructed, while the Baiyushan and Gujiashan plants are still under development. Other key sewage systems in the East Lake core area include the Shahu, Erlangmiao, Luobuzui, and Baiyushan treatment systems.

Rapid urban expansion has resulted in land-use conflicts with existing wastewater treatment infrastructure, posing challenges for harmonious urban development. The Shahu Wastewater Treatment Plant (WWTP) was established in the 1980s, originally sited at the edge of urban development. With urban expansion and the growth of the city’s central area, the Shahu WWTP gradually became part of the urban core and is now the only wastewater treatment plant within Wuhan’s second ring road. The plant’s presence is increasingly incompatible with the surrounding functional zoning, affecting regional development. The Erlangmiao Wastewater Treatment Plant is located within the urban economic center, which restricts the potential value of surrounding land and affects the quality of residential and commercial development in the vicinity. The Luobuzui WWTP was constructed northeast of Wuhan Railway Station, adjacent to the Third Ring Road. With the completion of Wuhan Railway Station and the gradual development of the Yangchunhu urban sub-center, the potential future relocation of the Wugang and Qingshan thermal power plants is expected to promote land value appreciation in the surrounding area. It can therefore be anticipated that the Luobuzui WWTP will face similar land-use conflicts in the future.

In addition, as environmental regulations become more rigorous, discrepancies between current sewage treatment facilities and ecological preservation goals have become more pronounced. The East Lake area features a well-developed water network, including six major lakes: East Lake, Shahu, Yangchunhu, Yanxi Lake, Yandong Lake, and Beihu. These lakes discharge into the Yangtze River through the Xingsheng Road Pumping Station, Luojiagang, Qingshan Port, Beihu Sluice, and Beihu Grand Port, respectively. According to the 2013 Wuhan Environmental Status Bulletin, water quality in Shahu is classified as Class V, while the other lakes in the East Lake region are classified as Class IV. In addition, the Shahu Wastewater Treatment Plant is surrounded on three sides by residential areas, but no odor-control facilities have been implemented. The Erlangmiao WWTP is also adjacent to residential neighborhoods; all open structures have been covered and equipped with deodorization systems, which temporarily mitigate odor impacts on nearby residents. The Luobuzui WWTP is located near two schools and numerous residential communities. To address insufficient sanitary protection distances, its anoxic and anaerobic tanks have been constructed as enclosed structures, and deodorization measures have been implemented. The water quality status of the East Lake area is shown in

Table 2. Water Quality Status in the East Lake Area.

Lake	Administrative Districts	Primary Water Function Zone	Lake Area (km2)	Water Quality Management Targets	2014 Water Quality Assessment
East Lake	East Lake Ecotourism Scenic Area	East Lake Development and Utilization Zone	33.989	Class III	Class IV
Sha Lake	Wuchang District East Lake Ecotourism Scenic Area	Sha Lake Preservation Zone	3.078	Class IV	Below Class V
Yangchun Lake	Hongshan District	Yangchun Lake Preservation Zone	0.576	Class IV	Class IV
Yanxi Lake	East Lake High-Tech Development Zone	Yanxi Lake Preservation Zone	14.231	Class III	Class IV
Yandong Lake	East Lake High-Tech Development Zone	Yandong Lake Preservation Zone	9.111	Class III	Class IV
Bei Lake	Qingshan District	Bei Lake Development and Utilization Zone	1.933	Class V	Class IV

.

2.2. Overview of East Lake Drainage Scheme

The design of deep tunnel systems involves numerous key technologies, including the connection between deep drainage tunnels and surface collection pipelines, energy dissipation in vertical shafts, transient flow control, surge control in the main tunnel, effluent flow regulation, ventilation and odor control during operation, and sediment management in the main tunnel [10,11,12,13].

The deep tunnel system needs to be coupled with the existing shallow drainage network to form a three-dimensional urban drainage system, collectively handling the city’s stormwater and wastewater [14]. Operational modes vary slightly depending on functional requirements. The overall planning layout of the East Lake core area follows a structured “One Plant, Two Lines, Three Tunnels” framework. “One Plant” refers to the demolition of the existing wastewater treatment facility and the construction of the new Beihu Wastewater Treatment Plant. “Two Lines” denotes two drainage tunnels along the northern alignment and one along the southern alignment. “Three Tunnels” consists of the near-term northern wastewater conveyance tunnel, the near-term northern stormwater regulation tunnel, and the long-term southern combined sewer tunnel. The spatial configuration of these components is illustrated in Figure 1.

Accordingly, the main components and core Functions of the Wuhan East Lake Deep Tunnel Project are summarized as follows:

1. East Lake Core Area Deep Tunnel System: The system adopts a “main tunnel + branch tunnel” configuration, primarily designed to convey wastewater from the Wuchang district to the Beihu Wastewater Treatment Plant for centralized treatment. It includes: (1) The Main Tunnel starts from Erlangmiao Pretreatment Station and ends at Beihu WWTP. The main tunnel is laid in the deep underground space of Wuchang, Hongshan, and Qingshan districts, passing through areas previously served by Shahu, Erlangmiao, and Luobuzui treatment plants. The total length is approximately 17.5 km, constructed using the shield tunneling method, with an inner diameter of 3.0–3.4 m, burial depth of 30–42 m, and slope of 0.00065 (≈0.065%). (2) Branch Tunnel, which connects to the main tunnel at the Qinghua Interchange Node from the Luobuzui Pretreatment Station. It is approximately 1.7 km long, constructed using the jacking method, with an inner diameter of 1.5 m, burial depth of 20–21 m, and slope of 0.0005 (≈0.05%).

2. Four Pumping Stations: The systems which pressurize surface wastewater into the deep tunnel (e.g., Erlangmiao and Luobuzui stations), with local sections operating under pressure flow conditions.

3. Intelligent Control System: Real-time monitoring of flow, velocity, and sedimentation conditions allows dynamic adjustment of pump station operation.

4. Terminal Wastewater Treatment Plant: Wastewater is conveyed to the Beihu WWTP, currently the largest underground WWTP in Asia, for centralized treatment. The plant’s long-term capacity is planned to reach 1.5 million m³/day.

Therefore, the East Lake Deep Tunnel System integrates three core functions: efficient wastewater conveyance to replace aging pipelines and prevent combined sewer overflows during rainfall events; interception of first-flush runoff to mitigate non-point source pollution; and anti-silting operation achieved through intelligent flushing mechanisms and optimized hydraulic design to ensure long-term sediment-free performance.

2.3. Methodology and Models for East Lake Drainage Simulation

Before the completion of the Wuhan East Lake deep tunnel, the research team implemented a multi-level simulation and analysis framework following a macro-coupled-micro-intelligent approach. At the macro level, SWMM (Storm Water Management Model) was used to simulate rainfall-runoff and pollutant loads in the urban drainage system, identifying the extreme inflows and pollutant ranges that the deep tunnel must accommodate [15,16]. The urban drainage system was divided into 58 sub-catchments, and inflow boundary conditions were set based on the spatial distribution of wastewater treatment plants and pretreatment stations, including Erlangmiao and Luobuzui. The drainage area of each inflow node ranged from approximately 35 to 60 ha, with an imperviousness of about 75%. The rainfall input was designed with a 100-year return period and a duration of 3 h, which represents typical rainfall characteristics of Wuhan city. The initial loss and depression storage were set to 1.5 mm for impervious surfaces and 5.0 mm for pervious surfaces. The Horton infiltration model was applied with an initial infiltration rate of 80 mm/h, a minimum infiltration rate of 10 mm/h, and a decay coefficient of 4.0 h⁻¹. The Manning roughness coefficient of the pipeline was set to 0.014, and the computational time step was 5 s. Typical rainfall events were then simulated to calculate runoff, peak flow, and pollutant loads. From these results, the critical non-silting velocity under full-sewage conditions was derived.

Then, a water quality model was developed to simulate the spatial–temporal distribution of suspended solids (SS) and thus identify potential sedimentation zones within the deep tunnel system. Subsequently, SWMM was coupled with ICM (InfoWorks Integrated Catchment Modeling) to simulate the overall hydraulic distribution of the pipeline network, optimize inlet storage tank capacities and pump station layouts, and ensure effective integration between the deep tunnel and the existing urban drainage system [17,18]. Within the water quality module, pollutant buildup and washoff processes were incorporated, which were, respectively, modeled using an (1) exponential function and (2) power function:

$$B\left(t\right)=C_1(1-e^{-C_{2}t})$$

(1)

$$W=C_w{q}^{\alpha }B$$

(2)

where C₁ is the maximum buildup (kg/ha), C₂ is the buildup rate constant (day⁻¹), q is the unit runoff (mm/h), C_w is the washoff coefficient, and α is the exponent.

Based on typical parameter ranges reported in domestic and international studies, as well as measured data from the East Lake Basin, pollutant parameters were initially set as follows:

TSS: C₁ = 200 kg/ha, C₂ = 0.5, Cw = 0.2 and α = 1.0

COD: C₁ = 40 kg/ha, C₂ = 0.5, Cw = 0.2 and α = 0.8

TN: C₁ = 5 kg/ha, C₂ = 0.2, Cw = 0.2 and α = 0.7

TP: C₁ = 2 kg/ha, C₂ = 0.15, Cw = 0.5 and α = 0.7

Pollutant concentrations in rainfall were assumed as: TSS = 1.5 mg/L, COD = 0.5 mg/L, TN = 0.3 mg/L, and TP = 0.05 mg/L. Pollutant transport in the pipeline was modeled with first-order decay (rate = 0.2 day⁻¹) to account for sedimentation and biodegradation.

To characterize the micro-scale hydraulics and sediment deposition at the deep tunnel inflow shaft-tunnel interfaces, a CFD-DEM coupled micro-scale model was employed. The CFD component is based on transient RANS equations and the VOF multiphase flow method, combined with adaptive mesh refinement (near-wall Δx = 0.01 m), to calculate hydraulic jump heights, energy losses, and local cavitation risks within the shaft [19]. The DEM component, based on the Newton–Euler equations and Hertz–Mindlin contact model, tracks the motion of suspended particles to predict potential sedimentation zones, with GPU parallel computing used for acceleration [20]. Sediment resistance coefficients are dynamically predicted using a graph neural network (GNN), improving the accuracy of deposition and sedimentation simulations, thereby guiding shaft-tunnel interface design, energy dissipation device layout, and lining anti-scour measures [21].

Finally, emergency scenario simulations were conducted to evaluate the system’s resilience under extreme conditions, including WWTP shutdowns, nighttime low flows, and stormwater overflows [22,23]. This progressive framework, from macro to micro scales and from design to operation, could enable comprehensive validation and preemptive risk control before construction completion.

3. Result

3.1. Simulation Output Analysis and Intelligent Response

3.1.1. SWMM Modeling for East Lake Tunnel

Based on the overall SWMM calculations, a hydraulic model of the East Lake deep tunnel was established. In this model, the deep tunnel was idealized as a large-diameter conduit with a uniform slope, with a total length of 19.2 km. The model was constructed using pipeline node coordinates, elevations, and slope data, with Erlangmiao and Luobuzui pretreatment stations as inflow boundaries and the terminal pumping station as the outflow boundary. The pipeline Manning coefficient was set to 0.014, and the time step was 5 s. The hydraulic model primarily computed flow distribution, hydraulic grade lines, and energy losses under various design conditions, enabling the assessment of conveyance capacity under extreme rainfall events and verification of operational safety. Figure 2 illustrates the simplified hydraulic model of the East Lake tunnel.

The East Lake deep tunnel achieves a technological leap through the integration of a distributed fiber optic sensing network (physical layer), CFD-DEM-AI coupled model (algorithm layer), and edge-intelligent control (execution layer), representing a “Chinese paradigm” for global deep tunnel operation and maintenance. According to the Wuhan Water Authority (2024) Annual Assessment of the East Lake Deep Tunnel Intelligent System, comparisons between the ICM/CFD-DEM-AI model and conventional CFD-DEM model during the 2023 extreme rainfall event indicate significant improvements in predictive accuracy and operational control (see Figure 3 and

Table 3. Comparison between the ICM/CFD-DEM-AI Model and the Traditional CFD-DEM Model.

Indicator	Conventional CFD-DEM Model	ICM/CFD-DEM-AI Model of This Project	Performance Improvement
Sedimentation Prediction Error	38.5%	4.3%	89%
Dredging Localization Accuracy	±3.2 m	±0.5 m	84%
Response Delay	2.1 h	8 min	94%

).

3.1.2. Intelligent Operation and Anti-Sedimentation System

The East Lake deep tunnel drainage system implements an integrated anti-sedimentation strategy based on multi-scale modeling and intelligent control. At the macro scale, the SWMM model simulates rainfall-runoff processes and predicts peak inflows under typical storm scenarios. Incorporating the spatial distribution of pretreatment facilities such as Erlangmiao and Luobuzui, the model calculates the critical non-silting velocity and guides the tunnel’s hydraulic design. The critical non-silting velocity can be calculated according to the following equation:

$$V_c=\sqrt{{\displaystyle \frac{4g{d}_{50}({\rho }_s-{\rho }_w)}{3C_D{\rho }_w}}}\cdot K_T\cdot K_S$$

(3)

where d₅₀ is the median sediment particle diameter (0.15 mm for Wuhan), C_D is the drag coefficient (1.2), K_T represents the temperature correction (K_T = 1 + 0.02(T − 20)), and K_S is the pipe wall roughness correction (1.05 for concrete pipes). For the East Lake deep tunnel, V_C = 0.65 m/s was obtained with a safety factor of 20%.

At the micro-scale, the system simulates sediment transport through a coupled CFD-DEM model, which captures hydraulic jumps, vortex patterns, and particle trajectories at shaft-tunnel interfaces. Sediment processes are divided into two stages: deposition (4) and scour-resuspension (5) [24].

$$M_s\left(t\right)=M_{max}\times {\displaystyle \frac{v(t-t_0)}{h}}$$

(4)

$$M_w\left(t\right)=k_v\times M_s\left(t\right)$$

(5)

where M_s (t) is the mass of deposited sediment at time t (kg); M_max is the maximum sediment mass (kg); v is the sedimentation rate (m/s); h is the water depth (m); and t₀ is the initial time when deposition begins. M_w (t) is the mass of sediment scoured at time t (kg), and k_v is the scour coefficient, which can be calibrated using underwater monitoring data.

These models are augmented by an AI-based real-time corrector, which integrates LSTM for rainfall prediction, graph neural networks (GNNs) for sediment–flow interaction, and reinforcement learning for dynamic adjustment of control strategies. The full model architecture, combining physical sensing (fiber optic sensors, turbidity probes) and real-time analytics, is summarized in

Table 4. Model Input–Output Architecture.

Data Type	Example Parameters	Collection Method	Update Frequency
Flow Velocity/Discharge	Radar Flow Velocity Meter (v), Ultrasonic Flow Meter (Q)	50 m/point Fiber Optic Sensor Network	100 Hz
Water Quality Parameters	Turbidity (NTU), TP, COD, H2S	Multispectral Online Monitoring Instrument	5 min
Structural Condition	Lining Strain (με), Joint Displacement (mm)	Distributed Fiber Optic Sensing (DTS)	Deep Tunnel Improvement of 25%
Historical Operation and Maintenance	Dredging Volume (m3), Corrosion Depth (mm/year)	Robotic Scanning Database	Event Trigger
Rainfall Forecast	Rainfall Intensity for the Next 3 Hours (mm/h), Storm Return Period	Meteorological Radar + Machine Learning Forecasting	15 min
Topography and Geology	Pipe Slope (‰), Soil and Rock Permeability Coefficient (cm/s)	BIM Geological Model	Static Parameters

.

To operationalize this system, intelligent monitoring and sediment removal are triggered when either sediment thickness exceeds 0.3 m or flow velocity drops below 0.8 m/s. Sensors and robotic laser scanning (±1 mm accuracy) provide high-resolution feedback every 500 m. Performance validation in 2023 showed strong agreement between predicted and measured sediment loads (R² = 0.93).

Table 5. Processing Methods after Sedimentation Prediction.

Stage	Technical Scheme	Key Parameters
Triggering Mechanism for Desilting	Real-time Sediment Monitoring + Hydraulic Model Prediction	Initiated when sediment thickness ≥ 0.3 m or flow velocity < 0.8 m/s
Main Desilting Approach	High-pressure Hydraulic Flushing + Negative-pressure Suction (No Mechanical Entry into Tunnel)	Flushing pressure ≥ 35 MPa; suction flow rate 300 m3/h
Sludge Treatment	In situ Pipeline Dewatering + On-ground Centrifugal Dewatering	Moisture content reduced from 98% to below 60%
Resource Utilization	Dewatered Sludge Utilized for Lightweight Aggregate Production (Resource Utilization Rate > 85%)	Compliant with Sludge Disposal for Brick-making in Municipal Wastewater Treatment Plants (GB/T 25031-2010) [25]

provides an overview of processing methods after sedimentation prediction involved in tunnel maintenance.

The main desilting process is a high-pressure hydraulic flushing system comprising fixed nozzles (spaced every 200 m) and mobile flushing robots for shaft and bend regions. Each 8-h operation can remove 800–1200 m³ of sediment. Additionally, a patented spiral flow generator enhances scouring by injecting tangential jets in low-velocity zones, increasing effective velocity. This system demonstrated a 32% increase in actual flow velocity (from 0.5 m/s to 0.66 m/s), and successful flow recovery verification is achieved when velocity remains above 0.75 m/s for 2 consecutive hours.

Following desilting, the collected sludge undergoes in situ dewatering and is further processed on the surface using centrifugal dehydration. The treated sludge is then utilized for resource recovery, producing ceramsite lightweight aggregates for construction. With a resource utilization rate over 85%, the plant yields 35,000 m³ of aggregate annually, reducing landfill demand by 1.2 hectares. This approach complies with national standards (GB/T 25031-2010), supporting the sustainable transformation of urban tunnel systems.

Table 6. Technical comparison of Wuhan East Lake Deep Tunnel with other international Tunnel projects in terms of generational Monitoring Technologies.

Technology Aspect	London Thames Tideway Tunnel	Tokyo Metropolitan Outer Area Underground Discharge Channel	Wuhan East Lake Deep Tunnel
Velocity Monitoring Density	500 m per point	300 m per point	50 m per point
Sedimentation Prediction Model	Empirical Formula	Two-dimensional Hydrodynamic Model	CFD-DEM-AI Coupled Model
Response Timeliness	Manual Analysis (≥4 h)	Semi-automatic (≈1 h)	Fully automatic (≤10 min)
Desilting Accuracy	±5 m	±3 m	±0.5 m

shows the technical comparison of the East Lake deep tunnel with other international deep tunnel projects, which demonstrates its significant technological advance in intelligent monitoring and sediment management. The high velocity monitoring density (50 m per point) renders the East Lake deep tunnel project with a fine-grained, real-time assessment of flow dynamics across the entire network. By integrating physical simulation and data-driven learning, the CFD-DEM-AI coupled sedimentation prediction model allows adaptive and highly accurate forecasting of sediment transport and deposition. As a result, the East Lake deep tunnel project displays very low system response time (≤10 min) and super high desilting accuracy (±0.5 m). The tunnel achieves fully automatic response and precise self-optimizing operation, which establishes a new technological benchmark for smart, resilient, and sustainable deep tunnel infrastructure worldwide.

3.1.3. Operational Scenarios and Emergency Dispatch Simulation

Based on potential emergencies in actual operation, three emergency dispatch scenarios were selected for system simulation testing: partial shutdown of downstream wastewater treatment plants, insufficient nighttime inflow in upstream areas, and overflow risks in upstream pipelines caused by rainfall.

(1) Partial Shutdown of Downstream Wastewater Treatment Plant

The downstream wastewater treatment plant has a daily treatment capacity of 800,000 m³. When certain treatment lines are taken offline for maintenance, the overall treatment capacity will be significantly reduced. To ensure that the deep tunnel system continues to operate under design conditions, inflow rates at the upstream pretreatment stations must be adjusted accordingly. In this scenario, the system’s dispatch target is set to 400,000 m³/d, with all four pretreatment stations remaining available for coordinated scheduling.

(2) Insufficient Nighttime Inflow in Upstream Areas

During nighttime hours, inflow from upstream areas may decrease substantially, potentially causing the water level in the upstream pretreatment station’s pump sump to drop below the operational threshold. This may necessitate the temporary shutdown of inflow pumps at that station. To maintain system balance, inflows at other pretreatment stations must be adjusted to compensate. In this scenario, the system dispatch target remains at 800,000 m³/d, while the Shahu pretreatment station is designated as non-dispatchable, with its inflow set to 0 m³/d.

(3) Overflow Risk in Upstream Pipelines Due to Rainfall

Rainfall in upstream areas may cause a surge in combined sewage, raising water levels before the pretreatment station pumps and creating overflow risks in the upstream network. To mitigate this risk, the affected pretreatment station must temporarily increase its inflow and maintain the higher intake until water levels subside. Meanwhile, inflow at other stations should be adjusted to ensure the overall flow remains within the treatment capacity of the downstream wastewater plant. In this scenario, the system dispatch target is set to 800,000 m³/d, with the Erlangmiao pretreatment station designated as non-dispatchable and its inflow fixed at 600,000 m³/d.

3.2. Full-Process Sewage System Calculation and Analysis

3.2.1. Interception Points and Interception Methods

According to the overall layout of the wastewater system in the core area of the Great East Lake Area, the total sewage volume will be approximately 1.5 × 10⁵ m³/day after Wugang’s relocation. The total sewage is divided into three catchment zones: 7 × 10⁴ m³/day flows directly to the Beihu Sewage Treatment Plant via the surface system, 6.5 × 10⁴ m³/day is directed to Luobuzui Pretreatment Station, and 1.5 × 10⁴ m³/day flows into Wudong Pretreatment Station. The distribution maps of the near-term and long-term interception points are shown in Figure 4 and Figure 5.

3.2.2. Node and Pipeline Flow in the Sewage Transmission System

Based on the dry-season flow predictions for each sewage collection system in the East Lake core area and the anticipated combined (or mixed) flow for interception, the capacities of sewage lift stations and key nodes (pretreatment stations) were determined as follows and summarized in

Table 7. Node Flow in the Transmission System.

		Near-Term Capacity (2024)			Long-Term Capacity (2049)
		Average Dry Season Flow	Peak Dry Season Flow	Rainy Season Flow	Dry Season Capacity	Dry Season Capacity
North Line Sewage Deep Tunnel	Shahu Pumping Station	0.77	1.0	1.0	0.77	0.77
	Erlangmiao Node	5.67	7.37	9.8	6.37	8.28
	Luobuzui Node	2.3	3.0	3.0	4.4	5.72
	Wudong Node	0.4	0.52	2.4	0.81	1.05
	North Line Deep Tunnel Pumping Station	9.26	12	12	11.6	15
North of Baiyushan Qinghua Road		0.62	0.81	0.81	2.3	3.0
Longwangzui (South Line)		/	/	/	3.5	4.5
Design Capacity of Beihu Wastewater Treatment Plant		9.26	12	12	11.6	11.6

.

(1) Shahu Pump Station: lift capacity along the D1350 sewer on Shahu Avenue is 1.0 m³/s;

(2) Erlangmiao Pretreatment Station: dry-season capacity includes the combined scales of the Shahu and Erlangmiao sewage collection systems, and near-term rainy-season capacity equals the sum of its dry-season capacity and the rainy-season inflow

(3) Luobuzui Pretreatment Station: dry-season capacity includes the Luobuzui sewage collection system and part of the future inflow from Wugang relocation;

(4) Wudong Pretreatment Station: dry-season capacity covers the Wudong area (south of Qinghua Road) and part of the future Wugang inflow; rainy-season capacity includes dry-season capacity and one additional unit (n0 + 1);

(5) Separately treated inflow: dry-season capacity includes Baiyushan area sewage collection system (north of Qinghua Road) and part of the future Wugang inflow.

In the near term, the North Line Sewage Deep Tunnel exhibits a cumulative average dry season flow of 9.26 m³/s, primarily contributed by Erlangmiao (5.67 m³/s), Luobuzui (2.3 m³/s), and Wudong (0.4 m³/s) nodes. During peak dry and rainy seasons, this total increases to 12 m³/s, aligning with the design capacity of the Beihu Wastewater Treatment Plant, indicating the system operates near full capacity under stressed conditions. The long-term projections show significant increases, especially at Erlangmiao (from 6.37 to 8.28 m³/s) and Luobuzui (from 4.4 to 5.72 m³/s), driving the pumping station’s required capacity up to 15 m³/s. The system’s phased scalability is further demonstrated by the Beihu Plant’s consistent capacity of 12 m³/s, which remains adequate in the near term but may require optimization if actual inflows exceed projections in the long term. Overall, this arrangement ensures balanced hydraulics, treatment efficiency, and system safety under combined-flow conditions.

According to near-term and long-term sewerage prediction and node deployment, the design flow rates for the transmission system pipelines are obtained (

Table 8. Design Flow of Pipeline Sections in the Transmission System.

Operating Condition		Erlangmiao-Third Ring Road	Third Ring Road-Wudong	Wudong-Pumping Station	Luobuzui-Third Ring Road
		Q2 (m3/s)	Q3 (m3/s)	Q4 (m3/s)	Q5 (m3/s)
Near-term	Average Dry Season Flow	5.67	7.97	8.37	2.3
	Peak Dry Season Flow	7.37	10.37	10.89	3.0
	Rainy Season Flow	7.37	10.37	10.89	3.0
Long-term	Average Dry Season Flow	6.37	10.77	11.6	4.4
	Peak Dry Season Flow	8.28	14	15	5.72

). In the near term, average dry-season flows range from 2.3 m³/s (Luobuzui) to 8.37 m³/s (Wudong), with the highest rainy-season peak reaching 10.89 m³/s (Wudong), indicating the system’s main interception capacity during rainfall. Long-term design accounts for urban expansion and climate variability, with increased capacities, e.g., Wudong peak flow up to 15 m³/s and Luobuzui to 5.72 m³/s. Overall, the pretreatment stations distribution, near-and long-term planning ensures balanced hydraulic performance, effective interception, and prevention of overloading the wastewater treatment plant across both dry and rainfall conditions.

3.2.3. Pretreatment Stations

The pretreatment stations are designed for the maximum inflow during the rainy season, while the capacity for conveying sewage into the deep tunnel and the wastewater treatment plant is based on the maximum dry-season flow. Before the deep-tunnel pump stations or the initial rainwater treatment plants are fully operational, the entire captured rainwater flow will not be immediately transferred; however, the pump station capacity of the deep tunnel is sufficient to meet rainy-season transfer requirements. As shown in

Table 9. Wastewater Pretreatment Processes.

No.	Project Name	Pretreatment Facilities and Objectives
1	Hong Kong Deep Sewage Tunnel Pretreatment Station	Coarse Bar Screen: 20 mm Fine Bar Screen: 4 mm Grit Chamber: Removes 95% of grit particles larger than 0.2 mm
2	Domestic Urban Wastewater Treatment Plants	Coarse Bar Screen: 20–25 mm Fine Bar Screen: 6 mm Grit Chamber: Removes 95% of grit particles larger than 0.2 mm

, domestic urban wastewater treatment plants in China generally adopt 20–25 mm coarse screens and 6 mm fine screens, therefore obtaining a grit removal efficiency of around 95%. In contrast, Hong Kong’s deep tunnel pretreatment stations applied high-standard fine screens to minimize the entry of suspended solids and grit into the deep tunnel system and thus prevent deposition [26]. These experiences provide a helpful reference for evaluating and modifying the East Lake system’s pretreatment design.

Based on field surveys of the existing Sha Lake, Erlangmiao, and Luobuzui wastewater treatment plants within the service area of the East Lake core-area sewage transfer system, and considering characteristics such as combined flow areas, mixed sewage connections, and numerous construction sites, a new Erlangmiao pretreatment station was constructed in the southeast corner of the Erlangmiao wastewater treatment plant. The pretreatment stations perform functions that are generally similar to the primary treatment processes of municipal wastewater treatment plants, and therefore adopt largely comparable treatment technologies. The Sha Lake wastewater treatment plant employs a primary treatment process consisting of coarse screens, a lift pump station, fine screens, and a horizontal flow grit chamber. The Erlangmiao plant’s primary treatment includes coarse screens, a lift pump station, fine screens, and a swirl grit chamber, while the Luobuzui plant uses a front coarse screen, coarse and fine screens with a lift pump station, and a swirl grit chamber. This pretreatment station not only serves as the starting point of the main sewage tunnel but also plays a crucial role in “lightening the load” for the underground main tunnel in the future. Through precise filtration and deodorization facility design, the pretreatment station reduces the burden on the main tunnel while minimizing the demand for surface land. These rigorous pretreatment measures ensure that sewage entering the underground main tunnel is free of impurities.

3.2.4. Energy Dissipation and Noise Reduction in Inflow Shafts

Currently, the commonly used inflow shafts in deep tunnel projects include five structural types: vortex, spiral, baffle, drop, and boot-shaped shafts [27,28]. Energy dissipation in spiral shafts is mainly achieved through friction between the flow and contact surfaces as well as water drops, resulting in complex structures. Drop shafts dissipate energy primarily through the free fall of water, featuring simpler structures but poorer structural load conditions, severe air entrainment, and less favorable flow patterns [29]. Boot-shaped shafts build upon drop shafts by adding a large air removal chamber, improving both venting within the shaft and the flow regime at the tunnel entrance [30].

Vortex shafts have an inlet similar to that of a swirl grit chamber, causing water to spiral downward along the shaft walls under centrifugal force; energy dissipation occurs mainly through friction and water drops [31]. Baffle-type shafts feature multiple staggered horizontal baffles, dividing the shaft vertically into a series of drop sections, with energy dissipation achieved through sequential drops. As shown in Figure 6, both vortex and baffle shafts have relatively simple structures. To scientifically select the inflow shaft type, a three-dimensional model of the shaft was established using the realizable k-ε two-phase gas–liquid turbulence model. The governing equations were discretized using the control volume method, and velocity–pressure coupling was handled via the SIMPLER algorithm. A flow boundary was applied at the upstream open channel entrance, and a pressure boundary was applied at the outlet. Accordingly, three-dimensional models of vortex-type (Figure 7) and baffle-type inflow shafts (Figure 8) were established, and then the distribution of water phase volume fraction inside the tunnel was, respectively, simulated.

As shown in Figure 7b,c, the vortex shaft exhibits stable flow conditions and uniform velocity distribution, with a clear separation between water and air phases. A relatively stable wall-adhering spiral flow forms inside the shaft, resulting in favorable wall stress conditions. The central void ratio of the shaft is relatively high, and no air enters the deep tunnel at the design water level.

In contrast, the Baffle shaft shows unstable flow and uneven velocity distribution (Figure 8b,c). Although the air release performance is acceptable, there is no distinct water phase. The pressure distribution across the baffle layers is uneven, with higher pressures in the water impact zones and relatively lower pressures elsewhere. At the design water level, small amounts of air can accumulate at the deep tunnel entrance (Figure 8a).

Both shaft schemes meet the energy dissipation requirements of the shaft under design conditions. Considering domestic and international experience as well as the 3D model results, both vortex-type and Baffle shafts can satisfy land-use constraints and maintenance requirements. However, the Baffle shaft has poorer air venting, and long-term operation requires higher standards for structural stress uniformity, durability, and corrosion resistance. Additionally, the upper water level of the Baffle shaft is less stable, which is unfavorable for operational monitoring. Therefore, the vortex-type shaft is selected for this project.

4. Discussion

4.1. Comparative Analysis with the International Deep Tunnel Project

Based on the technical review presented above and a horizontal comparison with international deep tunnel practices (see

Table 10. Deep Tunnel Station: International Comparison and Wuhan East Lake Tunnel Drainage Project.

City/Project	Pretreatment/Preliminary Treatment	Shaft Design and Energy Dissipation	Intelligent Operation and Control
Hong Kong-HATS Stage 2A	Multi-stage bar screens and grit removal; grease separation; air venting for odor control; effluent delivered via deep tunnel to secondary treatment plant	Vortex-type drop shaft with air conduits and energy dissipation pool; vent shafts to prevent negative pressure [32]	Real-time monitoring and adaptive SCADA control, enabling multi-pump coordination and automated gas venting
Singapore-DTSS Phase 2 (Changi WRP)	Fully enclosed pretreatment units with high-efficiency swirl grit and grease removal; pre-membrane equalization tanks to buffer flow	Multi-stage swirl shaft with buffer chamber, ~90 m depth; optimized for energy dissipation and gas separation	AI-based predictive flow control; centralized operation platform coordinates all pumping stations for unattended operation [33]
London-Thames Tideway Tunnel	Multiple interception and storage shafts integrated with pretreatment; vortex and grit control measures at entry [34]	Vertical vortex-type shafts with energy dissipation chamber; vent shafts for air release; 60–70 m depth	Digital twin system for real-time hydraulic simulation, predictive maintenance, and AI-assisted emergency response
Chicago-TARP (Tunnel and Reservoir Plan)	Initial grit and storage function; captures combined sewer overflows for pumping to treatment plants	Baffle-type drop shafts; up to 90 m depth; retrofitted vortex sections to mitigate cavitation [10]	Real-time rainfall forecasting and pump station control; limited AI application
Tokyo-Edogawa Trunk Sewer	Rainwater and wastewater separation with grit removal; automated sediment cleaning; odor control [35]	Spiral vortex drop shaft with vent shaft; optimized for gas–liquid separation and wall erosion mitigation	Adaptive pump control with AI, IoT-based sediment prediction, and shaft monitoring
Wuhan-East Lake Deep Tunnel	Integrated pretreatment and online sediment removal, coupling pretreatment facilities with real-time cleaning to enhance operational efficiency	Multi-point coordinated shaft layout, optimizing shaft positioning and pump station coordination for improved emergency response	Intelligent operation and control system: real-time flow and pump monitoring, automated sediment cleaning, adaptive pump operation, and optimized gas venting

), the Wuhan East Lake Deep Tunnel has achieved significant breakthroughs in core technologies. First, the project integrates pretreatment with online sediment removal, overcoming the traditional separation of these processes and enabling real-time management of solids within the deep tunnel, thereby substantially improving operational efficiency and maintenance convenience. Second, the system employs a multi-point coordinated drop shaft layout with pump station collaboration, optimizing the spatial and functional configuration of shafts and pumping stations, and markedly enhancing overall hydraulic efficiency and emergency response capability. Finally, by incorporating AI-assisted optimization of pump start–stop strategies, flow regulation, and gas venting, the system achieves further improvements in autonomous operation through real-time monitoring and adaptive control. Together, these integrated technologies enable the Wuhan East Lake Deep Tunnel to surpass international benchmark projects in sediment management, shaft operation, and intelligent control, providing a replicable model for smart and resilient urban drainage systems in China.

4.2. Assessment of System Benefits

The construction of the Dazhonghu Core Area Deep Sewer Transmission System systematically addresses regional challenges related to sewage treatment, water environment protection, and flood control, while substantially improving the investment climate and ecological environment of the central urban area. By enhancing the region’s infrastructure and overall environmental quality, the project has strengthened the area’s business development capacity, attracting more commercial investments and institutions, and promoting the clustered growth of high-end service industries represented by the financial sector. These developments have produced remarkable social and economic benefits.

Meanwhile, projects such as the Dazhonghu Core Area Sewage Transmission System, the Beihu Wastewater Treatment Plant and its auxiliary facilities, together with planned stormwater deep tunnels, contribute to a comprehensive solution for the area’s water environment management. Through the relocation of urban core sewage treatment plants to peripheral zones for centralized treatment, these initiatives resolve the conflicts between limited treatment capacity and stricter effluent discharge requirements. This approach effectively safeguards the central lakes and canal systems, ensuring the achievement of water quality targets. By integrating sewage transmission, overflow control during rainy seasons, and stormwater drainage, the system takes a crucial step toward establishing an ecological water network for the Dazhonghu region, yielding significant environmental benefits.

The East Lake Deep Tunnel project also generates both direct and indirect economic returns. The relocation of the Shahu Wastewater Treatment Plant has released high-value urban land previously occupied by three central sewage facilities, with the land value freed by the Shahu site alone exceeding the total project investment of 3 billion CNY. Beyond these direct gains, the project has greatly enhanced the region’s urban development potential and ecological carrying capacity, as the deep tunnel system now serves as the backbone of a regional environmental governance framework. This transformation positions the area as a model for high-end financial services and integrated commercial–residential development in Wuhan.

With the continuous process of urbanization and growing public demand for improved environmental quality, upgrading central urban sewage treatment facilities has become increasingly urgent. The “deep tunnel transmission and centralized treatment” model, exemplified by this project, effectively releases scarce urban land resources, supports higher treatment standards, and preserves shallow underground space for future urban utilities. This study demonstrates the value of China’s first deep sewage transmission tunnel and offers a meaningful reference for future deep tunnel projects, providing a comprehensive response to the water environment challenges of the East Lake area.

4.3. Future Research Directions

Future research will prioritize the application of artificial intelligence (AI) to achieve real-time, intelligent operation and management of deep tunnel drainage systems. AI-based algorithms can integrate multi-source data from rainfall forecasts, flow monitoring, and energy use to enable self-learning and adaptive optimization of pumping, gate control, and sediment management. The development of AI-driven digital twin platforms will further enhance situational awareness, predictive maintenance, and emergency response through dynamic simulation of flow and system stress. These advances will lay the foundation for autonomous and predictive operational control, ensuring that deep tunnel systems can respond efficiently to fluctuating inflows and extreme weather events.

Furthermore, the continued evolution of AI technologies will create new opportunities for holistic and resilient drainage management. Integrating deep learning, reinforcement learning, and optimization frameworks into tunnel operation could enable long-term performance forecasting, multi-objective decision-making, and adaptive resilience assessment. Moreover, cross-disciplinary research combining hydraulics, informatics, and urban climate science will expand the theoretical frontiers of smart water engineering. Through these efforts, the East Lake Deep Tunnel can serve as a prototype for next-generation intelligent drainage infrastructure, offering valuable insights for sustainable urban development in the era of climate uncertainty.

5. Conclusions

This study established an integrated framework for the planning, hydraulic design, simulation analysis, and intelligent management of the East Lake Deep Tunnel drainage system. A pilot-scale pipeline experiment accurately determined the critical non-silting velocity under full-pipe flow, providing a solid empirical basis for hydraulic design. Hydrological prediction and flow analysis were applied to optimize the capacities of pumping stations and pretreatment facilities for efficient conveyance and storage.

A three-dimensional gas–liquid two-phase model was used to evaluate inflow shafts at Erlangmiao, Luobuzi, and Wudong stations. The vortex-type shaft demonstrated more stable hydraulic performance, uniform velocity distribution, and favorable wall stress compared with the baffle-type shaft, confirming the model’s accuracy and design reliability.

The implementation of smart operational control—featuring automated gas venting, sediment detection, and real-time flow adjustment—significantly enhanced system efficiency, safety, and sustainability. Overall, the East Lake Deep Tunnel represents a technically advanced, data-driven approach to urban wastewater management and provides a practical reference for similar projects in high-density cities.