Source-linked AI summary
Non-Invasive Inspection of Water Canals Using Dronar
Michael Zielinski, Zhizhan Wang, Benjamin Dymond, Reza Razavian, Zhongwang Dou
TL;DR
Canal defects are difficult to inspect without costly, infrequent dry-ups, limiting opportunities to target maintenance. This paper uses an unmanned surface vehicle carrying consumer-grade sonar to inspect canal beds without draining the water. Sedimented-segment depth profiles were up to 15 cm shallower than profiles after cleaning, and repeated clean-segment runs closely overlapped.
Problem
Canal sediment, cracks, and lining deformation are difficult to identify without draining the water, while dry-ups are costly and occur only once every four years per segment.
Source: INTRODUCTION AND BACKGROUND · pp. 1–2Method
The paper presents a non-invasive inspection system that integrates consumer-grade sonar with an unmanned surface vehicle to collect canal-bed data.
Source: INTRODUCTION AND BACKGROUND · p. 3, Integrated unmanned surface vehicle… · p. 3Results
Up to 15 cm offsets separated the consistently shallower sedimented-segment DownScan profiles from profiles after cleaning, while repeated clean-segment runs closely overlapped.
Source: CONCLUSIONS · p. 14Takeaways & Limitations
The field tests support the dronar as a viable proof-of-concept tool for non-invasive canal-bed inspection.
Source: CONCLUSIONS · p. 14Validation compared only sedimented and clean canal conditions, and the system identifies only approximate sediment-deposit size and depth.
Source: Limitations and Future Work · p. 14
Abstract
from arXiv · showhide
Open concrete canals play a vital role in water transportation, serving as primary water infrastructure for millions of people across the Phoenix, Arizona, metro area. Over time, the concrete canals can experience a range of issues, including canal lining deformation, cracked concrete, and sediment buildup on the canal floor. Identifying such critical issues is a resource-intensive process, which currently happens only during four-year dry-up cycles. This prevents the maintenance crew from prioritizing operations on the most affected canal segments. To address this issue, the research team has developed and verified an easily deployable and non-invasive method to inspect canal beds without draining the water. This inspection system integrates affordable, off-the-shelf drone and sonar technology (termed dronar). This dronar system includes a consumer-grade sonar system integrated into an unmanned surface vehicle (USV) that carries the sonar transducer just under the surface of the canal water. This paper presents a proof-of-concept demonstration of the dronar system across three field tests on the Arizona Canal in Phoenix. DownScan depth profiles from the sedimented canal segment were consistently shallower than profiles from the same segment after cleaning, with offsets of up to 15 cm observed along the track. Repeated runs over the clean segment produced closely overlapping DownScan depth profiles, confirming that the dronar yields repeatable measurements across the natural variation of the canal bed. These results establish the dronar as a viable proof-of-concept tool for non-invasive canal bed inspection.
INTRODUCTION AND BACKGROUND
SRP canals serve Phoenix homes and agriculture, but sediment, cracks, and lining deformation are difficult to inspect without costly, infrequent dry-ups. The paper introduces dronar, a USV-based sonar approach intended to inspect canal beds without draining them.
INTRODUCTION AND BACKGROUND
- SRP’s seven canals span over 211 km and serve nearly 2.5 million Phoenix homes and the region’s agriculture, while sediment, cracks, and lining deformation can impair water transport.Visual inspection identifies only defects above water; sediment can cover cracks.Source: INTRODUCTION AND BACKGROUND · p. 1
- Dry-ups are costly and each canal segment is maintained through them only once every four years, motivating non-invasive bed inspection to help target maintenance.The stated targets include sediment buildup and potential liner deformation or concrete cracks.Source: INTRODUCTION AND BACKGROUND · p. 2
- Aerial sonar-tethered drones pose airspace, transducer-positioning, and endurance challenges for canal surveys, particularly near Phoenix-area airports.The Arizona Canal passes near multiple commercial airports, making airspace approval complex.Source: INTRODUCTION AND BACKGROUND · p. 3
- The paper presents a sonar-equipped unmanned surface vehicle and data verifying its operation and utility for non-invasive concrete-canal inspection.Prior work on consumer-grade sonar and drone-tethered systems motivates the approach.Source: INTRODUCTION AND BACKGROUND · pp. 2–3
MATERIALS AND METHODS
The sonar setup combines a georeferenced Lowrance transducer and hub with DownScan and SideScan coverage suited to the canal dimensions.
Sonar transducer
- The 800 kHz transducer provides DownScan coverage up to ±30° and SideScan coverage up to ±58° for canals approximately 15–20 m wide and 1–2 m deep.The transducer is paired with an HDS Pro 9 sonar hub and onboard GPS.Source: Sonar transducer · p. 3
- Simultaneous DownScan and SideScan can cover most of the canal cross section in a single lengthwise centerline pass.The passage describes this coverage for the specified canal dimensions.Source: Sonar transducer · p. 3
- The sonar automatically stores data on an internal SD card, while onboard GPS enables georeferenced returns for post-processing and mapping.The passage also notes the system’s low power draw and compact form factor as supporting lightweight-platform integration.Source: Sonar transducer · p. 3
Integrated unmanned surface vehicle and sonar
The dronar pairs a custom-built, remotely controlled twin-pontoon USV with an integrated sonar payload and locally stored survey data.
Integrated unmanned surface vehicle and sonar
- The custom-built twin-pontoon USV uses differential thrust from two electric thrusters and a CubePilot controller running ArduPilot rover firmware.The frame combines kayak stabilizer pontoons and aluminum components.Source: Integrated unmanned surface vehicle… · p. 3, Figure 2 · p. 4
- A 4-cell 14.8 V 20 Ah battery powers the approximately 16 kg vehicle, which uses modular commercial components and 3D-printed brackets for reconfiguration.Source: Integrated unmanned surface vehicle… · p. 4
- The sonar transducer is suspended below the waterline and beneath the thrusters, reducing turbulence and improving survey image clarity.Source: Integrated unmanned surface vehicle… · p. 4
- The untethered system stores sonar and controller logs locally; in Auto mode, the USV follows a predefined path even after radio-signal loss.Manual piloting was available up to 1 km, and loss of connection triggered Loiter mode with GPS position holding.Source: Integrated unmanned surface vehicle… · p. 4
Dronar transportation and deployment
A cart was designed to protect and handle the dronar during launch and recovery on canal walls. The first field test assessed deployment and operation on a recently cleaned Arizona Canal segment.
Dronar transportation and deployment
- The cart cradles the dronar with its transducer suspended until water entry and enables rope-guided launch and recovery down 45°–60° canal walls.Holes in the cart base reduce buoyancy and facilitate descent and ascent.Source: Dronar transportation and deployment · p. 5, Figure 3 · p. 5
Test 1: dronar operation verification
- Test 1 covered a 1.6 km Arizona Canal segment recently cleaned and inspected for liner damage, with limited to no sediment or wall/bed deformation.The test assessed vehicle performance in moving water and the cart-based launch and recovery method.Source: Test 1: dronar operation verification · p. 5, Figure 4 · p. 6
- During Test 1, the USV traveled downstream in Manual mode, then self-piloted upstream briefly in Auto mode while sonar data and location were logged.Recovery involved maneuvering the USV upstream of the cart and allowing it to drift onto the cart at the canal wall.Source: Test 1: dronar operation verification · p. 6
Test 2: pre- and post-maintenance surveys of a sedimented canal
The Arizona Canal segment west of North Mesa Drive was selected because storms had left it heavily sedimented and it was scheduled for maintenance. Pre- and post-maintenance surveys covered the same segment, with six post-maintenance runs spanning three operating conditions.
Test 2: pre- and post-maintenance surveys of a sedimented canal
- The segment west of North Mesa Drive was chosen after storms caused heavy sedimentation and before scheduled maintenance.The site provided pre- and post-maintenance data for evaluating the sonar inspection technique.Source: Test 2: pre- and post-maintenance… · p. 6
- The November 7, 2025 pre-maintenance survey used Manual mode downstream and logged sonar and location data for post-processing.The survey used the same launch and recovery procedure, with a gaff pole added to guide the USV onto the cart.Source: Test 2: pre- and post-maintenance… · p. 6
- Six downstream-only post-maintenance runs covered the same segment, with two runs each at full speed, canal-current speed, and in Auto mode.The survey followed thorough cleaning and canal-liner repair.Source: Test 2: pre- and post-maintenance… · p. 7
RESULTS
The field tests assessed dronar operation over a 1.6 km canal segment and across repeat surveys of a shorter segment. Runs were completed in manual and autonomous modes without telemetry loss.
Speed, endurance, and test duration
- The USV completed the full 1.6 km Test 1 downstream run in approximately 25 minutes at 8 km/hr at full throttle, using 38% battery.It traveled upstream at 1.6 km/hr at full throttle; loiter mode held position within 1 m of its assigned GPS coordinate.Source: Speed, endurance, and test duration · p. 7
- There was no loss of telemetry during the 1.6 km Test 1 run or any of the Test 2 runs.The USV traveled upstream in Manual mode at the end of Test 1.Source: Speed, endurance, and test duration · pp. 7–8, Figure 6 · p. 8
- The pre-maintenance Test 2a survey consisted of one approximately 400 m downstream run that took approximately 10 minutes.The USV was manually controlled while sonar and GPS data were logged onboard for post-processing.Source: Speed, endurance, and test duration · p. 8
- Test 2b included six downstream runs: two each at full throttle, canal-current speed, and in Auto mode.The first four runs ended with 71% charge remaining on a 16 Ah battery; the final two ended with 75% on a 20 Ah battery.Source: Speed, endurance, and test duration · p. 8
Deployment and recovery
DownScan profiles distinguished the sedimented canal bed from the same segment after maintenance, while repeated clean-segment runs produced closely matching profiles. Field tests also documented recovery procedures and their improving timing with operator experience.
Deployment and recovery
- Test 1 recovery took four attempts because of cart drift and timing complexity, but the USV was secured without damage.Figure 7 shows the staged recovery, including the cart hook holding the USV secure.Source: Deployment and recovery · p. 8, Figure 7 · p. 9
- During Test 2, recovery time fell from approximately 4 minutes on the first recovery to approximately 1 minute by the final run.Six sequential downstream runs involved recovery and redeployment between runs; Figure 8 shows recovery with a gaff pole.Source: Deployment and recovery · p. 9, Figure 8 · p. 10
Sonar Data
- Depth-profile analyses used DownScan data exclusively; SideScan imagery was included for illustration rather than quantitative analysis.Figure 9 presents SideScan as a top-down view and DownScan as a side view of the canal.Source: Sonar Data · p. 10, Figure 9 · p. 11
- Repeated clean-segment DownScan profiles stayed within 7 cm of one another, against an approximately 12 cm total depth range across the canal bed.Figure 10 compares two canal-current-speed runs; apparent shifts at 200–300 m were attributed to distance-calculation inaccuracy rather than sonar inaccuracy.Source: Sonar Data · p. 11, Figure 10 · p. 12
- Pre-maintenance raw depths were approximately 15 cm shallower than post-maintenance depths, and filtered residuals reached 15 cm while the post-maintenance trace stayed within approximately 5 cm of zero.Figure 11 shows the cropped representative segment contrast; filtering suppresses trends longer than 50 m and noise shorter than 0.05 m.Source: Sonar Data · p. 12, Figure 11 · p. 13
Operational Observations
Field tests found that recovery reliability improved with operator familiarity and a gaff pole, while hot-weather operation produced a controller warning. Upstream travel was severely constrained, but downstream-only surveys were practical; multi-kilometer endurance was sufficient, with larger-scale deployment requiring battery planning.
Recovery procedure
- Recovery reliability improved across Test 2, with both operator familiarity and adding a gaff pole contributing.The evidence does not separate the effects of these two changes.Source: Recovery procedure · p. 13
Thermal management
- During the hot first test, the controller overheated and warned; sustained high ambient temperatures may require thermal mitigation.No controller overheating occurred during the cooler November and February tests.Source: Thermal management · p. 13
Speed and endurance
- 1.6 km/hr upstream versus 8 km/hr downstream in Test 1; Test 2 therefore used an effective downstream-only approach.Downstream-only operation also conserved battery.Source: Speed and endurance · p. 13
- Endurance supported multi-kilometer downstream surveys, but scaling to SRP’s 211 km canal system will require battery management.Auto mode appeared more power-intensive per run than Manual mode.Source: Speed and endurance · p. 13
Sonar Performance and Canal Condition Assessment
DownScan surveys distinguished sedimented from cleaned canal-bed conditions and produced repeatable depth profiles. The validation, manual processing, and sonar coverage leave important limits on the system’s current detection scope and workflow.
Sonar Performance and Canal Condition Assessment
- Sedimented-segment DownScan profiles were shallower than post-cleaning profiles, and repeated clean-segment runs overlapped closely.The results support sediment detection and repeatable measurements under the tested canal conditions.Source: Sonar Performance and Canal Condition… · p. 14
Limitations and Future Work
- Validation compared only known sedimented and clean conditions, and the system estimates only approximate sediment size and depth.Further validation is needed for liner deformation, cracked concrete, other sediment types, and foreign objects.Source: Limitations and Future Work · p. 14
- Sonar post-processing was performed manually with Python scripts; automated or semi-automated processing remains future work.The proposed future direction includes machine-learning classification of canal-bed conditions.Source: Limitations and Future Work · p. 14
- The analysis used DownScan alone, whose shallow-water footprint leaves canal margins and wall-adjacent areas unreachable.SideScan data were collected but not analyzed; SideScan analysis or multi-pass surveys are proposed for broader lateral coverage.Source: Limitations and Future Work · p. 14
CONCLUSIONS
The dronar was tested for non-invasive canal-bed inspection on the Arizona Canal. Its DownScan profiles distinguished sediment accumulation from a cleaned segment and showed repeatability, supporting proof-of-concept use in active canals.
CONCLUSIONS
- Across Arizona Canal field tests, sedimented-segment profiles were up to 15 cm shallower than profiles after cleaning.Repeated runs over the clean segment produced closely overlapping depth profiles.Source: CONCLUSIONS · p. 14
- The off-the-shelf USV system was reliably deployed and recovered, and downstream-only operation supported multi-kilometer surveys with modest battery consumption.The system operated in active canals without draining the water for inspection.Source: CONCLUSIONS · p. 14
- Affordable, modular hardware and operation in active canals position the dronar for routine inspection across a large concrete canal network.The findings establish a proof of concept for detecting sediment accumulation and producing repeatable measurements under operational canal conditions.Source: CONCLUSIONS · p. 14