Quay crane trolley track monitoring system

Abstract

The invention discloses a quay crane trolley track monitoring system, and relates to the technical field of quay crane trolley tracks, the system comprises a visual monitoring module, an offset monitoring module, a trolley control module and a main control module; the visual monitoring module is used for monitoring a body structure fault, a connection fault and a component fault of the trolley track; the offset monitoring module is used for monitoring an offset fault of the trolley track; the trolley control module is used for controlling operation of the trolley; the master control module obtains video data of the visual monitoring module and offset parameters of the offset monitoring module through different interfaces, is connected to the trolley control module and is used for monitoring the state of the trolley track through the video data and the offset parameters. And under the condition that the trolley track is determined to be in the fault state, the trolley is controlled through the trolley control module, and fault reminding is triggered. The method has the advantages that the efficiency and accuracy of trolley track monitoring are improved, and safe operation of a trolley system is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of shore crane trolley track technology, in particular to a shore crane trolley track monitoring system. BACKGROUND

[0002] The small track trolley is a key component in the shore crane trolley mechanism, and plays a crucial role in the normal operation of the system. If the shore crane trolley cannot travel parallel to the track, it will cause the trolley structure to bear excessive internal stress, causing deformation of the steel structure, causing irreversible damage to the shore crane. Especially when the trolley is running at high speed under heavy load, the impact force will directly act on the track tread, causing the track and the rail beam to crack, the track and the wheel to abnormally wear, and the trolley frame to crack. Therefore, ensuring the flatness and integrity of the trolley track and its gasket is a necessary condition for the smooth operation of the trolley.

[0003] In the related art, the maintenance and monitoring of the shore crane trolley track mainly rely on strengthening daily manual inspection and improving maintenance technology. This method lacks systematicness and comprehensiveness, and cannot effectively monitor and analyze the damage of the trolley track. In addition, manual inspection is time-consuming and labor-intensive, has inspection blind spots, and has high risk of high-altitude work. SUMMARY

[0004] The present application aims to at least solve one of the technical problems existing in the prior art, and provides a shore crane trolley track monitoring system, which aims to improve the efficiency and accuracy of trolley track monitoring and ensure the safe operation of the trolley system.

[0005] The shore crane trolley track monitoring system provided by the present application comprises a visual monitoring module, an offset monitoring module, a trolley control module and a main control module. The visual monitoring module is used to monitor the body structure failure, connection failure and component failure of the trolley track. The offset monitoring module is used to monitor the offset failure of the trolley track. The trolley control module is used to control the operation of the trolley. The main control module acquires video data of the visual monitoring module and offset parameters of the offset monitoring module through different interfaces respectively, and is connected to the trolley control module. The main control module is used to monitor the state of the trolley track through the video data and the offset parameters, and to control the trolley through the trolley control module and trigger a failure reminder when the trolley track is determined to be in a failure state.

[0006] According to the technical scheme of the embodiment of the present application, the following beneficial effects are achieved: the visual monitoring module and the offset monitoring module are arranged at the trolley track, so that the trolley track can be automatically and comprehensively monitored without manual inspection, and the monitoring efficiency and accuracy are greatly improved; and the main control module can automatically analyze the state of the trolley track according to the video data of the visual monitoring module and the offset parameter of the offset monitoring module, and control the trolley through the trolley control module when the trolley track is determined to be in a fault state, so as to ensure the safe operation of the trolley system.

[0007] According to some embodiments of the present application, the system further comprises a monitoring management module, the main control module comprises a first interface and an analysis unit, the analysis unit comprises a second interface, the analysis unit is configured to acquire the video data through the second interface and perform fault analysis to obtain visual analysis results, and the main control module is further configured to acquire the video data through the first interface and upload the video data to the monitoring management module after adding the visual analysis results in the video data.

[0008] According to some embodiments of the present application, the triggering of the fault reminder comprises sending a fault reminder to the monitoring management module, and the fault reminder comprises a fault type and corresponding fault data.

[0009] According to some embodiments of the present application, the visual monitoring module is further configured to perform preliminary fault analysis and event detection to obtain preliminary visual analysis results, and the analysis unit is further configured to perform fault analysis in combination with the preliminary visual analysis results and the video data to obtain visual analysis results.

[0010] According to some embodiments of the present application, the system further comprises a monitoring gateway module, the monitoring gateway module is connected to the offset monitoring module and the main control module respectively, and the monitoring gateway module is configured to acquire the offset parameter of the offset monitoring module and send the offset parameter to the main control module.

[0011] According to some embodiments of the present application, the system further comprises a POE switch module, the POE switch module is connected to the visual monitoring module and the main control module respectively, and the POE switch module is configured to supply power for the visual monitoring module, acquire the video data of the visual monitoring module, and send the video data to the main control module.

[0012] According to some embodiments of the present application, the body structure fault comprises a crack fault and a fracture fault, the connection fault comprises a connection error fault of a track joint part, a joint gap error fault, and a centering error fault, and the component fault comprises a pressing plate missing fault and a bolt missing fault.

[0013] According to some embodiments of the present application, the master module is further configured to control the trolley to decelerate or stop according to a fault type of the trolley track in a case where it is determined that the trolley track is in a fault state.

[0014] According to some embodiments of the present application, the master module further comprises a fault feature database configured to collect the video data and the offset parameter monitoring, and to classify a fault feature state of the track and to regularly judge a fault development trend, so as to realize fault early warning.

[0015] According to some embodiments of the present application, the fault feature database is further configured to generate a data continuous tracking record of a fault formation start, development and maintenance quality of the trolley track.

[0016] Other features and advantages of the present application will be set forth in the description that follows, and in part will be apparent from the description, or can be learned by practice of the present application. The purposes and other advantages of the present application will be realized and attained by the structure particularly pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings are included to provide a further understanding of the technical scheme of the present application, and constitute a part of the specification, and are used to explain the technical scheme of the present application together with the embodiments of the present application, and do not constitute a limitation to the technical scheme of the present application.

[0018] The present application will be further described below in conjunction with the drawings and embodiments; Figure 1 is a structural block diagram of a trolley track monitoring system of a shore crane provided by an embodiment of the present application; Figure 2 is a structural diagram of a trolley track monitoring system of a shore crane provided by an embodiment of the present application. DETAILED DESCRIPTION

[0019] This part will describe the specific embodiments of the present application in detail, and the preferred embodiments of the present application are shown in the drawings, and the drawings are used to supplement the description of the textual part of the specification, so that one can intuitively and visually understand each technical feature and the overall technical scheme of the present application, but it cannot be understood as a limitation to the protection scope of the present application.

[0020] In the description of the present application, it should be understood that, in relation to the orientation description, for example, the orientation or position relationship indicated by up, down, front, back, left, right, etc. is based on the orientation or position relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the indicated device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation to the present application.

[0021] In the description of this application, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0022] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0023] The various embodiments of the quay crane trolley track monitoring system method of this application will be further described below with reference to the accompanying drawings.

[0024] like Figure 1 As shown, Figure 1 This is a structural block diagram of a quay crane trolley track monitoring system provided in one embodiment of this application. The quay crane trolley track monitoring system method may include a visual monitoring module, an offset monitoring module, a trolley control module, and a main control module. The visual monitoring module is used to monitor structural faults, connection faults, and component faults of the trolley track. The offset monitoring module is used to monitor offset faults in the trolley track; The vehicle control module is used to control the operation of the vehicle; The main control module acquires video data from the visual monitoring module and offset parameters from the offset monitoring module through different interfaces, and connects to the trolley control module. The main control module monitors the status of the trolley track through video data and offset parameters, and controls the trolley through the trolley control module and triggers fault alerts when it determines that the trolley track is in a faulty state.

[0025] For example, the visual monitoring module is a comprehensive perception and diagnostic system that processes image information. It includes dustproof, waterproof, and shockproof industrial-grade high-definition cameras deployed at key points along the track, as well as an auxiliary lighting system that provides illumination for the cameras to ensure high-quality images can be acquired under any lighting conditions. The visual monitoring module can transmit the acquired high-definition image data in real time to a server or edge computing node located in the central control room via industrial Ethernet or fiber optic network. In addition, through intelligent analysis software on the server, digital image processing algorithms and deep learning models are integrated to form a multi-task concurrent analysis engine. This engine can simultaneously perform multi-dimensional analysis on the incoming image data, such as crack identification, deformation measurement, corrosion assessment, bolt condition inspection, and weld defect detection. The analysis results are then bound to the original image evidence to generate a structured monitoring report, which is stored in the database. An alarm is automatically triggered when the threshold is exceeded.

[0026] For example, the offset monitoring module is used to monitor the dynamic deviation between the actual position of the track and the ideal design position during the operation of the trolley. Therefore, the offset monitoring module can employ high-precision geometric measurement sensors, and may include a measurement system based on a total station or laser tracker. Measurement stations are set up at reference positions such as the start and end points of the track. By emitting laser beams and receiving reflected signals, the three-dimensional coordinates of specific target points on the track can be accurately measured. Through periodic automated scanning, the straightness, elevation, and centerline deviation of the track in the global coordinate system can be calculated. Alternatively, it can include a distributed, fixedly installed displacement sensor network, for example, installing a large number of laser ranging sensors in sections sensitive to elevation changes. The system is vertically aligned with a fixed reference beam or the opposite track, and measures the height change of the track's top surface relative to the reference in real time to construct the track's real-time deflection curve. Laser rangefinders installed on the track's side can measure the track's lateral displacement relative to the reference. It can also include installing an inertial measurement unit and odometer on the trolley body, combined with high-precision GPS, and using data fusion algorithms such as Kalman filtering to calculate the trolley's three-dimensional position and attitude in real time. Since the trolley wheels run close to the track, the trolley's trajectory directly reflects the actual track alignment. By comparing the real-time trajectory with the designed trajectory, the longitudinal slope deviation, horizontal curvature deviation, and track gauge change can be directly obtained.

[0027] Understandably, structural failure refers to problems with the base material of the track beam itself, which may cause damage under long-term heavy loads, impacts, and natural environments. This includes fatigue cracks, plastic deformation, and corrosion. Fatigue cracks are microscopic cracks that propagate under cyclic loading until they form macroscopic visible cracks, such as at weld edges, around holes, or at abrupt changes in cross-section. Plastic deformation refers to permanent bending, crushing, or denting of the track beam, either locally or as a whole, under overload or accidental impact. This directly changes the straightness and elevation of the track, affecting the smooth operation of the trolley. Corrosion is caused by the electrochemical corrosion of the steel structure surface in a marine high-salt and high-humidity environment. This leads to a gradual thinning of the effective cross-section of the components, a decrease in load-bearing capacity, and in severe cases, the formation of rust pits, which exacerbates the initiation of fatigue cracks.

[0028] Connection failures refer to the failures at the connection points that fix the various track sections together and the track to the supporting structure. These include bolt connection failures, such as bolt loosening, falling off, shear fracture, and slippage between the connected plates. Loose bolts can cause the loss of preload in the connection pair, leading to relative displacement between components and compromising the integrity of the structure. Weld connection failures also include weld failures, such as cracking, slag inclusion, and lack of fusion in the welds between the track butt joint, the track and the pad plate, and the track and the web plate. These defects may expand under load, or the weld heat-affected zone material may become brittle, leading to fracture.

[0029] The components in the context of component failure refer to relatively independent standardized parts in the track system, rather than the track beam itself. These include track clamps and track pads, as well as safety limiting devices such as stops and guards. Track clamps press the bottom of the track onto the supporting beam to prevent lateral movement and upward tilting. Breakage, severe wear, or failure of the clamps themselves can lead to track instability. Track pads are installed between the track and the supporting beam, serving to level, buffer, and insulate. Cracks, crushing, or aging of the pads can cause changes in the layout height and additional dynamic impacts. Deformation or damage to safety limiting devices such as stops and guards can cause the trolley to derail at the end of its journey, leading to a serious accident.

[0030] For example, in monitoring fatigue cracks in structural failures, the vision system uses a high-resolution, high-contrast industrial camera, along with a suitable lighting system to eliminate shadows and reflections. It uses edge detection and texture analysis algorithms to identify discontinuous, elongated, and often irregularly oriented linear features in the image. After training on a large number of crack images, the semantic segmentation model in deep learning can accurately segment the crack region from a complex background at the pixel level, even detecting very subtle early cracks. Furthermore, the system can not only identify the existence of cracks but also accurately calculate the crack's length, width, area, and other geometric parameters, and compare them with preset safety thresholds to provide early warnings.

[0031] For monitoring plastic deformation, three-dimensional vision technology can be used, such as stereo vision-based binocular or multi-view camera systems, or structured light / laser scanning technology. By reconstructing the three-dimensional point cloud model of the track and accurately registering and comparing it with the standard CAD model during design, the amount of deformation of the track in the vertical and horizontal directions, such as bending, twisting or local concavity, can be clearly quantified, and the structural deformation can be displayed intuitively.

[0032] For corrosion monitoring, vision systems can analyze color and texture features. Through color space transformation and region segmentation, rusted areas can be effectively separated from the background, and then the proportion of the rusted area to the entire analysis area can be calculated, i.e., the corrosion rate, as a quantitative indicator to assess the severity of corrosion.

[0033] For monitoring bolt loosening or detachment in connection faults, cameras at fixed angles can be deployed at each key connection point. The standard shape of the bolt and nut can be located using a template matching algorithm, and the status can be determined by analyzing their image features. For loosening, a visible gap can be detected between the nut and the surface of the connected part, or the rotational displacement can be indirectly determined by measuring the angular deflection of the nut's edge relative to the initial reference image. For detachment, the bolt can be identified as detached when its contour features cannot be found through pattern recognition within a preset bolt installation area. For weld crack monitoring, a deep learning model trained on the weld area can learn the essential difference between normal weld texture and abnormal crack features, thereby identifying transverse, longitudinal, or fusion-line-promoted cracks in complex weld backgrounds.

[0034] For example, the main control module can perform deep fusion and correlation analysis on the video data provided by the visual monitoring module and the offset parameters provided by the offset monitoring module, thereby forming a comprehensive, dynamic, and accurate assessment of the health status of the vehicle track. The main control module will establish a hierarchical status assessment model for the track, including two dimensions: global geometric status and local structural status. Video data and offset parameters are the main data inputs for these two dimensions. In the initial stage of data integration, the data preprocessing unit within the main control module will perform spatiotemporal alignment of the two types of data to ensure that the track segment in the video being analyzed matches the track mileage position corresponding to the offset parameters, while all data are stamped with a unified timestamp. Subsequently, when the offset monitoring module detects a continuously increasing lateral offset or settlement in a certain section of the track, the main control module will automatically call and analyze the real-time and historical video data of the same section. For example, through visual analysis algorithms, it may discover large-area loosening of connecting bolts, failure of track pressure plates, or expanding cracks on the track beam web in this section, completing the precise positioning from macroscopic anomalies to microscopic ones. This ensures that the status judgment is not only supported by quantitative data but also by intuitive image evidence. In addition, the visual monitoring module can report... A tiny fatigue crack was discovered at a track weld. However, the static size of the crack alone may not be enough to immediately trigger the highest level alarm. In this case, the main control module retrieves long-term offset monitoring data of the weld location and analyzes the vertical or lateral displacement trend of that point. If the data shows that the deformation rate at that point has significantly accelerated after the crack appeared, even if the crack itself has not yet reached the preset threshold, the system can determine that the damage is in an active development phase and poses a substantial threat to the overall stability of the track structure, thereby raising its status risk level. In addition, by continuously receiving global offset parameters, the main control module can monitor the smoothness, elevation, and centerline position of the entire track line in real time, while visual data performs multiple local fine-tuning inspections of the baseline to detect spots and defects.

[0035] Therefore, the main control module overlays the video data provided by the visual monitoring module with the offset parameters provided by the offset monitoring module to construct a comprehensive health status profile of the track at the current moment. This profile shows whether the track is straight overall, what type of damage exists at specific locations, and how these damages affect the track geometry. For example, the system can determine that at kilometer marker K15+350, due to the loosening of three sets of connecting bolts, a 3-millimeter lateral offset and a slight change in track gauge have occurred at that point.

[0036] For example, the main control module has a tiered alarm threshold library for various fault characteristics. The thresholds are divided into several levels. The lowest level is the warning threshold, which is used to alert potential risks that need attention. The highest level is the fault threshold or shutdown threshold, which is used to indicate that the track is in an unsafe state and intervention measures need to be taken.

[0037] For example, when the visual monitoring module identifies a specific type of damage, such as a 50mm long crack in the track body, the system first queries a rule base that defines the basic risk level of such cracks at different locations and lengths. It then correlates this with data from the offset monitoring module to check the track geometry deformation in the cracked area. If the offset data is stable and shows no anomalies, the system may classify it as localized damage, a lower-level condition requiring planned maintenance. If, at the same time, there is sustained, excessive vertical displacement at the crack location, the system will infer from a preset mechanical model that the coexistence of cracks and settlement indicates a severe decrease in the track beam's load-bearing capacity and rapid damage development, triggering a higher-level structural fault alarm.

[0038] For example, if the offset monitoring module detects that the lateral offset of a certain track section suddenly exceeds the safe operating threshold of 10 mm, the main control module, upon receiving the signal, will instruct the trolley control module to immediately decelerate or stop, and request the vision monitoring module to perform focused scanning and analysis of the abnormal section. If the direct cause of the offset is found in the video analysis, such as the complete breakage of a set of critical track pressure plates or the loosening of bolts, the fault condition is finally confirmed. This mode of offset exceeding the limit triggering emergency braking plus visual confirmation of the root cause of the fault can ensure the timeliness and accuracy of the system response. In addition, for progressive faults, the main control module can also perform trend analysis. For example, if the vision system detects that the number of loose bolts is constantly increasing, and the offset data shows that the lateral instability of the track is also slowly intensifying, even if neither has reached the independent maximum fault threshold, their deterioration trends are highly correlated and continue to develop. The trend prediction algorithm of the main control module may calculate that the system will enter a dangerous state in the next few hours, which may trigger a pre-fault or serious warning state in advance and recommend maintenance in advance.

[0039] For example, the main control module can be an industrial server or industrial computer deployed in an electrical room or central control room, with multi-core processing capabilities, large-capacity memory and high-speed network interface to meet the parallel processing needs of video stream data and sensor data. The main control module may include an integrated software platform with a modular architecture, including a data acquisition and communication interface layer for establishing data connections with the cameras of the vision monitoring module, the sensors of the offset monitoring module, and the trolley control module via different protocols such as industrial Ethernet, fiber optics, or fieldbus; it may also include a multi-source data fusion and processing core with a built-in time-series database for storing massive amounts of monitoring data, integrating image recognition algorithms, point cloud processing algorithms, data filtering, and trend analysis tools to clean, align, correlate, and extract features from the input heterogeneous data; it may also include a fault diagnosis and inference engine that encapsulates a series of rules, models, and thresholds related to track structure mechanics, fault modes, and safety standards, enabling status assessment and fault diagnosis based on the fused information; it may also include a safety decision and alarm management unit for determining the control commands and alarm information to be issued based on the diagnostic results and according to preset safety policies, ensuring the priority and timeliness of the commands; and it may also include a human-machine interface that graphically displays the health status map of the entire track, the data stream of all monitoring points, historical trend curves, and all generated alarms and logs in real time, facilitating monitoring and intervention by maintenance personnel.

[0040] For example, the trolley control module can be a PLC or motion controller, directly installed in the trolley body or a nearby electrical cabinet, used to control the trolley's traction motor to control the trolley's speed, acceleration, and position. Under normal conditions, the trolley control module receives operating instructions from the quay crane central management system. In the safety closed loop of the entire monitoring system, when the main control module determines that the track is in a fault state based on fusion analysis, it sends emergency control instructions to the trolley control module via hard-wired connection or high-priority real-time network messages. These instructions could include immediately executing a smooth emergency stop, limiting the speed to below 30 m / min, or prohibiting movement in the X direction. Upon receiving these instructions, the trolley control module interrupts its current normal task, calls the underlying motion control program, and executes the main control module's commands in a safe and stable manner by adjusting the motor's torque and speed. Simultaneously, it feeds back its own status to the main control module in real time, forming a complete control closed loop. For example, the main control module can be networked northward via Ethernet or a wireless public network, and its internal communication interface with the backend meets the standard Ethernet network requirements and MQTT communication protocol needed for monitoring data exchange. To the south, it uses a wireless LoRa local area network to receive data from the monitoring host on the bracket trolley side. Furthermore, the main control module has a dry contact signal output interface, as well as data analysis and processing functions, supporting OCR recognition, structured OCR recognition, detection algorithms, classification algorithms, hybrid (detection + classification) algorithms, video event analysis algorithms, image comparison algorithms, semantic / instance segmentation algorithms, algorithm orchestration, and other algorithm capabilities. In addition, the main control module has two built-in GPUs, supports a virtual engine, and supports the mixed running of different algorithm capabilities.

[0041] For example, the main control module includes a 2U standard rack-mount 8-bay edge intelligent server that supports hot-swappable hard drives; it has 2 HDMI ports, 2 VGA ports, dual heterogeneous outputs, and supports dual 4K output; it supports a full configuration of 12TB hard drives (total capacity up to 96TB); 2 10M / 100M / 1000Mbps network ports; 2 USB 2.0 ports, 1 USB 3.0 port; 1 eSATA port; and alarm I / O interfaces: 16 alarm inputs and 4 alarm outputs. For example, the main control module has an input bandwidth of 320Mbps, an output bandwidth of 256Mbps, an access capability of 32 channels of H.264 and H.265 format high-definition streams, a maximum decoding capability of 16×1080P, RAID modes including RAID0, RAID1, RAID5, RAID6, and RAID10, and supports a global hot spare disk.

[0042] For example, the visual monitoring module includes an intelligent high-definition camera for image and video data acquisition and reporting, that is, acquiring video image data of the short rails of the trolleys on the left and right sides of the quay crane, and reporting the data to the main control module through wired communication. The intelligent high-definition camera can be installed at the corresponding positions of the short rails of the trolleys on the left and right sides of the quay crane.

[0043] For example, the intelligent high-definition camera, based on the retinal imaging principle, employs a dual-sensor architecture and dual-light fusion technology to provide a superior image visual experience, presenting color image quality as if it were daytime even in extremely low brightness. The intelligent high-definition camera also embeds a deep learning algorithm, using massive image and video resources as a foundation, and extracts target features through its own processing to form a deep, learnable target image, greatly improving the target detection rate. Furthermore, the intelligent high-definition camera supports boundary crossing detection, area intrusion detection, area entry detection, area departure detection, loitering detection, people gathering detection, and rapid movement detection. Features include detection of parking, abandoned objects, and objects being retrieved; the intelligent HD camera also has a built-in high-efficiency and gentle supplementary light to ensure normal target capture at night and supports sound and light alarms; the intelligent HD camera also supports mixed supplementary lighting, target capture distance of 3-6m, ordinary monitoring distance of 30m, supports 2560×1440@25fps real-time frame rate, supports fog penetration and has multiple white balance modes, supports three-stream technology, dual-channel HD, supports simultaneous 20-channel streaming, supports 1 pair of alarm input / output (alarm output supports a maximum of DC12V, 30mA), and 1 pair of audio input / output.

[0044] It is understood that the system provided in this application embodiment, through advanced visual recognition technology and sensor data, realizes intelligent detection and management of trolley track faults, including early fault detection, monitoring of the development process, and quality feedback after maintenance, forming a closed-loop intelligent automated management system. This ensures the goal of early fault handling and effective maintenance, greatly improving maintenance efficiency and quality. Furthermore, the system establishes a large database and algorithm model. By continuously inputting information on fault characteristics and development status, the learning algorithm can self-optimize and improve, realizing a shift from the traditional passive maintenance mode to a preventative maintenance mode, thereby improving the predictability and accuracy of maintenance. Moreover, utilizing advanced algorithms and visual recognition technology, the system can formulate preventative quay crane maintenance plans, avoiding unexpected quay crane downtime due to sudden faults, thus ensuring the continuity and stability of production. Through preventative maintenance, the system effectively reduces the failure rate, reduces maintenance costs, and improves the overall operating efficiency and reliability of the quay crane.

[0045] In one embodiment of the present application, the quay crane trolley track monitoring system further includes a monitoring management module. The main control module includes a first interface and an analysis unit. The analysis unit includes a second interface. The analysis unit is used to acquire video data through the second interface and perform fault analysis to obtain visual analysis results. The main control module is also used to acquire video data through the first interface, add visual analysis results to the video data, and then upload it to the monitoring management module.

[0046] For example, when the analysis unit is an algorithm program inside the main control module, it is essentially a set of automated software code that runs on the general-purpose or dedicated computing hardware of the main control module and is embedded in the software architecture of the main control module. It actively acquires the raw video data stream or captured key frame images transmitted from the visual monitoring module through the second interface, such as the function call interface or the subscription interface of the internal message queue. At this time, after the raw video data enters the analysis unit through the second interface, it starts the preset image processing and intelligent analysis process to obtain the visual analysis results. First, the algorithm preprocesses the input raw video frames, including Gaussian or median filtering to suppress environmental noise, adjusting image contrast and brightness to enhance feature visibility, image distortion correction to ensure the accuracy of geometric measurements, and possibly converting color images to grayscale for processing by specific algorithms. Then, for structural faults, edge detection algorithms can be used to initially locate all possible linear contours in the image. This is then combined with texture analysis or a deep learning-based semantic segmentation model to segment the cracked area from the complex background and calculate geometric parameters such as length, width, and direction. For connection faults such as loose bolts, the algorithm uses template matching to locate the reference position of the nut in the image, and then determines whether it is loose by analyzing its shape or calculating its rotational displacement relative to the initial installation angle. For bolt detachment, feature matching failure can be used for identification. For component faults, contour analysis can be used to check the integrity of its shape.

[0047] For example, when the analysis unit serves as an independent algorithm unit within the main control module, it can be a specially configured high-performance computing device, such as an industrial control computer or server with a built-in GPU, or a vision processing device integrated with a dedicated integrated circuit. This device undergoes hardware-level optimization for specific neural network models, achieving extremely high energy efficiency and speed. The independent unit has its own operating system and software environment specifically for running vision analysis algorithm libraries. The second interface between it and the main control module is not an internal software function call, but a physical network interface. It connects to the main control module via a high-speed industrial Ethernet or a dedicated data bus. Data exchange between the two uses standardized communication protocols, such as MQTT for message passing, gRPC or RESTAPI for service calls, or a custom TCP / IP-based protocol for transmitting large amounts of image data and structured analysis results.

[0048] Because the analysis unit is an independent algorithm unit within the main control module, the execution mode of the entire visual analysis process differs from that of an algorithm program within the main control module. The main control module sends the received raw video data as an external request to the independent algorithm unit via a second interface. Upon receiving the task, the independent unit utilizes all its computing resources to execute the analysis task. For example, for deep learning models, thousands of GPU cores can be used to process millions of calculations on a single image simultaneously. This ensures the real-time performance of visual analysis is not affected by the main control module's data recording, interface refresh, or communication with the offset module. Furthermore, when improved analysis capabilities are needed, the hardware or software of the independent algorithm unit can be upgraded separately. Additionally, if visual analysis requires lengthy algorithm iterations or model retraining, this can be performed on the independent unit without affecting the stable operation of the main control system.

[0049] For example, the main control module maintains a dynamic, timestamped, relational database. When the analysis unit completes its analysis through the second interface and returns structured visual analysis results—such as a JSON object containing fault type, location, confidence level, and image coordinates—the main control module precisely correlates the visual analysis results with the corresponding video frames or video clips at the time point being processed. Subsequently, the main control module, through the first interface, retrieves the original video data stream originally intended for upload to the monitoring management module. It then calls the graphics rendering engine to draw prominent markers on the original video frame images based on the original image coordinate areas provided in the visual analysis results. For example, it might use a red highlighted box to circle the identified crack area, a yellow arrow to indicate a loose bolt, and add a concise text label next to it. This image is then combined with the original video footage to generate a new enhanced image containing diagnostic information. In this way, monitoring center managers can see the precise location and nature of the fault at a glance on the video screen without having to consult data reports.

[0050] For example, metadata processing can be used to add metadata, which is data about data, i.e., structured visual analysis results. The main control module packages the analysis results data in a standardized format and transmits it through a video stream metadata channel that can be transmitted in parallel with the video stream. The analysis results data are sent out together through this video stream metadata channel, and the metadata in the monitoring and management module can maintain precise synchronization with the video frames. Alternatively, a data encapsulation container can be used, for example, packaging the video stream and analysis results together into a container format. Or, in some web-based technical solutions, the WebRTC data channel can be used to synchronously transmit video and JSON format analysis results.

[0051] For example, when the analysis unit is an algorithm program inside the main control module, it can read track video frames through the YUV service interface and execute the intelligent recognition process. The algorithm program can also report metadata (including the recognition results of the intelligent algorithm and captured images, etc.) to the monitoring platform through the SDC data channel. The main control module can obtain the original real-time video recording through the PGSP video interface and overlay the intelligent recognition results on the video to achieve complete intelligent recognition function.

[0052] In one embodiment of the quay crane trolley track monitoring system provided in this application, triggering a fault alert includes sending a fault alert to the monitoring management module, and the fault alert includes the fault type and the corresponding fault data.

[0053] For example, when a track is confirmed to be in a fault state, a fault alert system task is activated. Internally, the main control module dynamically generates a structured fault alert data object, instantiated in memory according to a predefined machine-readable data structure, including the fault type and corresponding fault data. The fault type is a standardized classification identifier that uses enumeration values ​​or specific codes to indicate the nature of the fault. Standardization avoids ambiguity in natural language, facilitating accurate parsing and classification by the receiving program. The corresponding fault data can be a nested data structure containing all key data points related to the fault type extracted from various analysis modules, as well as associated parameters extracted from the offset monitoring module. For example, the main control module can send the structured fault alert data through a communication interface agreed upon with the monitoring and management module, via a network-based API call. For instance, the main control module, acting as an HTTP client, sends a request to a specific RESTful API endpoint exposed by the monitoring and management module, transmitting the encapsulated fault alert JSON object in the request body.

[0054] In a quay crane trolley track monitoring system provided in one embodiment of this application, the visual monitoring module is further used to perform preliminary fault analysis and event detection to obtain preliminary visual analysis results, and the analysis unit is further used to combine the preliminary visual analysis results and video data to perform fault analysis to obtain visual analysis results.

[0055] For example, for preliminary fault analysis, the visual monitoring module can adopt fast feature extraction methods based on traditional computer vision. For instance, it can use efficient background subtraction algorithms to detect sudden changes in the scene, or use simple edge detection and contour analysis to find areas in the image that are significantly different from the contours of normal track structures. At the same time, it can also integrate a small convolutional neural network to perform fast binary or multi-class classification tasks. For instance, it can simply classify image frames as normal or suspected abnormal, or further distinguish between categories such as suspected cracks or suspected bolt abnormalities. Its output is a preliminary classification label with confidence.

[0056] For event detection, the visual monitoring module can analyze the differences between consecutive frames. For example, the motion vector of pixels in the image can be estimated by optical flow. If an unexpected moving object that does not conform to the normal operation mode of the vehicle is detected in a specific area of ​​the track, or if abnormal periodic displacement of the track pressure plate is detected during vibration, it will be marked as a suspicious event. In addition, for specific phenomena such as smoke and sparks, a dedicated detector based on color features and motion patterns can be integrated for rapid identification.

[0057] After completing the initial analysis, the preliminary visual analysis result generated by the visual monitoring module can be a data packet, including the timestamp of the alarm trigger, the camera number, the preliminary determined fault or event type, the approximate bounding box coordinates of the abnormal area in the image, an initial confidence score, and the associated raw video data segment.

[0058] For example, after receiving preliminary visual analysis results from the visual monitoring module, the analysis unit can perform credibility assessment and task parsing on the received preliminary results. Based on the camera ID that sent the preliminary results, the historical false alarm rate of the fault type, and the confidence score of the preliminary results, it determines which analysis strategy to initiate and how much computing resources to allocate. For instance, a high-confidence alarm for a suspected severe crack from a critical area will trigger the highest-priority analysis thread. Subsequently, the analysis unit can locate and load the specific video segment pointed to in the preliminary result data packet and extract the bounding box coordinates of the abnormal area provided in the preliminary results, forming the region of interest for subsequent refined analysis. This is equivalent to transforming a wide-angle search task into a fixed-point magnification detection task, significantly improving efficiency.

[0059] For example, when performing fault analysis by combining preliminary visual analysis results and video data, the analysis unit can call multiple deep learning models or traditional visual algorithms with higher computational requirements to analyze the region of interest. For instance, for suspected cracks indicated by the preliminary analysis, a pixel-level semantic segmentation model can be run in the region to accurately calculate the true length, maximum width, average width, branching and other geometric parameters of the crack. At the same time, a material defect classification model can be run to help determine whether it is a fatigue crack, stress corrosion crack or surface scratch.

[0060] For example, for bolt anomalies indicated by preliminary analysis, the analysis unit can first use more precise feature point detection or template matching algorithms to accurately locate each bolt and nut within the initially defined area. Then, it can use attitude estimation algorithms to measure the rotation angle of the nut, or use super-resolution technology to enhance image details to check for minute cracks in the bolt neck.

[0061] refer to Figure 2 , Figure 2 This is a structural diagram of a quay crane trolley track monitoring system provided in an exemplary embodiment of this application. In the quay crane trolley track monitoring system provided in an embodiment of this application, the system also includes a monitoring gateway module. The monitoring gateway module is connected to the offset monitoring module and the main control module respectively. The monitoring gateway module is used to obtain the offset parameters of the offset monitoring module and send them to the main control module.

[0062] For example, the monitoring gateway module can perform protocol conversion and data unification. The sensors used by the offset monitoring module may come from different manufacturers, and the communication protocols used may be traditional fieldbuses such as Modbus and Profibus, or various proprietary protocols based on serial ports. If the main control module were to directly adapt to these complex protocols, it would greatly increase the software complexity and maintenance cost of the main control module. However, the monitoring gateway module integrates driver libraries for various mainstream industrial protocols. It can actively poll or passively receive data from different offset sensors, then parse these heterogeneous data to extract meaningful offset parameters, and finally encapsulate them into a unified and standardized data format. This data is then sent out through common protocols that the main control module can process, such as Ethernet and MQTT, providing the main control module with a unified and easy-to-understand data interface, which greatly simplifies system integration.

[0063] For example, in addition to the offset monitoring module, the monitoring gateway module can also connect to various other types of monitoring modules. For instance, it can connect to environmental monitoring sensors, such as temperature and humidity sensors installed along the track and on the quay crane structure, to perform temperature compensation on the offset monitoring data using ambient temperature data. It can also connect to anemometers, allowing the main control module to combine wind speed and track offset to formulate more scientific speed limits for the trolley. Furthermore, it can connect to vibration sensors, which can detect potential faults such as loose bolts and worn components by analyzing the vibration spectrum characteristics of the track structure.

[0064] For example, the monitoring gateway module can also integrate an electrical and drive system monitoring module, and connect to a power analyzer that monitors the current, voltage, and power factor of the trolley's traction motor. Abnormal current fluctuations may indicate increased track resistance, motor malfunction, or transmission mechanism jamming. It can also access signals from encoders or RFID positioning systems to accurately calibrate the trolley's position and combine it with offset monitoring data to analyze whether a specific track deformation always occurs at a specific location, thus achieving accurate fault location.

[0065] For example, the monitoring gateway module can be an IoT gateway for sensor data acquisition, data reporting and processing, that is, to collect the anti-deviation sensor data of the short rails of the quay crane on both sides and report the data to the main control module through wired communication. The monitoring gateway module can be installed at the corresponding position of the short rails of the quay crane trolley.

[0066] For example, the monitoring gateway module can communicate with the main control module via wired Ethernet to the north, and directly connect to the anti-deviation sensor to collect sensor data to the south.

[0067] In one embodiment of this application, the quay crane trolley track monitoring system further includes a POE switch module. The POE switch module is connected to the visual monitoring module and the main control module, respectively. The POE switch module is used to power the visual monitoring module and to acquire video data from the visual monitoring module and send it to the main control module.

[0068] For example, in addition to connecting the cameras of the visual monitoring module, the PoE switch module can also connect to PoE-enabled smart sensor nodes. For instance, it can deploy high-precision PoE network audio sensors to collect sound characteristics of the track area during operation. By analyzing the audio signals, it can effectively identify abnormal noises during trolley operation, impact noises at track joints, and the periodic friction sounds characteristic of early-stage bearing damage. This complements the visual images, enabling the detection of unseen faults. It can also connect to PoE-enabled miniature weather stations to obtain real-time, precise wind speed, wind direction, temperature, and humidity data for the local area where the track is located, allowing the main control module to comprehensively determine the cause of track deviation. Furthermore, the PoE switch can also connect to a monitoring gateway module, which is powered by the switch.

[0069] In the quay crane trolley track monitoring system provided in one embodiment of this application, the main structure faults include crack faults and fracture faults, the connection faults include connection error faults, joint gap error faults, and alignment error faults at the track joint, and the component faults include pressure plate missing faults and bolt missing faults.

[0070] In one embodiment of the quay crane trolley track monitoring system provided in this application, the main control module is further used to control the trolley to decelerate or stop according to the fault type of the trolley track when it is determined that the trolley track is in a fault state.

[0071] For example, structural failures include crack failures and fracture failures. The identification and tracking of track cracking phenomena, particularly at the welding points of long track interfaces, are addressed. Structural failures may include: General cracks: Cracks with an opening of less than 0.5mm will trigger a crack warning in the background system. To control the normal speed of the trolley, it is necessary to tighten bolts and adjust shims in a timely manner to ensure the technical requirements of the trolley track are met. Severe cracks: When the cracking degree is between 0.5 and 3 mm, the background system will issue a crack alarm and control the trolley to slow down. It is necessary to prepare relevant maintenance plans and implementation plans to repair the trolley track cracks as soon as possible. Track breakage: When the track breaks by more than 3mm, the background system will issue a track breakage fault warning and control the trolley to stop running.

[0072] For example, the track joint is a unique connection point of the trolley track. To maintain its separated state, it must be divided into two sections. To ensure the smooth operation of the trolley, the connection error, joint gap, and alignment accuracy of the two track sections must meet the corresponding requirements. Connection faults include connection error faults at the track joint, joint gap error faults, and alignment error faults. Connection error fault: The connection error shall not exceed ±0.5mm. If the error exceeds the standard, the background system will show a joint height deviation fault, control the trolley to decelerate, and adjust the connection part as soon as possible to restore its technical condition, ensure the smooth operation of the trolley, and avoid the expansion of the fault range. Joint gap error fault: The joint gap must not exceed 4±0.5mm. If the error exceeds the standard, the background system will indicate a joint gap fault, control the trolley to slow down, and adjust the joint as soon as possible to restore its technical condition, ensure the smooth operation of the trolley, and avoid the expansion of the fault range.

[0073] Centering error fault: The centering accuracy must not exceed ±0.5mm. If the error exceeds the standard, the control trolley will decelerate and the background system will show a track centering deviation fault. The track centering part needs to be adjusted as soon as possible to restore its technical condition, ensure the smooth operation of the trolley, and avoid the expansion of the fault range.

[0074] For example, the track fixing module is located on both sides of the track, and its function is to fix the trolley track. It consists of two main parts: pressure plates and bolts. If the pressure plates or bolts are missing, the track will shift, causing a malfunction. Therefore, component malfunctions include pressure plate missing malfunctions and bolt missing malfunctions. Pressure plate missing fault: If a pressure plate is detected missing in the pressure plate monitoring area, the background system will issue a pressure plate missing fault warning, control the trolley to stop running, and the missing pressure plate needs to be replaced as soon as possible.

[0075] Bolt missing fault: If a bolt is detected missing in the bolt detection area, the background system will issue a bolt missing fault warning, control the trolley to slow down, and the missing bolt needs to be replaced as soon as possible.

[0076] In one embodiment of the quay crane trolley track monitoring system provided in this application, the main control module further includes a fault feature database. The fault feature database is used to collect video data and monitor offset parameters, and to classify track fault feature states and judge the patterns of fault development trends in order to achieve fault early warning.

[0077] For example, the fault feature database can classify track fault feature states through a multi-level classification model system. This system collects historical data and expert knowledge, and continuously collects and stores structured analysis results from the visual monitoring module and quantitative parameters from the offset monitoring module. For instance, maintenance experts can label the data based on mechanical principles and historical experience. For example, a crack length greater than 20mm accompanied by a vertical displacement exceeding 5mm at that point can be marked as high-risk structural damage. A condition where multiple bolts on one side of the track show slight loosening and lateral offset fluctuates within 2-5mm can be marked as decreased connection reliability. The database system then uses unsupervised machine learning clustering algorithms to automatically explore massive amounts of unlabeled data, discovering naturally occurring fault pattern combinations. For example, the algorithm might automatically identify a feature cluster characterized by a periodic increase in lateral offset in specific sections of the track under high-temperature weather conditions, without visual damage, thus defining a new type of fault-induced geometric instability.

[0078] For known fault categories, the system employs a supervised classification algorithm, using labeled data for training, to generate a model that can automatically classify real-time incoming monitoring data. When new data arrives, the model calculates the probability of it belonging to each fault category and assigns it to the most probable category, thereby achieving automated state classification of fault characteristics.

[0079] To determine the patterns in fault development trends, the fault characteristic database establishes an independent time-series data archive for each key monitoring point, recording the complete history of parameter changes over time. Through in-depth analysis of the time-series data, the system can perform various pattern judgments. For example, trend analysis uses algorithms such as linear regression, exponential smoothing, or ARIMA (autoregressive integral moving average model) to fit trend lines of parameter changes over time, thereby determining whether the fault is in a stable state, a slow development phase, or an accelerated deterioration phase. For instance, the system might calculate that a crack is steadily expanding at a rate of 0.5 mm per month, while the settlement at another point is continuously sinking at an acceleration of 0.1 mm per week. For correlation analysis, the database uses statistical methods to analyze the linkages between different parameters, thereby discovering hidden patterns. For example, it might find a strong positive correlation between the lateral offset at a certain point on the track and the square of the daily average wind speed, thus determining that the offset at that point is mainly controlled by wind load; or it might find that a certain type of bolt loosening always preferentially appears on the outside of the track curve, characterizing the uneven distribution of load.

[0080] In one embodiment of the quay crane trolley track monitoring system provided in this application, the fault feature database is also used to generate continuous tracking records of the fault formation initiation, development, and maintenance quality of the trolley track.

[0081] For example, the fault feature database does not wait for a fault to fully manifest before recording it. Instead, based on real-time analysis capabilities, when the visual monitoring module's preliminary analysis or the analysis unit's in-depth analysis first identifies an abnormal feature that exceeds the normal fluctuation range—for example, detecting a microcrack only 0.5 mm long at a track weld for the first time, or the offset monitoring module detecting that the settlement rate at a certain point accelerates from 0.1 mm per month to 0.3 mm per month—this time point, along with all the initial parameters of the abnormal feature, is captured by the database and created as the starting event of a new fault event, establishing a timeline origin and a quantitative baseline for the initial state of the fault. All subsequent developments will be compared based on this. At the same time, the database will associate this event with historical normal state data at the same location, forming the beginning of a complete story of before-and-after comparison.

[0082] After capturing the starting point, the fault feature database automatically and frequently tracks and records the fault development stages. According to a preset sampling period or based on event triggers, the latest monitoring data of the fault point is continuously added to the tracking record, including the evolution of fault body parameters, such as the growth curve of crack length, changes in width, whether branches appear, the increase in the number of loose bolts, the expansion of the loosening angle, the cumulative value of offset and the fluctuation of instantaneous rate, etc. It will also record the related environmental and load context simultaneously, such as the ambient temperature during the fault development, the trolley passage and load at the recording time, and the real-time wind speed. All these data are sorted by timestamp to form a multi-dimensional time series of fault development. The trend analysis algorithm embedded in the database will run on these sequences in real time to automatically calculate key development indicators such as the fault expansion rate and acceleration, and may identify different stages of development.

[0083] When maintenance activities occur, maintenance personnel report the execution and completion of maintenance work orders through the system. The fault feature database records this maintenance as an event node in the fault record, detailing the maintenance time, maintenance process used, information on replaced parts, and responsible personnel. Within a specific verification period after maintenance, the fault feature database activates a monitoring mode, comparing the new data collected after maintenance with the fault state before maintenance and the ideal design standard state. For example, for repaired cracks, it continuously monitors whether cracks reappear in the area; for tightened bolts, it monitors whether the preload remains stable; for corrected track geometry deformation, it monitors whether the offset returns to and remains stable within the allowable tolerance zone in the long term. Through continuous comparison of post-maintenance data and pre-maintenance data, a quantitative evaluation of the maintenance quality can be automatically generated. For example, if the condition is stable after maintenance and parameters return to the normal range, the maintenance quality is excellent; if the condition improves in the short term after maintenance but the parameters show a trend of deterioration again, the maintenance effect needs to be observed or the fault characteristics are not eliminated after maintenance, the maintenance is invalid. The evaluation, along with all the verification data after maintenance, is then fully added to the tracking record.

[0084] In the several embodiments provided in this application, it should be understood that the disclosed systems, instruments, and methods can be implemented in other ways. For example, the instrument embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between instruments or units may be electrical, mechanical, or other forms. Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0085] It should also be understood that the various implementation methods provided in this application can be combined arbitrarily to achieve different technical effects.

[0086] The embodiments of this application have been described in detail above with reference to the accompanying drawings. However, this application is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of this application.

Claims

1. A track monitoring system for a quay crane trolley, characterized in that, include: The visual monitoring module is used to monitor structural faults, connection faults, and component faults of the trolley track. Offset monitoring module, used to monitor offset faults in the trolley track; The vehicle control module is used to control the operation of the vehicle; The main control module acquires video data from the visual monitoring module and offset parameters from the offset monitoring module through different interfaces, and is connected to the trolley control module. The main control module monitors the status of the trolley track through the video data and the offset parameters, and controls the trolley through the trolley control module and triggers a fault alert when it determines that the trolley track is in a faulty state.

2. The system according to claim 1, characterized in that, The system also includes a monitoring and management module. The main control module includes a first interface and an analysis unit. The analysis unit includes a second interface. The analysis unit is used to acquire the video data through the second interface and perform fault analysis to obtain visual analysis results. The main control module is also used to acquire the video data through the first interface, add the visual analysis results to the video data, and then upload it to the monitoring and management module.

3. The system according to claim 2, characterized in that, The triggering of the fault alert includes sending a fault alert to the monitoring and management module, and the fault alert includes the fault type and the corresponding fault data.

4. The system according to claim 2, characterized in that, The visual monitoring module is also used to perform preliminary fault analysis and event detection to obtain preliminary visual analysis results. The analysis unit is also used to combine the preliminary visual analysis results and the video data to perform fault analysis to obtain visual analysis results.

5. The system according to claim 1, characterized in that, The system also includes a monitoring gateway module, which is connected to the offset monitoring module and the main control module respectively. The monitoring gateway module is used to obtain the offset parameters of the offset monitoring module and send them to the main control module.

6. The system according to claim 1, characterized in that, The system also includes a PoE switch module, which is connected to the visual monitoring module and the main control module respectively. The PoE switch module is used to power the visual monitoring module and to acquire video data from the visual monitoring module and send it to the main control module.

7. The system according to claim 1, characterized in that, The structural faults include crack faults and fracture faults; the connection faults include connection error faults, joint gap error faults, and alignment error faults at the track joint; and the component faults include missing pressure plate faults and missing bolt faults.

8. The system according to claim 1, characterized in that, The main control module is also used to control the trolley to decelerate or stop according to the fault type of the trolley track when it is determined that the trolley track is in a fault state.

9. The system according to claim 1, characterized in that, The main control module also includes a fault feature database, which is used to collect the video data and the offset parameter monitoring, and to classify the track fault feature status and judge the pattern of fault development trend in order to realize fault early warning.

10. The system according to claim 9, characterized in that, The fault feature database is also used to generate continuous tracking records of the fault formation, development, and maintenance quality of the trolley track.

Images