Automatic elevator hoist rope oiling system and method based on machine vision

Abstract

The invention discloses an elevator hoist rope automatic oiling system and method based on machine vision. The system comprises an industrial camera, a development board integration module, an oil supply mechanism, an oil scraping mechanism and the like. The system integrates the NVIDIA image segmentation device, the power supply lithium battery (24V), the voltage conversion module and the relay module, and is fixed above a camera support. The industrial camera is fixed to the camera support, located above the traction rope and the traction wheel, connected with NVIDIA equipment and used for detecting the oil film coverage rate on the traction rope and judging whether the oil film coverage rate reaches the standard or not. And an oil supply pump in the oil supply mechanism is driven by a signal sent by the development board to oil and maintain the hoist rope. The rotation of the stepping motor is controlled through the signal sent by the development board, the combination and separation of the traction rope oil scraping plate are realized through the synchronous belt mechanism and the lead screw nut mechanism which are connected with each other, and the traction wheel oil scraping plate and the traction wheel are in slight contact through the small spring force, so that the oil removal purpose is achieved. And finally, redundant oil liquid removed from the traction system can flow into the oil storage tank.

Description

Technical Field

[0001] The present invention relates to the fields of elevator safety protection devices and machine vision technology, and in particular to a machine vision-based automatic oiling system and method for an elevator traction rope. Background Art

[0002] With the continuous advancement of science and technology, the elevator industry, as a specialized equipment sector within the manufacturing industry, has been at the forefront of the industry's development and is closely intertwined with people's daily lives. By 2024, the total number of special equipment in China will reach 19.5525 million units, including 9.6446 million elevators. The popularity of residential elevators in China has grown significantly in recent years, particularly amidst the rapid pace of urbanization. More and more residential buildings are equipped with elevators. However, despite this increase in elevator coverage, many installed elevators still face significant issues in terms of safety, convenience, and intelligence. The performance of elevator ropes, as the core transmission component of elevator systems, directly impacts the safety and service life of elevators. Currently, elevator rope lubrication and maintenance primarily rely on manual labor, which presents challenges such as poor lubrication uniformity, low oil control accuracy, and safety hazards associated with high-altitude operations. Traditional manual lubrication methods struggle to ensure uniform oil film penetration, leading to localized dry friction and wire wear. Oil quantity deviations can reach as high as ±40%, and maintenance procedures can result in long downtime and high operational risks. Existing automated solutions still have significant shortcomings: contact detection is susceptible to interference from oil films, infrared sensors have a high misjudgment rate in oily environments, and timed and quantitative lubrication modes cannot adapt to varying operating speeds, resulting in a significant decrease in lubrication effectiveness at high speeds. Furthermore, traditional mineral greases exhibit unstable performance at high temperatures, and existing systems lack interoperability with IoT platforms, making intelligent maintenance difficult. Therefore, an automated lubrication system capable of real-time monitoring, precise control, and adaptability to complex operating conditions is urgently needed.

[0003] The invention patent application publication number CN212712278U is titled "A Steel Wire Rope Maintenance and Oiling Device for an Elevator Traction System." In this steel wire rope maintenance and oiling device for an elevator traction system, the limiting pulley moves along with the steel wire rope, and due to the arrangement of the slide groove and the slider, the drip rack moves along with the movement of the limiting pulley, so that the oil dripping port at the other end of the drip rack can evenly oil the steel wire rope just wound on the reel. This method has higher oiling accuracy and does not cause waste, solving the problem that the existing steel wire rope maintenance and oiling device for an elevator traction system cannot evenly oil the steel wire rope. Compared with traditional methods, this method improves the uniformity problem of traditional manual oiling through mechanical linkage, but it still belongs to passive lubrication and cannot be intelligently adjusted, adapted to complex working conditions, or achieve data management.

[0004] The utility model patent application, "An Elevator Rope Oiling Auxiliary Device," with publication number CN220976193U, describes a mechanical structure for an auxiliary oiling device. This device, equipped with a drive mechanism, a driven mechanism, and a sliding portion, draws excess lubricating oil from the surface of the wire rope into the main drainage pipe as it revolves around the drive rod. The oil then flows along a barrier plate to an oil storage tank connected to the branch drainage pipe, thus avoiding some lubricating oil waste. While mechanical lubricating oil recovery solutions can reduce some waste, they lack intelligent detection, dynamic adjustment, and data analysis capabilities.

[0005] Therefore, how to better improve the safety of community owners when taking elevators, better meet the community owners' elevator experience, and improve the convenience and intelligence of community elevators are key issues that need to be urgently addressed at this stage. Summary of the Invention

[0006] The purpose of this patent is to provide a machine vision-based automatic lubrication system and method for elevator traction ropes, addressing the aforementioned issues with the prior art. This technology can be widely applied to residential elevator traction systems. Machine vision is used to inspect the traction ropes in the traction system to determine whether the oil film coverage meets the required standards. Image segmentation technology is used to segment the captured rope image, determining the number of oil-covered and uncovered pixels. The oil film coverage is then determined to determine whether the standard is met. If the oil film coverage exceeds the required standards, the lubrication system does not need to be activated. If the oil film coverage falls below the required standards, the oil supply pump is activated to apply oil to the traction ropes using a spray nozzle, and a rubber scraper is used to remove excess oil from the ropes and traction sheaves. This system and method are efficient, accurate, and applicable in a variety of scenarios, particularly with respect to lubrication efficiency. This not only effectively improves the safety of residential residents but also reduces human intervention, improving the elevator ride experience.

[0007] To achieve the above-mentioned purpose, the present invention provides the following solution: The present invention provides an elevator traction rope automatic oiling system and method based on machine vision, including an industrial camera, a light source, an oil supply pump, an oil nozzle, an oil scraper, an oil storage tank, and a development board integrated module.

[0008] The development board integrated module includes an NVIDIA Jetson Xavier NX image segmentation device, a 24V lithium battery, a voltage conversion module, and a relay module. The development board integrated module is fixed to the housing of the traction machine with different screws.

[0009] The industrial camera is fixed above the traction sheave and traction rope via a fixture and connected to the NVIDIA image segmentation device. It identifies the oil film coverage on the traction rope and determines whether the oil film coverage meets the standard. If it exceeds the standard, the oiling system does not need to be activated. If it does not, the oil supply pump is activated to oil the traction rope using the oil nozzle for maintenance, and a rubber scraper is used to remove excess oil from the traction rope and traction sheave.

[0010] The NVIDIA image segmentation device is fixed to the development board integrated module by screws, connected to the industrial camera, the power supply lithium battery (24V), the relay module and the voltage conversion module, and mainly runs the relevant detection algorithm.

[0011] The oil supply pump is fixed to the upper side of the traction wheel and the traction rope through a fixing device, is started and stopped by the signal given by the NVIDIA image segmentation device, and is connected to the left fuel injector and the oil storage tank located on the lower side of the traction wheel using an oil pipe.

[0012] There are two types of oil scrapers. One is fixed in the middle of the oil storage tank. The oil scraper is located on the lower side of the traction wheel. A small force is applied by a small spring to make the scraper and the traction wheel fit together to remove the oil stains on the traction wheel. The other is fixed on both sides of the oil storage tank. The oil scraper is in a closed state when the oiling system is working and fits tightly with the traction rope. When the oiling system stops working, the oil scraper is separated from the traction rope.

[0013] The oil nozzle is fixed to the upper side of the traction wheel and the traction rope through a fixing device, and is connected to the oil supply pump through an oil pipe, so as to guide the oil to flow accurately to each traction rope.

[0014] The oil reservoir is secured to the hoist housing via a fixture, located below the traction sheave. A first-type oil scraper is located in the center of the reservoir, along with second-type oil scrapers on either side, respectively, for removing excess oil from the traction sheave and traction rope and returning it to the reservoir. The bottom of the reservoir is connected to an oil pump via an oil pipe, which supplies oil to the fuel injector.

[0015] The present invention discloses the following technical effects: the present invention can realize real-time detection of the oil film on the surface of the traction rope in the traction system of a residential elevator, and judge whether the oil film coverage rate meets the standard through an image segmentation algorithm; then, the traction rope is oiled and maintained in combination with the oil supply system, and the excess oil on the traction wheel and the traction rope is removed through an oil storage tank with two types of oil scrapers, and the excess oil is recovered into the oil storage tank; the traction rope is automatically oiled and maintained by controlling the elevator control system.

[0016] Compared with the prior art, the present invention has the following advantages:

[0017] (1) Through machine vision technology, the system can automatically detect the oil film coverage of the traction rope, achieving manual and contactless measurement. Compared with traditional manual contact detection, this greatly improves the convenience of maintenance.

[0018] (2) The DeepLabV3+ network segmentation model has a high accuracy in image segmentation, and the model using MobileNetV2 as the backbone is relatively small, making it easy to deploy to the above-mentioned development board, making the system's segmentation detection speed far exceed traditional manual detection.

[0019] (3) High-resolution industrial cameras + AI visual algorithms are used to detect the oil film thickness, coverage and wear status on the traction rope surface in real time, and the oil output is dynamically adjusted with an error of ≤±5% (the traditional manual method has an error of ±40%).

[0020] (4) Machine vision replaces contact probes to avoid mechanical wear interfering with detection accuracy and increase service life by more than 3 times; closed-loop feedback system: real-time image analysis → oil pump PID control → secondary detection and verification, forming a continuous optimization closed loop.

[0021] (5) This system is not only suitable for ordinary residential areas, but can also be used in other places with high security and efficiency requirements, and has a wider range of application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 The figure is an overall structural diagram of an automatic lubrication system and method for elevator traction ropes based on machine vision in an embodiment of the present invention.

[0024] Figure 2 This is a structural diagram of the detection and oil supply parts of a machine vision-based elevator traction rope automatic oiling system and method in an embodiment of the present invention.

[0025] Figure 3 This is a structural diagram of a traction rope oil scraper of an elevator traction rope automatic oiling system and method based on machine vision in an embodiment of the present invention.

[0026] Figure 4 This is a structural diagram of the traction sheave scraper of an elevator traction rope automatic oiling system and method based on machine vision in an embodiment of the present invention.

[0027] Figure 5The figure is an overall workflow diagram of an automatic lubrication system and method for elevator traction ropes based on machine vision in an embodiment of the present invention.

[0028] Figure 6 This is a network structure diagram of the DeepLabV3+ image segmentation model of a machine vision-based elevator traction rope automatic oiling system and method in an embodiment of the present invention.

[0029] Among them, 1 is an industrial camera, 2 is an oil nozzle, 3 is an oil pump, 4 is an oil pipe, 5 is a traction rope oil scraper, 6 is a traction wheel oil scraper, 7 is an oil storage tank, 8 is a synchronous belt mechanism, 9 is a screw nut mechanism, and 10 is a stepper motor. DETAILED DESCRIPTION

[0030] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0031] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] The present invention provides a machine vision-based automatic oiling system and method for elevator traction ropes, as shown in FIG1 to FIG6 , comprising: a machine vision detection system, an oil supply mechanism, and an oil scraping mechanism.

[0033] The machine vision inspection system consists of an industrial camera 1, secured to the underside of the traction sheave housing bracket with screws; a development board integrated module 2, which includes an NVIDIA image segmentation device, a 24V lithium battery, a relay module, and a voltage conversion module. The development board integrated module 2 is secured to the top of the traction sheave housing bracket (on the back of the industrial camera 1) with various screws.

[0034] The NVIDIA edge detection device is fixed to the development board integrated module 2 by screws, and is connected to the industrial camera 1, the power supply lithium battery (24V), the relay module and the voltage conversion module, and mainly runs the relevant detection algorithm.

[0035] like Figure 1As shown, industrial camera 1 is fixed above the traction sheave and traction rope via a fixture and connected to an NVIDIA image segmentation device. The improved DeepLabV3+ image segmentation algorithm deployed on the NVIDIA image segmentation device determines the oil film coverage on the traction rope surface and determines whether the traction rope requires oiling and maintenance. To improve the efficiency and accuracy of oil film segmentation on the traction rope surface, the original DeepLabV3+ network was modified based on the characteristics of the dataset. The main focus is on reducing the backbone network weight, improving the ASPP, and adding an attention mechanism. The details are as follows.

[0036] (1) Since the system needs to detect the oil film on the surface of the traction rope in real time and does not require high segmentation accuracy, in order to further improve the segmentation speed of the network, only the first 8 layers are selected from the 21 layers in the original backbone Mobilenetv2 network. This can significantly reduce the model size and improve the segmentation efficiency while slightly reducing the segmentation accuracy. Figure 6 shown.

[0037] (2) Considering the irregularity of the oil film distribution on the traction rope and the slight decrease in segmentation accuracy caused by the compensation step (1), especially in the specific scene of the elevator shaft with insufficient lighting, the first CBAM attention mechanism is inserted between the backbone network (mobilenetv2) of the original DeepLabV3+ network and the input and output of ASPP, and the second CBAM attention mechanism is inserted before the first upsampling after ASPP. The overall structure of the CBAM attention mechanism can be divided into a channel attention module and a spatial attention module. The channel attention module is used to assign different weights to different feature channels; the spatial attention module is used to allocate attention in the spatial dimension of the feature map and select important spatial regions. The adaptive fusion mechanism will perform weighted fusion on the outputs of the above two to generate the final attention feature. The specific formula is as follows.

[0038] (1) Channel Attention Module

[0039] The channel attention module is used to dynamically adjust the weight of each channel based on global information and obtain the global information of the channel through global average pooling.

[0040]

[0041] in, represents the pixel value of the input feature map, and Represents the height and width of the feature map (corresponding to the width and height of the prior box). The global information is processed through a fully connected network, and the generated weights are activated using the ReLU function. With input features Multiply them together to get the weighted channel features.

[0042]

[0043] in, is the ReLU activation function, is the Sigmoid function, Expressed as weight.

[0044] (2) Spatial Attention Module

[0045] The spatial attention module is responsible for selecting important regions in the feature map. It fuses channel features through max pooling to obtain spatial information. It then performs a convolution operation on the fused spatial features to generate spatial attention weights. Represented as spatial information.

[0046]

[0047] The resulting spatial weights With input features Multiply them together to get the weighted spatial features. The specific formula is shown below.

[0048]

[0049] (3) In order to compensate for the slightly decreased segmentation accuracy caused by step (1) and better retain the underlying image features of the oil film on the traction rope, and at the same time integrate multi-scale features, a multi-branch cascade ASPP structure is selected, such as Figure 6 As shown, the output of each branch is concatenated with the input of all subsequent branches, and the update formula is as follows.

[0050]

[0051]

[0052]

[0053]

[0054] Input feature map Give the model, through The dilated convolution is used to connect to the next feature map and so on to form a multi-branch cascade structure.

[0055] Adding the above three methods to the original DeepLabV3+ network structure can improve the detection efficiency and accuracy of the segmentation network for the oil film coverage of the traction rope to a certain extent.

[0056] The main workflow of the improved DeepLabV3+ image segmentation algorithm is as follows:

[0057] (1) Construct a traction rope image dataset, use Labelme software to annotate targets and key points, and divide the dataset into training and validation sets according to a certain ratio;

[0058] (2) The constructed dataset is trained through the improved DeepLabV3+ image segmentation network to obtain a model for real-time rope oil film coverage. Since the improved training model has a small number of parameters, it can be directly deployed;

[0059] (3) Deploy the model to NVIDIA edge detection equipment to detect the oil film coverage of the elevator traction rope in real time;

[0060] If the oil film coverage of the traction rope is higher than the standard, there is no need to start the automatic oiling system; if the oil film coverage of the traction rope is lower than the standard, the oiling system and the oil scraping system will cooperate with each other to achieve the purpose of providing automatic oiling maintenance for the traction rope.

[0061] The oil supply mechanism includes an oil nozzle 3, an oil pump 4, an oil supply pipe 5, and an oil reservoir 8. The oil pump 4 is activated and deactivated by signals from the machine vision inspection system. The pump 4 draws oil from the reservoir, supplies it to the oil nozzle 3, and sprays it onto the traction rope. The oil nozzle 3 and oil pump 4 are fixed above the traction sheave and traction rope. The oil supply pipe 5 connects the oil pump 4 to the oil reservoir 8, which is fixed below the traction sheave.

[0062] There are two types of oil scraping mechanisms. The first type is Figure 3 As shown, it includes a traction rope scraper 6 (on both sides of the traction rope), a synchronous belt mechanism 9, a screw nut mechanism 10 and a stepper motor 11. The rotation of the stepper motor 11 is controlled by the signal provided by the machine vision detection system. The stepper motor 11 is connected to the left side of the screw nut mechanism 10 through the synchronous belt mechanism 9, and the right synchronous belt mechanism 9 is connected to the two screw nut mechanisms 10 to ensure that the rotation angles of the two screw nut mechanisms are the same. The two traction rope scrapers 6 are fixed to the two screw nuts 10 by screws, and the merging and separation are achieved by the rotation of the screw nut mechanism; the second type is as follows Figure 4 As shown, the traction sheave oil scraper 7 is located between the oil storage tank 8 and the traction sheave. By applying a small spring force, the traction sheave oil scraper 7 is slightly in contact with the traction sheave to remove oil from the traction sheave.

[0063] In the description of the present invention, it should be understood that the fixed or positional relationship indicated in the relevant description is based on the orientation or positional relationship shown in the accompanying drawings, which is only for the convenience of describing the patent method, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0064] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A machine vision-based automatic lubrication system and method for elevator traction ropes, characterized by: Industrial cameras, development integrated boards, oil supply mechanisms, oil scraping mechanisms, etc.

2. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The development board integrated module includes an NVIDIA Jetson Xavier NX image segmentation device, a power supply lithium battery (24V), a voltage conversion module, and a relay module.

3. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The development board integrated module is fixed to the back of the camera bracket by different screws.

4. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The industrial camera is fixed above the traction wheel and the traction rope through a fixing device and is connected to the NVIDIA Jetson Xavier NX image segmentation device to segment the collected traction rope image and determine whether the oil film coverage on the traction rope meets the standard.

5. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 4, characterized in that: The image segmentation algorithm uses an improved DeepLabV3+ network model and proposes a lightweight backbone extraction network method to reduce the number of backbone layers and the size of backbone parameters, thereby improving the overall efficiency of the segmentation model. An attention mechanism is used, with the first attention mechanism inserted between the input and output of the backbone network and the ASPP, and the second attention mechanism inserted before the first upsampling after the ASPP. To better preserve the underlying image features of the traction rope oil film and simultaneously integrate multi-scale features, a multi-branch cascade ASPP structure is selected. This approach improves the DeepLabV3+ network's detection efficiency and accuracy for the traction rope dataset.

6. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The NVIDIA Jetson Xavier NX image segmentation device is fixed to the development board integrated module by screws, connected to the industrial camera, the power supply lithium battery (24V), the relay module and the voltage conversion module, and mainly runs the relevant detection algorithm.

7. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The oil in the oil storage tank below the traction sheave is transported to the oil nozzles above the traction sheave and the traction rope through the oil supply pump to accurately complete the oiling maintenance task of the traction rope.

8. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The traction rope oil scrapers are located on both sides of the traction rope and are in a separated state when not working. When working, the stepper motor drives the synchronous belt mechanism and the screw nut mechanism to close the oil scrapers on both sides of the traction rope, thereby achieving the purpose of removing excess oil on the surface of the traction rope.

9. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The traction sheave oil scraper is located below the traction sheave and above the oil storage tank. A slight spring force is used to make the traction sheave oil scraper slightly contact the traction sheave, thereby achieving the effect of degreasing the traction sheave.

10. The automatic lubrication system and method for elevator traction ropes based on machine vision according to claim 1, characterized in that: The oil storage tank is located below the traction wheel, with a traction wheel oil scraper installed in the middle and traction rope oil scraper devices on both sides. The inclined plates on both sides are provided with holes for the traction rope to pass through.

Images