This project relies on IoT and AI vision technology to build a special equipment safety monitoring system for China Railway 12th Bureau: it covers the full lifecycle management of tower cranes, gantry cranes, bridge erecting machines, construction elevators, and other equipment, worker control, real-time monitoring and abnormal alarms for lifting weight/ropes/wind speed, moving construction safety management from "post-event tracing" to "real-time prevention and control," and comprehensively strengthening the operational safety defense line.
China Railway 12th Bureau is a large comprehensive construction enterprise under China Railway Construction Corporation, mainly engaged in railway, highway, municipal, building construction, urban rail, and other engineering construction, and undertakes construction tasks for a large number of difficult and key bridge, tunnel, and station building projects. Special equipment such as tower cranes, gantry cranes, bridge erecting machines, crawler cranes, and construction elevators are core equipment at railway and bridge construction sites, and their operational safety is directly related to project progress, human life, and corporate reputation.
For a long time, the safety management of special equipment has generally relied on the traditional model of "manual inspection + operator experience": paper ledgers, equipment status observed by the human eye, violations traced after the fact, and environmental factors such as wind speed difficult to link in real time, so hidden accident hazards are often "discovered only afterward." At the same time, laws and regulations such as the Special Equipment Safety Law and the Provisions on Safety Supervision and Administration of Construction Hoisting Machinery set clear compliance requirements for equipment registration, inspection, and personnel certification, and traditional methods are difficult to implement efficiently.
Under the national policy background of promoting "smart construction sites" and "intelligent construction," China Railway 12th Bureau introduced IoT, AI vision, intelligent sensing, and big data technologies, with Shenzhen Xiangming Technology Co., Ltd. as the contractor, to build a special equipment safety monitoring system covering "full equipment lifecycle + worker control + real-time monitoring of lifting weight/ropes/wind speed + abnormal alarms," realizing the upgrade of safety management from "experience-based judgment" to "data perception."
Before the system was built, special equipment safety management mainly faced six typical pain points:
| No. | Pain Point | Specific Manifestation |
|---|---|---|
| 1 | Equipment ledgers are scattered and the lifecycle is hard to trace | There are many types of equipment such as tower cranes and gantry cranes, widely distributed, and ledgers rely on Excel or paper, unable to run through the entire process of "entry—use—inspection—maintenance—scrapping" |
| 2 | No one reminds when inspection and testing expire | Regular inspection plans rely on manual memory, missed inspections and overdue inspections are prone to regulatory penalties, and compliance risk is high |
| 3 | Personnel qualifications are hard to verify as genuine or fake | Certificates of special operations personnel are scattered, mismatches between people and certificates and expired certificates occur from time to time, and unauthorized machine operation is hard to intercept in real time |
| 4 | Key parameters rely on manual observation | Key parameters such as lifting weight, rope wear, height, and inclination depend on the driver's naked-eye judgment, and hidden hazards are discovered with serious delay |
| 5 | Environmental factors such as wind speed lack linkage | Wind speed has a huge impact on high-altitude equipment such as tower cranes, yet there is a lack of real-time monitoring and automatic braking when thresholds are exceeded |
| 6 | Accident evidence collection is difficult | There is a lack of black boxes, operation records, and video linkage, making it difficult to determine responsibility after an accident occurs |
Centering on the four dimensions of "equipment—personnel—environment—operation," the system establishes four major construction goals:
Digitalization of the full equipment lifecycle.Realize closed-loop management of special equipment from entry registration, inspection and testing, maintenance, to scrapping and deregistration, with the entire data process traceable.
Personnel qualifications and operating behavior controllable.Realize dynamic verification of personnel qualifications, face verification before machine operation, and real-time identification of violations.
Real-time monitoring and intelligent alarms for key parameters.Conduct 7×24h real-time monitoring of core indicators such as lifting weight, ropes, wind speed, and operating parameters, with multi-terminal alarms when thresholds are exceeded and linkage braking when necessary.
Visualized control and regulatory compliance.Achieve one-screen overview through the safety monitoring large screen, and connect to the government regulatory platform to meet compliance requirements.
The system is organized according to the four major links of "full equipment lifecycle—worker control—real-time safety monitoring—intelligent alarms and visualization," with a total of 14 core functional modules.
(1) Full Equipment Lifecycle Management
1. Special Equipment Archives and Type Management
What it is:Establish a unified equipment ledger covering multiple types such as tower cranes, gantry cranes, bridge erecting machines, crawler cranes, and construction elevators, linked to factory serial numbers, technical parameters, affiliated projects, and current locations.
What problem it solves:Solve the problems of many equipment types, wide distribution, and scattered ledgers that are difficult to unify, and provide a unique data foundation for subsequent inspection, maintenance, and transfer.
Target audience:Project department equipment administrators, group company asset administrators.
2. Equipment Registration and Inspection & Testing Management
What it is:Aligned with the requirements of the "Special Equipment Safety Law", it records key milestones such as equipment registration numbers, installation notifications, use registration, first inspection, and periodic inspections, with automatic reminders as deadlines approach.
What problem it solves:Completely eliminates relying on memory for inspection dates, prevents missed inspections and overdue inspections, and reduces the risk of regulatory penalties.
Target audience:Safety directors, equipment administrators, regulatory inspection declarants.
3. Maintenance and Servicing Management
What it is:Automatically generates maintenance plans based on equipment type, operating conditions, and working hours; maintenance processes are recorded by scanning codes and uploading photos, and key components can be linked to spare parts consumption.
What problem it solves:Shifts from "repair after failure" to "scheduled maintenance", significantly reducing unexpected downtime.
Target audience:Equipment maintenance teams, workshop directors.
4. Equipment Scrapping Management
What it is:For equipment that has reached its service life or is no longer assessed as meeting safety requirements, it supports a full closed-loop process for scrapping applications, technical appraisal, and deregistration.
What problem it solves:Avoids equipment operating beyond its service life while in poor condition, and clarifies the asset disposal process.
Target audience:Equipment administrators, asset management department.
(II) Operator Control
5. Special Operations Personnel Qualification Management
What it is:Uniformly enters certificate information for operators, commanders, riggers, and other personnel, supports certificate images, validity periods, and review records, with automatic yellow/red warnings upon expiration.
What problem it solves:Prevents unlicensed or expired personnel from operating machines, and intercepts compliance risks of "person-certificate mismatch" in advance.
Target audience:Safety directors, human resources, project managers.
6. Pre-operation Facial Identity Verification
What it is:Before the driver ascends the tower or personnel lift, facial recognition compares the prestored photo and qualification information in the system, and only after verification is passed is the equipment allowed to start.
What problem it solves:Solves "borrowed certificate operation" and "substitute operation", ensuring that each piece of equipment corresponds to a specific responsible person.
Target audience:Operating drivers, project department safety officers.
7. Identification of Illegal Operation Behavior
What it is:Using AI vision models to recognize behaviors such as smoking, phone calls, fatigue, and leaving the post in the cockpit in real time; when abnormal, local sound and light alarms are triggered and pushed to the backend.
What problem it solves:Extend safety supervision to the "work behavior itself" and eliminate hidden dangers in the bud.
Target audience:Drivers, safety officers, project managers.
(III) Real-time safety monitoring
8. Monitoring of lifting weight and hoisting mechanism
What it is:Real-time collection of the weight of the lifted object through tension/weight sensors; warning when approaching 90% of the rated load, and automatic locking and alarm when reaching 100%.
What problem it solves:Eliminate tower collapse and rope break accidents caused by overloaded lifting from the source.
Target audience:Operating drivers, construction workers, safety officers.
9. Monitoring of wire rope/sling status
What it is:Through means such as rope diameter wear sensors, broken wire detection, and vibration spectrum analysis, determine the service status and remaining life of the wire rope.
What problem it solves:Turn "seeing whether a wire rope still looks okay" into "using data to determine whether it can continue to be used."
Target audience:Equipment maintenance team, safety officers.
10. Wind speed and meteorological monitoring
What it is:A high-precision anemometer is integrated at the top of the tower crane to collect wind speed, wind direction, and temperature in real time; when approaching the rated wind speed, it automatically issues sound and light alarms, and when the threshold is reached, it links to equipment locking.
What problem it solves:Turn the uncertainty of "being afraid when the wind gets strong" into automatic response based on evidence.
Target audience:Drivers, on-site dispatchers, safety directors.
11. Comprehensive monitoring of operating parameters
What it is:High-frequency collection of operating parameters such as height, slewing, amplitude, inclination, and speed, with automatic alarms for abnormal postures.
What problem it solves:Continuously record the "posture data" that drivers cannot see, helping operations become smoother and more controllable.
Target audience:Operating drivers, safety officers.
(IV) Intelligent alarms and visualization
12. Multi-terminal alarms and closed-loop handling
What it is:Alarms are distributed by severity level to multiple terminals such as cab sound and light, team leader APP, dispatch large screen, and supervision platform, with the entire process of alarm response, handling, and closure documented.
What problem it solves:Ensure every alarm has a "recipient," "handler," and "handling result," avoiding alarms becoming mere formality.
Target audience:Team leaders, safety officers, dispatch supervisors.
13. Video AI and Black Box
What it is:Video streams are deployed at key points such as the cockpit, hook, and tower arm. AI identifies illegal operations and obstacles in real time, and all data is stored in the cloud as a "construction black box."
What problem it solves:It not only provides a data source for AI algorithms, but also provides a tamper-proof evidence chain for accident reconstruction and liability determination.
Target audience:Safety officers, project departments, insurance companies, regulatory authorities.
14. Safety Monitoring Large Screen and Regulatory Integration
What it is:A 3D GIS + digital twin large screen that displays project distribution, equipment status, personnel location, and alarm events in real time; supports data reporting to government smart construction site/special equipment regulatory platforms.
What problem it solves:Project departments, groups, and regulators get a "single-screen overview," enabling both decision-making and compliant reporting.
Target audience:Project managers, group leaders, regulatory authorities.
The system adopts a "cloud-edge-device" four-layer architecture, balancing on-site real-time performance and platform scalability:
Perception layer (terminals):IoT terminals such as anemometers, weighing sensors, inclinometers, vibration collectors, facial recognition terminals, and AI cameras are responsible for raw data collection.
Edge layer (on-site):Black boxes for tower cranes/gantry cranes/construction elevators and industrial control gateways perform local alarm judgment and braking linkage, and can "stand guard independently" even when the network fluctuates.
Platform layer (cloud):The data middle platform, AI middle platform, and business middle platform based on the industrial internet platform provide unified storage, computing, model inference, and external data services.
Application display layer:PC management backend, mobile APP, digital twin large screen, and regulatory integration interfaces, respectively serving different roles.
Unified access for multiple types of equipment.Heterogeneous equipment such as tower cranes, gantry cranes, bridge erecting machines, crawler cranes, and construction elevators are connected based on a unified protocol, and the cost of adding new types later is low.
"Equipment-personnel-environment-operation" four-dimensional integration.It is not just equipment operation data, but incorporates personnel qualifications, environmental meteorology, and work behavior into the same risk model.
IoT + AI dual protection.Sensors tell "what it is," and AI vision tells "what it is doing." The two cross-validate each other, significantly reducing the false alarm rate.
Alarms can be linked to equipment locking.Unlike traditional monitoring that "only warns but does not act," the system can directly cut off dangerous actions when key thresholds are triggered, "nipping risks in the bud."
Closed loop for regulatory compliance.Inspection records, personnel qualifications, and alarm events are fully traceable, and data can be reported with one click, reducing corporate compliance costs.
After the system was launched and operated in multiple key engineering projects of China Railway 12th Bureau Group, it brought significant improvements in the three dimensions of safety, compliance, and efficiency:
The accuracy of identifying illegal operations in key work scenarios has improved significantly, and risk events such as unauthorized machine operation and mismatch between person and certificate have dropped substantially.
High-risk events such as overload and wind speed exceeding limits achieve "real-time perception—automatic locking—post-event traceability", and major safety hazards are significantly reduced.
Compliance matters such as equipment inspection expiration and personnel certificate expiration are automatically reminded by the system, and human missed inspections and overdue phenomena are basically eliminated.
After an accident occurs, the operation process is quickly reconstructed through the "black box + video AI", responsibility determination is clearer, and dispute handling cycles are significantly shortened.
The group / project department / regulator achieve multi-level collaboration through the same digital twin large screen, and management communication costs drop significantly.
The essence is safety.Turn special equipment accidents from "passive response" to "active defense", directly protecting the lives of on-site workers.
Compliance reduces risk.Use digital means to meet the requirements of the Special Equipment Safety Law and other regulations, significantly reducing regulatory penalties and brand reputation risks.
Improve quality and efficiency.The entire process of inspection / maintenance / alarms is digitized, reducing manual inspection workload and human error.
Management upgrade.Accumulate special equipment operation data assets for construction enterprises, which can later connect to larger blueprints such as smart construction sites and smart construction.
Strong replicability.The four-dimensional model of "equipment—personnel—environment—operation" is suitable for multiple types of construction scenarios such as railway, highway, municipal, and building construction.
Special equipment safety management has never been a pile of single-point technologies, but an integrated project of "equipment full life cycle + operating personnel + real-time monitoring + compliance reporting". Shenzhen Xiangming Technology Co., Ltd. will continue to focus on the core concept of "perception + intelligence + linkage", and together with China Railway 12th Bureau and more construction enterprises, bring every construction site, every piece of equipment, and every operator into a perceptible digital defense line, so that "safety" no longer depends on personal experience, but is guarded by the system.
——Shenzhen Xiangming Technology Co., Ltd. · Case Library
Q1: What is the China Railway 12th Bureau special equipment safety monitoring system?
A: This is a full-element safety management and control platform for special equipment relying on IoT Internet of Things and AI vision technology. It is aimed at engineering construction enterprises such as railway, highway, municipal, and building construction, and covers the full life cycle and real-time safety monitoring of multiple types of equipment such as tower cranes, gantry cranes, bridge erecting machines, crawler cranes, and construction elevators.
Q2: What core functional modules does the system include?
A: There are 14 core modules in total, organized by four major links: equipment full life cycle (archives and types, registration and inspection, maintenance, scrapping), operating personnel control (qualifications, face verification, violation identification), real-time safety monitoring (lifting weight, ropes, wind speed, operating parameters), intelligent alarms and visualization (multi-terminal alarms, video AI and black box, monitoring large screen and regulatory connection).
Q3: What problems do personnel qualifications and face verification solve?
A: The system uniformly files certificates for operators, commanders, riggers and other operating personnel and automatically warns before expiration; drivers must pass face recognition comparison before climbing the tower, and only those with valid qualifications can start the equipment, thereby eliminating violations such as "borrowing certificates to operate", "substitute operation", and "operating without a certificate".
Q4: How do real-time monitoring functions such as wind speed and lifting weight work?
A: A high-precision anemometer is integrated at the top of the tower crane, the hoisting mechanism is equipped with a weighing sensor, the wire rope is deployed with wear and vibration monitoring, and parameters such as height / inclination / slewing are continuously collected; when approaching the rated threshold, the system warns in advance, and when reaching the dangerous threshold, it can automatically trigger equipment locking, moving "risk discovery" from after the event to real time.
Q5: Besides reminders, what else can intelligent alarms do?
A: Alarms are distributed by severity level to multiple terminals such as sound and light in the driver's cab, the team leader APP, the dispatch large screen, and the regulatory platform. Key alarms can directly link to equipment locking; the entire process of receiving, handling, and closing alarms is traceable, and each alarm has a "handler" and "handling result", avoiding alarms becoming mere formality.
Q6: What technical architecture does the system adopt?
A: It adopts a four-layer "cloud—edge—device" architecture: the perception layer consists of terminals such as anemometers, weighing, inclination, face recognition, and AI cameras; the edge layer consists of tower crane / elevator black boxes and industrial control gateways, with local alarm judgment and braking linkage, operating independently under weak network conditions; the platform layer consists of a data middle platform, AI middle platform, and business middle platform; the application layer provides a PC management backend, mobile APP, digital twin large screen, and regulatory connection interfaces.
Q7: Which enterprises or projects are suitable for deploying this system?
A: It is suitable for construction enterprises that need to use special equipment such as tower cranes, gantry cranes, bridge erecting machines, crawler cranes, and construction elevators on a large scale, covering engineering scenarios such as railways, highways, municipal works, building construction, urban rail transit, bridges, and tunnels. It is especially suitable for owners and general contractors with multiple projects running in parallel, high regulatory requirements, and the need to "preemptively intercept" safety risks.
Room 1409, Building C, Shenzhen Century Industrial Center, No. 743 Zhoushi Road, Hezhou Community, Hangcheng Sub-district, Bao'an District, Shenzhen
