Double-crane lifting dynamic balance control method based on real-time load feedback

Abstract

The invention provides a double-machine lifting dynamic balance control method based on real-time load feedback, which comprises the following steps of: after a system is initialized, synchronously acquiring load, height, attitude, position, motion and environment data of two machines, and fusing and filtering; load, height, angle, position and motion coordination multiple deviations are calculated, and balance and safety are evaluated; generating a speed, position, attitude and safety comprehensive control strategy by using multi-objective optimization according to an evaluation result, and setting a priority and a look-ahead mechanism; executing instructions in a distributed manner and monitoring the instructions in real time; self-adaptive parameter adjustment and machine learning continuous optimization are carried out based on a feedback closed loop; and multi-stage safety monitoring and emergency processing are synchronized. According to the invention, the precision, safety and efficiency of hoisting operation can be improved, the operation difficulty and energy consumption are reduced, and the adaptability and reliability of the system in a complex environment are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial automation control, and more particularly, to a double-machine lifting dynamic balance control method based on real-time load feedback. BACKGROUND

[0002] In modern industrial lifting operations, double-machine lifting is a common large equipment lifting method widely used in construction, shipbuilding, petrochemical industry and other fields. Traditional double-machine lifting operation mainly relies on the experience and manual control of operators, and the lifting task is completed by manually observing and coordinating the actions of the two cranes. However, this method has many problems. First, manual operation is difficult to accurately control the synchronization of the two cranes, which can easily lead to load imbalance, increase the risk of shaking and collision during lifting, and thus affect the efficiency and safety of lifting. Second, operator fatigue and human error cannot be ignored, which can lead to serious safety accidents. In addition, the traditional manual control method cannot obtain and process complex environmental information such as wind speed and direction in real time, making it difficult to make effective safety assessments and decisions in complex environments.

[0003] With the continuous development of sensor technology, automation control technology and computer technology, some improved double-machine lifting control methods have been proposed. For example, sensors are installed on the cranes to monitor parameters such as load weight, lifting height and motion state, achieving a certain degree of automation control. However, most of these methods can only achieve local automation control and cannot effectively fuse and analyze multi-source data, making it difficult to fully assess the load balance state and safety during lifting. Moreover, existing control methods lack adaptive optimization mechanisms and cannot dynamically adjust control strategies based on real-time feedback data to respond to complex lifting environments and task requirements.

[0004] In the implementation of the embodiments of the present application, at least the following problems or defects exist in the prior art: the prior art cannot achieve comprehensive, real-time and dynamic balance control of the double-machine lifting process, and it is difficult to meet the high-precision and high-safety requirements of lifting operations in complex industrial environments. SUMMARY

[0005] The present application provides a double-machine lifting dynamic balance control method based on real-time load feedback, comprising: Step S100, system initialization and parameter presetting stage, based on the preset system control requirements, set the core control parameters including load deviation threshold, proportional coefficient, integral coefficient and differential coefficient, and configure the operation parameters such as system sampling frequency, safety angle threshold and height compensation coefficient, and complete the startup and self-checking program of all sensor systems; Step S200, multi-source data synchronous acquisition and fusion processing stage, based on the cooperative work of the multi-sensor system, synchronously acquiring real-time load data, lifting height data, spatial attitude angle data, position coordinate data, motion state data and environmental parameter data of the double cranes, and performing time synchronization, coordinate unification and filtering processing on the collected multi-source heterogeneous component data through data fusion technology; Step S300, load state multi-computation and analysis stage, based on the multi-dimensional data after fusion processing, calculating real-time load deviation value, height coordination deviation value, spatial angle deviation value, position synchronization deviation value and motion coordination degree, etc. multiple state parameters, and using multi-condition judgment algorithm for load balance state evaluation and safety analysis; Step S400, intelligent decision and control instruction generation stage, based on the state evaluation result, using multi-objective optimization algorithm to generate comprehensive control strategy including speed adjustment instruction, position correction instruction, attitude adjustment instruction and safety control instruction, and establishing instruction priority mechanism and forward-looking control mechanism; Step S500, multi-system cooperative execution and real-time monitoring stage, based on the generated control instruction set, coordinating each execution subsystem to complete parameter adjustment task through distributed control architecture, and real-time monitoring instruction execution state and execution effect, ensuring the coordination and safety of system operation; Step S600, dynamic feedback and adaptive optimization stage, based on the feedback data collected in the execution process, using closed-loop control strategy to adjust control parameters in real time, and continuously optimizing system performance through machine learning algorithm, establishing adaptive control mechanism; Step S700, safety monitoring and emergency handling stage, based on real-time safety monitoring data, identifying system abnormal state through multi-level safety monitoring system, and starting corresponding emergency handling mechanism according to preset emergency handling plan.

[0006] Further, the step S200 comprises: Step S210, based on high-precision weight sensor system, synchronously acquiring real-time load data of the first crane and the second crane at a preset sampling frequency, ensuring the synchronization and accuracy of data acquisition; Step S220, based on high-resolution encoder system, acquiring real-time lifting height data of the first crane and the second crane, and eliminating measurement noise by using data filtering technology; Step S230, based on the inclination sensor system, accurately measuring the angle data of the hoisting rigging and the vertical direction, providing data support for spatial attitude analysis; Step S240, based on GPS positioning module and laser ranging system, acquiring real-time spatial position coordinate data of the two cranes, and establishing a unified spatial coordinate system; Step S250, based on the motion sensor system, collect the running speed data and acceleration data of the crane, and provide data basis for motion coordination analysis; Step S260, based on the environmental monitoring sensor, real-time collect environmental parameter data such as wind speed, wind direction and environmental temperature, and provide environmental factor data for safety evaluation; Step S270, based on the time synchronization system, perform timestamp alignment processing on the multi-source heterogeneous component data to ensure the time consistency of the data; Step S280, based on multi-level filtering algorithm and data fusion technology, perform noise elimination, outlier processing and data fusion on the collected original data to improve data quality and reliability.

[0007] Further, the step S300 includes: Step S310, based on real-time load data, calculate the load deviation value of the two cranes by difference calculation, and smooth the data fluctuation by using sliding average filtering technology; Step S320, based on the lifting height data, calculate the height coordination deviation value, and eliminate the influence of height difference on load balance through height compensation algorithm; Step S330, based on the spatial attitude angle data, calculate the angle deviation value, and convert the angle deviation to spatial position deviation by using coordinate transformation technology; Step S340, based on the spatial position coordinate data, calculate the position synchronization deviation value by relative position, and establish the mapping relationship between position deviation and motion control; Step S350, based on the running speed and acceleration data, calculate the motion coordination degree through kinematic analysis to evaluate the motion synchronization performance of the two cranes; Step S360, based on multiple state parameters, use fuzzy logic algorithm to judge the load balance state, and generate system state evaluation report; Step S370, based on environmental parameters and system state parameters, identify safety hazards through risk assessment model, and calculate system safety level; Step S380, based on the comprehensive evaluation result, generate state evaluation report containing state rating and adjustment priority, and provide basis for control decision.

[0008] Further, the step S300 also includes: Step S390, based on the weighted fusion algorithm, fuse multiple state parameters into a comprehensive load deviation index, which comprehensively considers the mutual influence of load deviation, height deviation, angle deviation and position deviation; Step S391, based on the state parameters, construct a multi-dimensional state evaluation matrix, and use principal component analysis method to extract key state features; Step S392, based on fuzzy logic algorithm, the system state is divided into multiple levels, and the corresponding relationship between state level and control strategy is established; Step S393, based on time series analysis method, the trend of historical operation data is predicted, and the state change trend in the future time period is predicted; Step S394, based on probability statistics method, the risk coefficient of the system at present is calculated, and the corresponding control strategy is made for different risk levels; Step S395, based on Markov model, state transition probability matrix is established, transition possibility between states is predicted, and basis for preventive control is provided.

[0009] Further, the step S400 includes: Step S410, based on state rating results, decision tree algorithm is used to select the most suitable control strategy generation algorithm; Step S420, based on multi-objective optimization algorithm, multiple objectives such as load balancing, motion stability and energy consumption optimization are considered comprehensively, and the optimal speed regulation instruction is generated; Step S430, based on position synchronization deviation value, PID control algorithm is used to generate position correction instruction, and accurate position control is realized; Step S440, based on angle deviation value, space geometric transformation is used to generate attitude adjustment instruction, and spatial symmetry of hoisting system is ensured; Step S450, based on risk coefficient evaluation results, hierarchical safety control instruction is generated, including multiple levels of safety measures such as early warning prompt, speed limit control and emergency shutdown; Step S460, based on cooperative control algorithm, cooperative control instruction set is generated to ensure that the control instructions of two cranes match each other; Step S470, based on real-time state of the system, adaptive algorithm is used to generate parameter adjustment instruction, and control parameters are optimized in real time.

[0010] Further, the step S400 further includes: Step S480, based on state emergency degree evaluation, priority mechanism of control instruction is established, and important instructions are executed preferentially; Step S481, based on model predictive control algorithm, prospective control instruction is generated according to state trend prediction results, and system response speed is improved; Step S482, based on fault-tolerant control algorithm, when sensor failure or data anomaly is detected, reliable fault handling instruction is generated by using state estimation value; Step S483, based on energy consumption optimization algorithm, energy saving optimization instruction is generated under the premise of ensuring system safety, and system operation energy consumption is reduced; Step S484, based on the input of the human-computer interaction system, generate auxiliary decision-making instructions to provide control suggestions for the operator; Step S485, based on the security verification algorithm, generate verification code for each control instruction to ensure the integrity and security of the instruction.

[0011] Further, the step S500 includes: Step S510, based on the scheduling algorithm of the instruction distribution system, accurately send the speed adjustment instruction to the corresponding variable frequency control system; Step S520, based on the high-precision positioning function of the servo control system, execute the position correction instruction to ensure the precise positioning of the walking mechanism; Step S530, based on the coordinated control ability of the multi-axis controller, execute the posture adjustment instruction to realize the smooth movement of the boom; Step S540, based on the fast response characteristics of the safety protection system, execute the safety control instruction to ensure the safe operation of the system; Step S550, based on the real-time data acquisition ability of the distributed sensor network, collect the feedback data of each execution mechanism; Step S560, based on the real-time analysis function of the state monitoring system, detect abnormal conditions in the instruction execution process; Step S570, based on the multi-index comprehensive evaluation function of the effect evaluation algorithm, calculate the success rate and effect index of the instruction execution; Step S580, based on the quantitative scoring standard of the quality evaluation system, objectively score the execution effect of each control instruction.

[0012] Further, the step S500 also includes: Step S590, based on the communication protocol of the distributed control architecture, establish a real-time data exchange mechanism between each subsystem; Step S591, based on the verification logic of the collaborative verification algorithm, ensure the coordinated action of each execution mechanism; Step S592, based on the recording function of the instruction tracing system, record the execution process and timestamp information of each instruction; Step S593, based on the real-time monitoring function of the safety monitoring system, prevent the occurrence of instruction conflict and execution anomaly; Step S594, based on the scheduling strategy of the efficiency optimization algorithm, reduce the time delay in the instruction transmission and execution process; Step S595, based on the backup scheme of the fault-tolerant control mechanism, automatically start the backup control scheme when part of the execution mechanism fails.

[0013] Further, the step S600 includes: Step S610, based on the distributed architecture of the multi-dimensional feedback data collection system, real-time collection of feedback information of each execution link; Step S620, based on the parameter adjustment strategy of the adaptive control algorithm, real-time optimization of the operating parameters of the control system; Step S630, based on the forward-looking feature of the model predictive control algorithm, adjusting the control strategy according to the dynamic characteristics of the system; Step S640, based on the data analysis capability of the machine learning algorithm, establishing a system performance optimization model and continuously improving it; Step S650, based on the adaptive characteristics of the online parameter setting algorithm, automatically adjusting the proportional, integral and differential parameters of the controller; Step S660, based on the comprehensive evaluation function of the multi-dimensional evaluation index system, regularly evaluating the control effect of the system; Step S670, based on the search ability of the optimization algorithm, continuously improving the control strategy to improve the system performance; Step S680, based on the storage and retrieval function of the knowledge base system, saving the optimized control parameters and strategy scheme.

[0014] Further, the step S700 includes: Step S710, based on the distributed architecture of the multi-level safety monitoring system, real-time monitoring of the safety state of each link of the system; Step S720, based on the prediction function of the risk assessment model, generating an early warning signal in time when potential risks are detected; Step S730, based on the inference mechanism of the decision tree model, automatically selecting the optimal processing scheme according to the abnormal type; Step S740, based on the emergency handling process of the hierarchical response mechanism, starting the corresponding emergency program according to different emergency levels; Step S750, based on the protection function of the safety interlocking mechanism, ensuring the reliability and safety of the emergency handling process; Step S760, based on the statistical analysis of the emergency exercise results, regularly updating and optimizing the emergency plan; Step S770, based on the data collection function of the safety event record system, recording all relevant information of safety events; Step S780, based on the evaluation criteria of the continuous improvement mechanism, regularly evaluating and optimizing the safety measures.

[0015] The above embodiments according to the present application have at least the following beneficial effects: 1. Through multi-source data synchronous acquisition and fusion processing, the present application can obtain real-time multi-dimensional data such as load, height, attitude, position, motion state and environmental parameters during the lifting process of the double machine, and perform time synchronization, coordinate unification and filtering processing. This not only ensures the accuracy and reliability of the data, but also provides comprehensive data support for subsequent load balance state evaluation and safety analysis, solving the problem of inaccurate load balance control and insufficient safety evaluation caused by inaccurate or incomplete data in traditional methods.

[0016] 2. Load balance state evaluation is performed using multi-state parameter calculation and fuzzy logic algorithm, the present application can comprehensively consider load deviation, height deviation, angle deviation and position deviation, and accurately judge and classify the system state. At the same time, based on risk assessment model and probability statistical method, it can identify safety hazards in time and develop corresponding control strategy. This makes the load balance control in the lifting process more precise and intelligent, improves the safety and reliability of the lifting operation, effectively avoids the accident risk caused by load imbalance or sudden safety problems.

[0017] 3. The present application generates comprehensive control strategy through multi-objective optimization algorithm, and establishes instruction priority mechanism and forward-looking control mechanism, which can quickly generate optimal control instruction according to real-time state evaluation result. At the same time, based on dynamic feedback and adaptive optimization mechanism, the system can adjust control parameters and continuously optimize control strategy in real time to adapt to complex lifting environment and task requirements. This not only improves the efficiency and precision of lifting operation, but also reduces energy consumption, enhances the adaptability and stability of the system, effectively solves the problems of low control precision, high energy consumption and difficulty in adapting to complex environment in traditional manual control and simple automatic control methods. BRIEF DESCRIPTION OF DRAWINGS

[0018] The above and other objects, features and advantages of the exemplary embodiments of the present application will be more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which: Figure 1 The flowchart of the double-machine lifting dynamic balance control method based on real-time load feedback provided by an embodiment of the present application. DETAILED DESCRIPTION

[0019] The technical solutions in the present application will be described clearly and completely in the present application combined with the drawings in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. The components of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application. It should be noted that: similar numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0020] In the traditional double-machine lifting operation, the problem of insufficient load balancing control accuracy caused by multi-source heterogeneous data fusion failure is significant. Due to the lack of cooperative processing mechanism for real-time load data, spatial attitude data, position coordinate data and environmental parameter data, the system cannot accurately construct a multi-dimensional state evaluation model, resulting in the inability to trigger dynamic compensation in time when the load deviation threshold is broken. This problem directly leads to the mismatch of double-machine motion trajectory, the stress of lifting rigging exceeding the limit and the decrease of system safety margin, seriously affecting the stability and reliability of the lifting operation.

[0021] For example, in the scene of lifting large offshore platform modules, two cranes are respectively equipped with weight sensors, inclination sensors and GPS positioning modules, but each sensor system operates independently and the sampling frequencies are inconsistent. When the wind speed suddenly changes and causes the spatial angle of the lifting rigging to change, due to the lack of time synchronization mechanism and coordinate unification model, the system cannot accurately calculate the coupling relationship between angle deviation and load deviation. At this time, the high coordination deviation value continues to accumulate, but the traditional control strategy lacks multi-objective optimization capability and can only compensate through single parameter adjustment, resulting in the position synchronization deviation value exceeding the safety threshold, ultimately causing the lifting system to oscillate.

[0022] If the above problems are not solved, the double-machine lifting system will be in a sub-stable state for a long time, and uneven load distribution will cause accelerated fatigue damage of key mechanical parts. In a dynamic operating environment, uneliminated measurement noise and data delay will cause control command lag, causing continuous deterioration of motion coordination. More seriously, the system without closed-loop feedback mechanism cannot real-time correct control parameter deviation, when the sensor drift is caused by environmental temperature change, it may directly lead to misjudgment of safety angle threshold, and finally cause catastrophic structural failure.

[0023] In the face of the above problems, the present application first realizes that the root cause of multi-source heterogeneous data fusion failure is the lack of systematic collaborative processing mechanism. In the traditional method, various sensors run independently and the data processing process is fragmented, which leads to the inability to establish a dynamic correlation model between multi-dimensional parameters. In view of this, the present application considers constructing a unified data acquisition and fusion framework, and eliminating the error accumulation caused by data heterogeneity through time synchronization and coordinate unification technology. At the same time, in view of the control lag problem in the dynamic environment, the present application realizes that it is necessary to establish a closed-loop feedback mechanism to combine real-time state evaluation and parameter optimization. Further analysis shows that the existing technology fails to effectively integrate safety monitoring and execution control links, resulting in insufficient emergency response capability. Based on this, the present application proposes to design a whole-process control system covering data acquisition, state analysis, decision making, execution monitoring and safety protection, and to realize the collaborative operation of each link through a distributed architecture.

[0024] As shown in Figure 1 The present application proposes a dual-machine lifting dynamic balance control method based on real-time load feedback, including the following steps: Step S100, system initialization and parameter presetting stage, based on the preset system control requirements, set the core control parameters including load deviation threshold, proportional coefficient, integral coefficient and differential coefficient, and configure the operation parameters such as system sampling frequency, safety angle threshold and height compensation coefficient, and complete the start-up and self-checking program of all sensor systems; Step S200, multi-source data synchronous acquisition and fusion processing stage, based on the collaborative work of multi-sensor system, synchronously acquire real-time load data, lifting height data, spatial attitude angle data, position coordinate data, motion state data and environmental parameter data of dual-crane, and through data fusion technology, carry out time synchronization, coordinate unification and filtering processing on the collected multi-source heterogeneous component data; Step S300, multiple load state calculation and analysis stage, based on the multi-dimensional data after fusion processing, through comprehensive calculation of real-time load deviation value, height coordination deviation value, spatial angle deviation value, position synchronization deviation value and motion coordination degree and other state parameters, and using multiple condition judgment algorithm, carry out load balance state evaluation and safety analysis; Step S400, intelligent decision and control instruction generation stage, based on the state evaluation result, use multi-objective optimization algorithm to generate comprehensive control strategy including speed adjustment instruction, position correction instruction, attitude adjustment instruction and safety control instruction, and establish instruction priority mechanism and forward-looking control mechanism; Step S500, multi-system collaborative execution and real-time monitoring stage, based on the generated control instruction set, through distributed control architecture, coordinate each execution subsystem to complete parameter adjustment task, and real-time monitor instruction execution state and execution effect, ensure the coordination and safety of system operation; Step S600, dynamic feedback and adaptive optimization phase, based on the feedback data collected during the execution process, the closed-loop control strategy is used to adjust the control parameters in real time, and the machine learning algorithm is used to continuously optimize the system performance, and the adaptive control mechanism is established; Step S700, safety monitoring and emergency handling phase, based on real-time safety monitoring data, through a multi-level safety monitoring system to identify system abnormal state, and according to the preset emergency handling plan to start the corresponding emergency handling mechanism.

[0025] The system initialization and parameter presetting stage refers to the process of basic configuration and parameter setting of the control system, which can be realized by using the parameter configuration interface built in the PLC controller or embedded system. The allowable load difference range is defined by presetting the load deviation threshold value. The basic parameters of the PID control algorithm are composed of the proportional coefficient, integral coefficient and differential coefficient. The data acquisition period is determined by the system sampling frequency. The maximum inclination angle of the hoisting equipment is limited by the safety angle threshold value. The height compensation coefficient is used to correct the mechanical error caused by the height difference. The sensor start and self-checking program completes the hardware state detection through the built-in diagnostic module of the sensor. The multi-source data synchronous acquisition and fusion processing stage refers to the process of obtaining multi-dimensional data of the hoisting system through multiple types of sensors. The sensor data synchronous transmission can be realized by using CAN bus or industrial Ethernet. The multi-source data time stamps are aligned through the time synchronization protocol. The space coordinate transformation algorithm is used to convert different measurement datum into a unified coordinate system. The Kalman filter or wavelet denoising algorithm is used for filtering processing to eliminate measurement noise. The load state multiple calculation and analysis stage refers to the process of calculating the dynamic state of the system based on the fusion data. Real-time load deviation value calculation can be realized by using matrix operation and dynamic modeling technology. The height coordination deviation value is obtained by multiplying the height difference and the compensation coefficient. The spatial angle deviation value is obtained by converting the spatial attitude data through the trigonometric function. The position synchronization deviation value is calculated through the relative coordinate difference value. The motion coordination degree is calculated by weighting the speed difference and acceleration difference. The safety evaluation is realized by using the logic threshold comparison and state machine model through the multiple condition judgment algorithm. The intelligent decision and control instruction generation stage refers to the process of generating control strategy according to the state evaluation. Multi-objective optimization can be realized by using genetic algorithm or particle swarm optimization algorithm. The speed adjustment instruction is generated through the PID algorithm. The position correction instruction is realized by using the closed-loop feedback control model. The attitude adjustment instruction is calculated based on the spatial geometric transformation. The instruction priority mechanism adopts the preemptive scheduling algorithm. The forward-looking control mechanism adopts the model predictive control technology. The multi-system collaborative execution and real-time monitoring stage refers to the process of control instruction distribution and execution verification. The instruction parallel execution is realized by using the distributed control system architecture. The subsystem communication is realized by using the OPC UA protocol. The execution state visualization is realized by using the data acquisition and monitoring control system. The coordination verification is completed by comparing the actual data with the motion trajectory simulation. The dynamic feedback and adaptive optimization stage refers to the process of adjusting the control parameters based on the execution effect. The feedback loop is constructed by using the closed-loop control theory. The machine learning algorithm can be used to optimize the parameters by using neural network or support vector machine. The adaptive control mechanism is realized by using the online parameter tuning algorithm. The safety monitoring and emergency handling stage refers to the process of risk identification and fault handling. The redundant detection mechanism is constructed by using the multi-level safety monitoring system. The automatic response is realized by using the pre-programmed logic for emergency handling plan. The safety interlocking mechanism adopts the double protection of hard-wired loop and software logic.

[0026] Specifically, first, system initialization and parameter presetting are performed, core control parameters and operation parameters are set, sensor system startup and self-checking are completed. Then multi-source data synchronous acquisition and fusion processing are performed, various types of data of the double cranes are collected, and the data are subjected to time synchronization, coordinate unification and filtering processing. Next, load state multiple calculation and analysis are performed, multiple state parameters are calculated, load balance state evaluation and safety analysis are performed. Based on the evaluation results, intelligent decision-making and control instruction generation are performed, a comprehensive control strategy is generated using a multi-objective optimization algorithm, an instruction priority mechanism and a forward-looking control mechanism are established. Subsequently, multi-system collaborative execution and real-time monitoring are performed, parameter adjustment tasks are completed by coordinating each execution subsystem through a distributed control architecture, and execution state and effect are monitored in real time. In the execution process, dynamic feedback and adaptive optimization are performed, a closed-loop control strategy is used to adjust control parameters in real time, and machine learning algorithms are used to continuously optimize system performance. Finally, safety monitoring and emergency handling are performed, system abnormal states are identified through a multi-level safety monitoring system, and corresponding emergency handling mechanisms are started.

[0027] Each step works cooperatively to achieve dynamic balance control of the double crane lifting. System initialization lays the foundation for subsequent operations. Multi-source data acquisition and fusion provide accurate input for state evaluation. Load state calculation and analysis achieve comprehensive evaluation of system state. Intelligent decision-making generates optimal control strategies. Multi-system collaborative execution ensures accurate execution of control instructions. Dynamic feedback and adaptive optimization achieve continuous improvement of the system. Safety monitoring and emergency handling ensure safe and reliable operation of the system.

[0028] The application further comprises: based on a high-precision weight sensor system, real-time load data of the first crane and the second crane are synchronously collected at a preset sampling frequency; based on a high-resolution encoder system, real-time lifting height data of the first crane and the second crane are obtained, and data filtering technology is used to eliminate measurement noise; based on an inclination sensor system, the inclination data of the hoisting rigging and the vertical direction are accurately measured; based on a GPS positioning module and a laser ranging system, real-time spatial position coordinate data of the two cranes are obtained, and a unified spatial coordinate system is established; based on a motion sensor system, running speed data and acceleration data of the crane are collected; based on an environmental monitoring sensor, environmental parameter data such as wind speed, wind direction and environmental temperature are collected in real time; based on a time synchronization system, time stamp alignment processing is performed on multi-source heterogeneous component data; based on a multi-level filtering algorithm and data fusion technology, noise elimination, outlier processing and data fusion are performed on the collected raw data.

[0029] The high-precision weight sensor system synchronously collects the force data of the hooks of the two cranes through a preset fixed sampling frequency, thereby eliminating the load calculation error caused by sampling time deviation. The high-resolution encoder system measures the number of rotations of the hoist by using an incremental encoder, and eliminates the measurement noise caused by the wire rope shaking by using a Kalman filtering algorithm. The inclination sensor measures the inclination of the sling by using a three-axis accelerometer, and improves the angle measurement accuracy by using a temperature compensation algorithm. The GPS positioning module and the laser ranging system work jointly, and the position of the crane base and the position of the hook are mapped to the same coordinate system by using a coordinate transformation algorithm. The motion sensor acquires acceleration data of the crane trolley by using a MEMS inertial measurement unit, and eliminates mechanical vibration interference by using a low-pass filter. The environmental monitoring sensor constructs a data acquisition network by using an ultrasonic anemometer and a temperature sensor, and realizes real-time data return by using a wireless transmission protocol. The time synchronization system uses the IEEE 1588 precision clock protocol to control the timestamp error of the sensor data to be in the order of microseconds. The multi-stage filtering algorithm includes a three-stage series structure of sliding average filtering, median filtering and Butterworth filtering, and can eliminate high-frequency noise and pulse interference in stages.

[0030] Specifically, during the data acquisition process, the weight sensor and the encoder realize sampling synchronization through a hardware trigger signal, thereby ensuring the time sequence consistency of the load data and the height data. The inclination sensor data is converted into the horizontal offset of the hook by using a coordinate transformation algorithm, and the absolute position data obtained by the laser ranging system forms a spatial posture composite parameter. The acceleration data acquired by the motion sensor is converted into a speed change by integral operation, and is cross-verified with the height change rate measured by the encoder. The environmental parameter data is integrated into the safety evaluation model by using a weighting algorithm, and the data acquisition frequency is automatically triggered to be improved when the wind speed exceeds a threshold. The time synchronization system sends a global clock signal through the main controller, and forces all sensor nodes to complete data uploading within a fixed time window. The multi-stage filtering algorithm first uses sliding average filtering to eliminate random noise, then uses median filtering to eliminate abnormal data points, and finally uses a Butterworth low-pass filter to suppress high-frequency interference. The data fusion technology uses the D-S evidence theory algorithm to weight and fuse the redundant data from different sensors, thereby generating a standardized data set with unified time reference, spatial coordinates and dimension specification, and providing a high-precision data basis for subsequent load state analysis.

[0031] As a preferred embodiment, the scheme of the application is implemented as follows: In the multi-source data synchronous acquisition and fusion processing stage, first, the real-time load data of the first crane and the second crane is synchronously acquired by a high-precision weight sensor system at a sampling frequency of 100 Hz. Then, the lifting height data of the two cranes is obtained by using a high-resolution encoder system, and the Kalman filter technology is used to eliminate measurement noise. Then, the angle of inclination sensor system is used to measure the angle of deviation of the hoisting rigging from the vertical direction, with an accuracy of 0.1°. At the same time, the real-time spatial position coordinate data of the two cranes is obtained by the GPS positioning module and the laser ranging system, and a unified spatial coordinate system is established with the ground as the reference. In addition, the motion sensor system collects the running speed and acceleration data of the crane, with a sampling frequency of 50 Hz. The environmental monitoring sensor collects real-time parameters such as wind speed, wind direction and environmental temperature, with an update frequency of 1 Hz.

[0032] To ensure the time consistency of the data, a time synchronization system based on IEEE 1588 precision time protocol is used to align the timestamps of the multi-source heterogeneous component data, with a synchronization accuracy of better than 1 ms. Finally, the original data collected is processed by multi-level filtering algorithm and data fusion technology. Specifically, the median filter is used to remove outliers, the wavelet transform is used to remove high-frequency noise, and then the Kalman filter is used for data fusion to improve the data quality and reliability.

[0033] The application further proposes the following steps: based on the real-time load data, the load deviation value of the two cranes is calculated by difference calculation, and the sliding average filtering technology is used to smooth the data fluctuations; based on the lifting height data, the height coordination deviation value is calculated, and the height compensation algorithm is used to eliminate the influence of height difference on load balancing; based on the spatial attitude angle data, the angle deviation value is calculated, and the coordinate transformation technology is used to convert the angle deviation into spatial position deviation; based on the spatial position coordinate data, the position synchronization deviation value is calculated by relative position, and the mapping relationship between position deviation and motion control is established; based on the running speed and acceleration data, the motion coordination degree is calculated by kinematic analysis to evaluate the motion synchronization performance of the two cranes; based on multiple state parameters, the fuzzy logic algorithm is used for load balancing state judgment to generate a system state evaluation report; based on environmental parameters and system state parameters, the risk assessment model is used for safety hazard identification to calculate the system safety level; based on the comprehensive evaluation result, a state evaluation report containing state rating and adjustment priority is generated to provide basis for control decision.

[0034] The moving average filtering technology adopts a moving average algorithm with a time window length of 10 sampling periods to eliminate high-frequency noise interference of the load sensor; the height compensation algorithm establishes a compensation coefficient matrix according to the proportional relationship between the length of the crane boom and the weight of the load, and the compensation coefficient value ranges from 0.8 to 1.2; the coordinate transformation technology decomposes the angle deviation into X / Y / Z axis displacement deviation through three-dimensional space projection, and the projection error is controlled within ±2mm; the position deviation mapping relationship is realized by establishing a position deviation-velocity correction amount table in a two-dimensional coordinate system, and the correction amount step is set to 5mm / s; the motion coordination degree calculation adopts a weighted average algorithm of velocity synchronization error and acceleration matching degree, and the weight coefficients are 0.6 and 0.4 respectively; the fuzzy logic algorithm sets five input parameters and three output levels, and the membership function adopts a triangular distribution; the risk assessment model integrates environmental wind speed, temperature fluctuation and equipment operating parameters, and generates a five-level safety warning through risk probability calculation; the state rating adopts a percentage quantitative scoring mechanism, and the priority adjustment is automatically sorted according to the scoring difference.

[0035] Specifically, when the system enters the load state calculation phase, first, the real-time load data of the two cranes is subjected to 20ms period moving average processing to eliminate ±50kg instantaneous deviation caused by sensor jitter. Subsequently, according to the height sensor data, the preset height compensation coefficient matrix is called to dynamically compensate the load deviation calculation result, for example, when the height difference reaches 300mm, the compensation algorithm is automatically activated. After coordinate transformation, the spatial angle data generates a position deviation of -15mm in the X-axis direction and +8mm in the Y-axis direction, which is cross-verified with the position deviation measured by the GPS positioning system. The motion coordination degree calculation module is updated every 500ms, and a yellow warning is triggered when the speed synchronization error exceeds 0.2m / s. The fuzzy logic system inputs five parameters of load deviation, height deviation, angle deviation, position deviation and motion coordination degree, and outputs three state levels of balance, critical and imbalance after processing by 32 judgment rules in the rule base. The risk assessment model combines the current 3.5m / s wind speed data and equipment operating parameters to calculate a risk coefficient of 0.25, corresponding to a secondary safety warning. The finally generated state evaluation report contains the scores of each subsystem, a comprehensive score of 85 points and a clear indication of the priority of the hydraulic system adjustment, providing data support for subsequent control instruction generation. This scheme integrates multi-dimensional parameters and quantitative evaluation, enabling the system to accurately identify minor state abnormalities and establish an operable priority adjustment mechanism, effectively solving the problems of fuzzy state judgment and insufficient adjustment basis in traditional methods.

[0036] The application further proposes to fuse multiple state parameters into a comprehensive load deviation index based on a weighted fusion algorithm, which comprehensively considers the mutual influence of load deviation, height deviation, angle deviation and position deviation; to construct a multi-dimensional state evaluation matrix based on state parameters, and extract key state characteristics using principal component analysis method; to divide the system state into multiple levels based on fuzzy logic algorithm, and establish the corresponding relationship between state level and control strategy; to predict the trend of state change in the future time period based on time series analysis method; to calculate the current risk coefficient of the system based on probability statistics method, and formulate corresponding control strategies for different risk levels; to establish a state transition probability matrix based on Markov model, and predict the transition possibility between states to provide basis for preventive control.

[0037] Among them, the weighted fusion algorithm uses entropy method to determine the weight coefficient of each deviation parameter, and generates a comprehensive index through linear weighting; the row vector of the multi-dimensional state evaluation matrix corresponds to the time series sampling point, and the column vector contains the normalized state parameters; the principal component analysis method realizes data dimension reduction through eigenvalue decomposition of the covariance matrix; the fuzzy logic algorithm establishes a triangular membership function to divide the state level; the time series analysis uses ARIMA model for trend prediction; the probability statistics method uses Monte Carlo simulation to calculate the risk coefficient; the Markov model constructs a probability transition matrix through historical state transition frequency.

[0038] Specifically, the load deviation, height deviation, angle deviation and position deviation data collected by multiple sensors are input into the weighted fusion module after normalization, and dynamic weights are assigned according to the influence of each parameter on system stability to generate a comprehensive load deviation index. This index, together with real-time motion parameters, constructs a multi-dimensional state evaluation matrix, and the first three principal components reflecting the main dynamic characteristics of the system are extracted through principal component analysis. The fuzzy logic system matches the preset membership function according to the principal component values, and outputs the current state level label. The time series analysis module trains the prediction model based on historical data, and outputs the state change trend curve in the next five seconds. The probability statistics module combines the current state parameters and environmental parameters to calculate the probability value of system imbalance in the next control cycle through random sampling. The Markov model predicts the possible state migration path and probability based on the current state level and historical transition records, and generates preventive control suggestions. The output data of the above modules are input into the decision system together to provide quantitative basis for multi-objective optimization control.

[0039] The application further proposes to select a control strategy generation algorithm based on the state rating results using a decision tree algorithm, generate an optimal speed adjustment instruction based on a multi-objective optimization algorithm, generate a position correction instruction based on a position synchronization deviation value using a PID control algorithm, generate an attitude adjustment instruction by spatial geometric transformation based on an angle deviation value, generate a hierarchical safety control instruction based on risk coefficient evaluation results, generate a cooperative control instruction set based on a cooperative control algorithm, and generate a parameter adjustment instruction based on the real-time state of the system using an adaptive algorithm.

[0040] In the decision tree algorithm, the load balance level, safety level and energy consumption level in the state rating results are used for branch selection, and each branch node corresponds to a different control strategy generation algorithm. The multi-objective optimization algorithm uses a genetic algorithm to perform multi-generation iterative optimization on the speed adjustment instruction, with the load balance weight coefficient set to 0.4-0.6 and the energy consumption optimization weight coefficient set to 0.2-0.3. In the PID control algorithm, the proportional coefficient is set to 0.8-1.2, the integral time is set to 5-8 seconds, and the differential time is set to 0.5-1 second. The spatial geometric transformation uses a homogeneous coordinate transformation matrix to convert the angle deviation into a boom joint angle adjustment amount, with a translation component of 0.05-0.15 meters and a rotation component of 1-3 degrees in the transformation matrix. In the hierarchical safety control instruction, the warning prompt corresponds to a risk coefficient of 0.3-0.5, the speed limit control corresponds to a risk coefficient of 0.5-0.7, and the emergency stop corresponds to a risk coefficient of 0.7 or above. The cooperative control algorithm generates an instruction set that satisfies a synchronization error of less than 2 centimeters by establishing a coupling constraint condition for the kinematic equations of the two cranes. The adaptive algorithm uses a gradient descent method to adjust the control parameters in real time with a learning rate of 0.1-0.3.

[0041] Specifically, when the system state rating is three, the decision tree algorithm selects the multi-objective optimization branch based on the genetic algorithm, simultaneously optimizes the load deviation and energy consumption indicators in the iteration process, and outputs the speed adjustment instruction after 20-30 generations of iteration. The position correction instruction independently adjusts the X / Y / Z three-axis deviation through the PID controller, with the proportional coefficient dynamically adjusted according to the deviation amount, and the proportional coefficient increased to 1.5 times when the deviation exceeds 5 centimeters. The attitude adjustment instruction converts the detected 3-degree angle deviation into a hydraulic cylinder stroke adjustment amount through the homogeneous transformation matrix, with the stroke adjustment accuracy controlled within ±2 millimeters. The safety control module calculates the risk coefficient in real time, and when the wind speed exceeds 8 meters / second and the risk coefficient reaches 0.55, the speed limit control instruction is automatically triggered to reduce the running speed to 70% of the rated value. The cooperative control algorithm solves the motion trajectory equations of the two crane hoisting points to generate a synchronization control instruction that keeps the distance error between the hoisting points less than 1.5 centimeters. The parameter adjustment module collects system response data every 2 seconds and updates the PID control parameters using the adaptive algorithm to ensure that the system response time is always controlled within 0.8 seconds.

[0042] The application further proposes to establish a control instruction priority mechanism based on state emergency degree evaluation, generate prospective control instructions based on a model predictive control algorithm, generate fault handling instructions based on a fault-tolerant control algorithm, generate energy-saving optimization instructions based on an energy consumption optimization algorithm, generate auxiliary decision-making instructions based on a human-machine interaction system, and generate instruction verification codes based on a safety verification algorithm.

[0043] The priority mechanism establishes an execution order queue by evaluating the emergency degree of each control instruction in real time, ensuring that critical instructions have priority in occupying system resources. The model predictive control algorithm predicts the state change trend in the future period by combining historical operation data and current state parameters, and generates control instructions in advance. The fault-tolerant control algorithm estimates the real system state using the Kalman filter algorithm to generate alternative control instructions when detecting abnormal sensor data. The energy consumption optimization algorithm analyzes the power consumption characteristics of each actuator by establishing an energy consumption model, and generates optimization parameters that balance safety and energy saving. The human-machine interaction system displays system state evaluation results in real time through a visual interface and receives adjustment suggestions from operators to generate auxiliary instructions. The safety verification algorithm uses a hash algorithm to generate a unique verification code for each control instruction, and ensures the integrity of the instruction transmission through code comparison.

[0044] Specifically, when the system detects that the load deviation exceeds the threshold, the safety control instruction is automatically promoted to the highest priority, immediately interrupting the execution of non-urgent instructions. At the same time, the model predictive control algorithm generates speed adjustment instructions for the next three seconds in advance based on the crane motion acceleration and wind speed change trend. When the tilt angle sensor experiences a transient fault, the fault-tolerant control algorithm calculates the current angle value based on the previous five seconds of attitude data and generates temporary attitude adjustment instructions. The energy consumption optimization algorithm analyzes the relationship between the frequency converter frequency and energy consumption, and selects the lowest energy consumption frequency parameter under the premise of ensuring lifting accuracy. The operator views the three adjustment schemes recommended by the system through the touch screen and generates the final control instruction after selecting and confirming. Each instruction is attached with an MD5 verification code generated by the current timestamp and instruction content before sending, and the receiving end can execute it only after verification.

[0045] As a preferred embodiment, the scheme of the application is implemented as follows: In the intelligent decision-making and control instruction generation stage, the priority mechanism of control instructions is established based on state emergency degree evaluation to ensure that important instructions are executed first. Specifically, a multi-level priority queue structure is used to divide control instructions into three priority levels according to their emergency degree. High-priority instructions include safety-related emergency stop instructions and load balancing adjustment instructions; medium-priority instructions include position correction and attitude adjustment instructions; low-priority instructions include regular speed adjustment instructions. The instruction execution system processes instructions in each level in order of priority from high to low, ensuring that critical instructions are executed in a timely manner.

[0046] Further, based on the model predictive control algorithm, the forward-looking control instructions are generated according to the state trend prediction results, improving the system response speed. In specific implementation, a dynamic model of the double-machine lifting system is established, including parameters such as load, crane movement, and environmental factors. The Kalman filter is used to predict the system state, and the prediction time window is set to 5 seconds. Based on the prediction results, control instructions such as speed adjustment and position correction are generated in advance to achieve early response to changes in system state.

[0047] Thus, based on the fault-tolerant control algorithm, when a sensor fault or data anomaly is detected, reliable fault handling instructions are generated using state estimation values. Specifically, a redundant sensor configuration is established, and triple redundancy is used for key parameters such as load and position. When a sensor data anomaly is detected, reliable state values are estimated using other normal sensor data and state observers to ensure the accuracy of control instructions.

[0048] Among them, based on the energy consumption optimization algorithm, energy-saving optimization instructions are generated to reduce system operating energy consumption under the premise of ensuring system safety. In specific implementation, an energy consumption model of the double-machine lifting system is established, considering factors such as crane motor power, movement trajectory, and load distribution. The dynamic programming algorithm is used to optimize the movement trajectory and speed curve of the two cranes under the condition of meeting load balance and safety constraints, minimizing total energy consumption.

[0049] Further, based on the input of the human-machine interaction system, auxiliary decision-making instructions are generated to provide control suggestions for the operator. Specifically, a graphical human-machine interface is designed to display key information such as system state, load distribution, and safety assessment results in real time. When the system detects potential risks or situations requiring human intervention, it prompts the operator through the interface with suggestions such as adjusting the lifting speed or changing the relative position of the crane. The operator can confirm or modify the suggestions, and the system generates corresponding control instructions based on the final decision.

[0050] Finally, based on the safety verification algorithm, a verification code is generated for each control instruction to ensure the integrity and safety of the instructions. In specific implementation, a hash algorithm is used to encrypt each control instruction to generate a unique verification code. Before instruction execution, the system automatically verifies the verification code, and only instructions that pass the verification can be executed, effectively preventing instructions from being tampered with or illegally interfered with.

[0051] The application further proposes that the scheduling algorithm based on the instruction distribution system accurately sends the speed adjustment instructions to the corresponding variable frequency control system; the high-precision positioning function based on the servo control system executes the position correction instructions to ensure the accurate positioning of the walking mechanism; the coordinated control ability based on the multi-axis controller executes the attitude adjustment instructions to realize the smooth movement of the boom; the fast response characteristics based on the safety protection system execute the safety control instructions to ensure the safe operation of the system; the real-time data acquisition ability based on the distributed sensor network collects the feedback data of each actuator; the real-time analysis function based on the state monitoring system detects the abnormal conditions in the instruction execution process; the multi-index comprehensive evaluation function based on the effect evaluation algorithm calculates the success rate and effect index of the instruction execution; and the quantitative scoring standard based on the quality evaluation system objectively scores the execution effect of each control instruction.

[0052] Among them, the instruction distribution system adopts a priority queue scheduling algorithm to transmit the speed adjustment instructions according to the equipment number and the instruction type, and the transmission delay is controlled within 5 milliseconds. The servo control system realizes positioning accuracy of ±0.1 millimeter through a closed-loop feedback mechanism, and uses a grating ruler to correct mechanical position deviation in real time. The multi-axis controller coordinates the movement of each joint through a motion trajectory interpolation algorithm to ensure that the attitude adjustment angle error is less than 0.5 degrees. The safety protection system is configured with double redundant processing units, and the abnormal response time is not more than 50 milliseconds. The distributed sensor network adopts a time-triggered acquisition mode, and the data sampling interval is uniformly set to 10 milliseconds. The effect evaluation algorithm establishes a comprehensive evaluation model including execution time deviation, position error and energy consumption index, and the weight coefficients are set to 0.4, 0.4 and 0.2 respectively. The quality evaluation system adopts a percentage scoring standard, and when the execution time deviation exceeds the set threshold, 2 points are deducted for every 1 millisecond over; 5 points are deducted for every 0.1 millimeter over the position error.

[0053] Specifically, during the instruction execution phase, the instruction distribution system first parses the instruction type and device identifier in the control instruction set, matches the communication address of the target device using a hash algorithm, and allocates different types of instructions to dedicated communication channels using time-division multiplexing technology. When the speed adjustment instruction arrives at the frequency converter control system, the servo control system synchronously receives the position correction instruction, achieving precise positioning of the walking mechanism through dual-loop control of the position and speed loops. Upon receiving the attitude adjustment instruction, the multi-axis controller decomposes it into target angles for each joint motor through inverse kinematics calculations and controls the motor movement using trapezoidal acceleration / deceleration curves. The safety protection system monitors the current and temperature parameters of each actuator in real time; when motor overload is detected, a safety control instruction is immediately triggered to cut off power output. The distributed sensor network periodically collects the actual position, speed, and current data of each actuator via the CAN bus. The status monitoring system compares the difference between the instruction target value and the actual value; when the position deviation exceeds the limit for three consecutive sampling cycles, an abnormal alarm is triggered. After each instruction execution, the performance evaluation algorithm extracts execution time, positioning accuracy, and energy consumption data and inputs them into the evaluation model to generate quantitative evaluation results. The quality evaluation system generates an execution quality report based on the evaluation results, and automatically triggers the instruction resending mechanism when the score is below 80.

[0054] As a preferred embodiment, the solution of this application is specifically implemented as follows: Based on the scheduling algorithm of the command distribution system, speed adjustment commands are accurately sent to the corresponding frequency converter control systems. Specifically, the command distribution system employs a priority-based task scheduling algorithm, sorting commands according to their urgency and importance to ensure that critical commands are transmitted to the frequency converter control systems first. Upon receiving the speed adjustment command, the frequency converter control system precisely controls the motor speed by adjusting the output frequency, thereby achieving precise speed regulation of the crane.

[0055] Furthermore, based on the high-precision positioning function of the servo control system, position correction commands are executed to ensure the accurate positioning of the traveling mechanism. The servo control system adopts a closed-loop control strategy, using a high-resolution encoder to provide real-time position information, combined with a PID control algorithm, to achieve sub-millimeter-level precise positioning of the traveling mechanism. Position correction commands are transmitted to the servo driver via the CAN bus, and the driver adjusts the speed and direction of the servo motor according to the commands, thereby achieving precise position adjustment of the crane.

[0056] Therefore, based on the coordinated control capabilities of the multi-axis controller, attitude adjustment commands are executed to achieve smooth boom movement. The multi-axis controller uses an interpolation algorithm to decompose the attitude adjustment commands into motion commands for each axis, ensuring synchronous movement of multiple axes and avoiding boom shaking or instability. By calculating the velocity and acceleration curves of each axis in real time, the controller achieves smooth boom start and stop, reducing impact on the load.

[0057] wherein, based on the fast response characteristics of the safety protection system, safety control instructions are executed to ensure safe operation of the system. The safety protection system adopts a redundant design, including a main controller and an independent safety monitor. Safety control instructions are transmitted through an independent safety communication network to ensure the reliability and real-time performance of the instructions. The system sets multiple safety thresholds, and when it detects a situation that exceeds the safety range, it immediately triggers the corresponding safety measures, such as deceleration, shutdown, or emergency braking.

[0058] Specifically, based on the real-time data collection capability of the distributed sensor network, feedback data from each actuator is collected. The sensor network adopts a star topology, including load sensors, angle sensors, position sensors, etc., and realizes high-speed data transmission through industrial Ethernet. The data collection system uses high sampling rate analog-to-digital converters to ensure the real-time performance and accuracy of the data.

[0059] For example, based on the real-time analysis function of the state monitoring system, abnormal situations during instruction execution are detected. The state monitoring system uses a rule-based expert system combined with fuzzy logic algorithms to perform real-time analysis and judgment on the collected data. The system predefines characteristic parameters of multiple abnormal patterns, quickly identifies potential abnormal situations through pattern matching, and generates corresponding alarm information.

[0060] Further, based on the multi-index comprehensive evaluation function of the effect evaluation algorithm, the success rate and effect indicators of instruction execution are calculated. The effect evaluation algorithm uses a weighted scoring method, considering multiple dimensions such as the accuracy, timeliness, and stability of instruction execution. The system compares the expected effect of the instruction with the actual execution result, calculates the execution success rate, and generates a detailed evaluation report.

[0061] As a preferred embodiment, based on the quantitative scoring standard of the quality evaluation system, the execution effect of each control instruction is objectively scored. The quality evaluation system uses a standardized scoring template, including instruction response time, execution accuracy, system stability, and other evaluation indicators. The scoring results are presented through a visual interface, making it easy for operators and managers to understand system performance in a timely manner and providing a basis for subsequent optimization.

[0062] The application further proposes a communication protocol based on a distributed control architecture to establish real-time data exchange mechanisms between subsystems; a verification logic based on a collaborative verification algorithm to ensure the coordinated action of each actuator; a recording function based on an instruction tracing system to record the execution process and timestamp information of each instruction; a real-time monitoring function based on a safety monitoring system to prevent instruction conflicts and execution abnormalities; a scheduling strategy based on an efficiency optimization algorithm to reduce time delays in instruction transmission and execution; and a backup scheme based on a fault-tolerant control mechanism to automatically start a backup control scheme when some actuators fail.

[0063] The real-time data exchange mechanism ensures the transmission of data between subsystems at a fixed cycle by defining a standardized communication protocol format and transmission rules, and the transmission cycle can be set to 10-50 milliseconds; the cooperative verification algorithm triggers a secondary verification process when the deviation exceeds the set threshold by comparing the feedback data of each actuator with the matching degree of the expected action trajectory; the instruction tracing system stores instruction execution logs in a distributed database, and each instruction record contains instruction content, execution timestamp and actuator identifier; the safety monitoring system detects whether there are mutually exclusive instructions or resource competition by real-time scanning of the instruction queue, and the detection cycle can be set to 5-20 milliseconds; the efficiency optimization algorithm adopts a dynamic priority scheduling strategy, and dynamically adjusts the transmission order according to the urgency of the instruction; the fault-tolerant control mechanism presets fault recognition conditions, and activates the backup scheme when the actuator response delay exceeds 500 milliseconds or the feedback data is abnormal for three times in a row.

[0064] Specifically, during the execution of the instruction, the communication protocol first encapsulates the control instruction as a standardized data packet, which is distributed to each subsystem at a cycle of 20 milliseconds through a high-speed bus. The cooperative verification algorithm receives the displacement sensor data of each actuator in real time, calculates the deviation between the actual displacement and the target value of the instruction, and triggers the inter-actuator cooperative calibration program when the deviation exceeds 3%. The instruction tracing system generates an operation record containing a millisecond-level timestamp after each instruction execution, which forms an associated data chain with the abnormal detection results of the safety monitoring system. The safety monitoring system identifies the mutually exclusive instruction combination that may cause mechanical interference by parallel scanning of the instruction queue, such as triggering the interlock mechanism when sending the walking mechanism forward and the boom rotation instructions at the same time. The efficiency optimization algorithm assigns transmission priority according to the type of instruction, and the priority of the position correction instruction is set to the highest level to ensure that its transmission delay does not exceed 10 milliseconds. The fault-tolerant control mechanism automatically switches to the backup hydraulic drive system and recalculates the motion trajectory when a walking mechanism servo controller failure is detected, and the switching process takes less than 200 milliseconds. Through the combined application of the above technical means, the system realizes seamless switching in the fault state while maintaining the execution accuracy, effectively improving the overall control reliability and response speed.

[0065] The present application further proposes a communication protocol based on a distributed control architecture to establish a real-time data exchange mechanism between subsystems, a verification logic based on a cooperative verification algorithm to ensure the coordinated action of each actuator, a recording function based on an instruction tracing system to record the execution process and timestamp information of each instruction, a real-time monitoring function based on a safety monitoring system to prevent the occurrence of instruction conflict and execution anomaly, a scheduling strategy based on an efficiency optimization algorithm to reduce the time delay in the instruction transmission and execution process, and a backup scheme based on a fault-tolerant control mechanism to automatically start the backup control scheme when some actuators fail.

[0066] The real-time data exchange mechanism adopts an industrial Ethernet communication protocol to realize millisecond-level data transmission, wherein the data frame structure includes a device identification code and a check code field; the cooperative verification algorithm verifies the action synchronization by comparing the displacement feedback values of the actuators of the two cranes, and triggers a re-verification process when the deviation exceeds a preset threshold; the instruction tracing system stores instruction execution logs using blockchain technology, and each operation record includes a timestamp accurate to microseconds and an operator electronic signature; the safety monitoring system sets a double-checking mechanism to identify abnormal operations by comparing control instructions with execution feedback data; the efficiency optimization algorithm uses a dynamic priority scheduling strategy to control the transmission delay of safety instructions within 50 milliseconds; and the fault-tolerant control mechanism presets three sets of redundant control schemes, and automatically switches to a backup controller when the actuator response times out.

[0067] Specifically, during the operation of the distributed control architecture, the industrial Ethernet communication protocol realizes real-time data interaction between the walking mechanism controller and the main control system by dividing dedicated data transmission channels, and the device identification code ensures that the data packet is accurately routed to the target subsystem. The cooperative verification algorithm synchronously collects the displacement amounts of the walking encoders of the two cranes when the servo motor executes the position correction instruction, and verifies the action synchronization through difference calculation. When a displacement deviation of more than 2 mm is detected, the compensation program is automatically triggered. The instruction tracing system records the sending time, execution status, and operator information of each control instruction using distributed storage nodes, forming an unalterable operation trajectory chain. The safety monitoring system identifies abnormal displacement that exceeds the safety boundary by comparing the theoretical motion trajectory with the actual feedback data in parallel, and immediately sends an emergency stop instruction. The efficiency optimization algorithm dynamically allocates network bandwidth resources according to the urgency of the control instruction, and compresses the transmission period of safety control instructions to 30 milliseconds. The fault-tolerant control mechanism automatically switches to a state estimation algorithm based on motor current values to maintain basic control functions when the hydraulic system pressure sensor fails, and the switching time of the backup control scheme does not exceed 100 milliseconds.

[0068] The present application further proposes a safety state of each link of the real-time monitoring system of the distributed architecture including a multi-level safety monitoring system; a prediction function based on a risk assessment model generates an early warning signal in a timely manner when a potential risk is detected; an inference mechanism using a decision tree model automatically selects the optimal processing scheme according to the type of anomaly; an emergency handling process based on a hierarchical response mechanism starts the corresponding emergency program according to different urgency; a protection function through a safety interlocking mechanism ensures the reliability and safety of the emergency handling process; statistical analysis based on emergency exercise results updates and optimizes emergency plans periodically; the data collection function of the safety event recording system records all relevant information of safety events completely; and evaluation criteria based on a continuous improvement mechanism evaluate and optimize safety measures periodically.

[0069] The multi-level safety monitoring system adopts a distributed sensor network to realize a three-level monitoring architecture of a physical layer, a control layer and a system layer. The physical layer sensors monitor the stress state of the mechanical structure, the control layer sensors collect the operating parameters of the actuator, and the system layer sensors monitor the environmental parameters. The risk assessment model integrates historical fault data and real-time monitoring data, calculates the risk probability value through a Bayesian network, and triggers a warning signal when the risk value exceeds a preset threshold. The decision tree model sets branch nodes corresponding to different abnormal types, and the terminal nodes are associated with a preset treatment scheme library. The accuracy rate of abnormal type identification reaches 98.6%. The hierarchical response mechanism sets three-level response procedures of yellow, orange and red, corresponding to the start of device speed reduction, local shutdown and full-system emergency braking measures. The safety interlocking mechanism ensures that the emergency command is executed preferentially and without conflict through hard-wire connection and software interlocking. The emergency drill system stores not less than 200 groups of historical fault scene data, and automatically generates an evaluation report including 12 indicators such as response time and disposal effect after each drill. The safety event recording system stores event data using blockchain technology to ensure that the data cannot be tampered with. The continuous improvement mechanism sets a quarterly evaluation cycle, and updates not less than 3 safety control parameters each time.

[0070] In implementation, when the distributed pressure sensor detects that the sling tension exceeds the safety threshold, the physical layer monitoring data is calculated by the risk assessment model to generate a red warning signal, triggering the decision tree model to automatically match the steel wire rope fracture emergency plan. The hierarchical response mechanism immediately starts the red emergency program, the safety interlocking system preferentially cuts off the hydraulic power source and activates the mechanical brake, while the position sensor monitors the falling trajectory of the hoisted object in real time. During the execution of the emergency command, the safety event recording system records 28 parameters such as brake pressure and displacement change at an interval of 0.1 seconds to form a complete event log. The post-event analysis module calls the data of the last 3 emergency drills to compare the disposal effect of this event and automatically optimize the brake response time parameter. During quarterly evaluation, the system calls the recorded 17 safety event data in the past 90 days, and finds through cluster analysis that there is a correlation between environmental temperature and sensor false alarm rate, and accordingly adjusts the temperature compensation coefficient from 0.85 to 0.92. The whole process through multi-level monitoring, intelligent decision-making, hierarchical response and data-driven continuous improvement shortens the system emergency response time to 0.8 seconds, and reduces the major accident rate to less than 0.03 times per thousand hours.

[0071] Through the above technical solutions, the application realizes multi-dimensional safety monitoring of the whole process of double-machine lifting operation, can quickly identify potential risks and automatically execute hierarchical emergency measures under complex working conditions, effectively reduces the risk of secondary accidents caused by human judgment delay, and continuously improves the reliability of the safety protection system through a closed-loop feedback mechanism.

[0072] The above merely provides an example of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A dual crane lifting dynamic balance control method based on real-time load feedback, characterized in that, Includes the following steps: Step S100, System initialization and parameter preset stage: Based on the preset system control requirements, set the core control parameters including load deviation threshold, proportional coefficient, integral coefficient and derivative coefficient, and configure the system sampling frequency, safety angle threshold and height compensation coefficient, and complete the startup and self-test procedures of all sensor systems. Step S200, Multi-source data synchronous acquisition and fusion processing stage, based on the collaborative work of the multi-sensor system, synchronously acquires real-time load data, lifting height data, spatial attitude angle data, position coordinate data, motion status data and environmental parameter data of the two cranes, and performs time synchronization, coordinate unification and filtering processing on the acquired multi-source heterogeneous component data through data fusion technology. Step S300, the multi-dimensional calculation and analysis stage of load status, based on the multi-dimensional data after fusion processing, comprehensively calculates multiple status parameters such as real-time load deviation value, height coordination deviation value, spatial angle deviation value, position synchronization deviation value and motion coordination degree, and uses a multi-condition judgment algorithm to evaluate the load balance status and perform security analysis. Step S400, Intelligent Decision and Control Command Generation Stage: Based on the state assessment results, a multi-objective optimization algorithm is used to generate a comprehensive control strategy that includes speed adjustment commands, position correction commands, attitude adjustment commands, and safety control commands. At the same time, a command priority mechanism and a forward-looking control mechanism are established. Step S500, Multi-system collaborative execution and real-time monitoring stage: Based on the generated control instruction set, the distributed control architecture coordinates each execution subsystem to complete the parameter adjustment task, and monitors the instruction execution status and execution effect in real time to ensure the coordination and security of system operation; Step S600, Dynamic Feedback and Adaptive Optimization Stage: Based on the feedback data collected during the execution process, a closed-loop control strategy is adopted to adjust the control parameters in real time, and the system performance is continuously optimized through machine learning algorithms to establish an adaptive control mechanism. Step S700, the safety monitoring and emergency response phase, based on real-time safety monitoring data, identifies abnormal system states through a multi-level safety monitoring system, and activates the corresponding emergency response mechanism according to the preset emergency response plan.

2. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 1, characterized in that, Step S200 includes: Step S210: Based on the high-precision weight sensor system, real-time load data of the first crane and the second crane are collected synchronously at a preset sampling frequency to ensure the synchronization and accuracy of data collection. Step S220: Based on the high-resolution encoder system, the lifting height data of the first crane and the second crane are acquired in real time, and data filtering technology is used to eliminate measurement noise. Step S230: Based on the tilt sensor system, accurately measure the deflection angle data of the hoisting sling with respect to the vertical direction to provide data support for spatial attitude analysis; Step S240: Based on the GPS positioning module and laser ranging system, acquire the real-time spatial position coordinate data of the two cranes and establish a unified spatial coordinate system; Step S250: Based on the motion sensor system, collect the crane's operating speed data and acceleration data to provide a data foundation for motion coordination analysis; Step S260: Based on environmental monitoring sensors, collect environmental parameter data such as wind speed, wind direction and ambient temperature in real time to provide environmental factor data for safety assessment; Step S270: Based on the time synchronization system, perform timestamp alignment processing on the multi-source heterogeneous component data to ensure the time consistency of the data; Step S280 involves performing noise reduction, outlier processing, and data fusion on the collected raw data based on multi-level filtering algorithms and data fusion technology to improve data quality and reliability.

3. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 2, characterized in that, Step S300 includes: Step S310: Based on real-time load data, the load deviation values ​​of the two cranes are calculated by difference, and the moving average filtering technique is used to smooth data fluctuations. Step S320: Based on the lifting height data, calculate the height coordination deviation value, and eliminate the impact of height difference on load balance through a height compensation algorithm; Step S330: Based on the spatial attitude angle data, calculate the angle deviation value, and use coordinate transformation technology to convert the angle deviation into a spatial position deviation; Step S340: Based on the spatial position coordinate data, the position synchronization deviation value is calculated through relative position, and a mapping relationship between position deviation and motion control is established; Step S350: Based on the running speed and acceleration data, calculate the motion coordination degree through kinematic analysis to evaluate the motion synchronization performance of the two cranes; Step S360: Based on multiple state parameters, a fuzzy logic algorithm is used to determine the load balance status and generate a system status assessment report. Step S370: Based on environmental parameters and system status parameters, identify safety hazards using a risk assessment model and calculate the system safety level; Step S380: Based on the comprehensive evaluation results, generate a status assessment report that includes status ratings and adjustment priorities to provide a basis for control decisions.

4. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 3, characterized in that, Step S300 further includes: Step S390: Based on the weighted fusion algorithm, multiple state parameters are fused into a comprehensive load deviation index, which comprehensively considers the mutual influence of load deviation, height deviation, angle deviation and position deviation. Step S391: Construct a multidimensional state evaluation matrix based on state parameters, and extract key state features using principal component analysis. Step S392: Based on the fuzzy logic algorithm, the system state is divided into multiple levels, and the correspondence between the state levels and the control strategy is established. Step S393: Based on time series analysis, perform trend prediction on historical operating data to predict the state change trend in future time periods; Step S394: Calculate the current risk coefficient of the system based on probability and statistics methods, and formulate corresponding control strategies for different risk levels; Step S395: Based on the Markov model, establish a state transition probability matrix to predict the transition probability between states, providing a basis for preventive control.

5. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 4, characterized in that, Step S400 includes: Step S410: Based on the state rating results, the decision tree algorithm is used to select the most suitable control strategy to generate the algorithm. Step S420: Based on a multi-objective optimization algorithm, the optimal speed adjustment command is generated by comprehensively considering multiple objectives such as load balancing, motion stability, and energy consumption optimization. Step S430: Based on the position synchronization deviation value, a PID control algorithm is used to generate a position correction command to achieve precise position control; Step S440: Based on the angle deviation value, an attitude adjustment command is generated through spatial geometric transformation to ensure the spatial symmetry of the hoisting system; Step S450: Based on the risk coefficient assessment results, generate graded safety control instructions, including multiple levels of safety measures such as early warning prompts, speed limit control, and emergency shutdown. Step S460: Based on the cooperative control algorithm, generate a cooperative control instruction set that ensures the control commands of the two cranes match each other; Step S470: Based on the real-time system status, an adaptive algorithm is used to generate parameter adjustment instructions to optimize control parameters in real time.

6. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 5, characterized in that, Step S400 further includes: Step S480: Based on the assessment of the urgency of the situation, establish a priority mechanism for control instructions to ensure that important instructions are executed first. Step S481: Based on the model predictive control algorithm, generate forward-looking control commands according to the state trend prediction results to improve the system response speed; Step S482: Based on the fault-tolerant control algorithm, when a sensor fault or data anomaly is detected, a reliable fault handling instruction is generated using the state estimation value. Step S483: Based on the energy consumption optimization algorithm, under the premise of ensuring system safety, generate energy-saving optimization instructions to reduce system operating energy consumption; Step S484: Based on the input from the human-computer interaction system, generate auxiliary decision-making instructions to provide control suggestions for the operator; Step S485: Based on the security verification algorithm, generate verification code for each control command to ensure the integrity and security of the command.

7. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 6, characterized in that, Step S500 includes: Step S510: Based on the scheduling algorithm of the instruction distribution system, the speed adjustment instruction is accurately sent to the corresponding frequency conversion control system. Step S520: Based on the high-precision positioning function of the servo control system, execute the position correction command to ensure the accurate positioning of the walking mechanism; Step S530: Based on the coordinated control capability of the multi-axis controller, execute the attitude adjustment command to achieve smooth movement of the boom; Step S540: Based on the rapid response characteristics of the security protection system, execute security control commands to ensure the safe operation of the system; Step S550: Based on the real-time data acquisition capability of the distributed sensor network, collect feedback data from each actuator; Step S560: Based on the real-time analysis function of the status monitoring system, detect abnormal situations during instruction execution; Step S570: Based on the multi-index comprehensive evaluation function of the effect evaluation algorithm, calculate the success rate and effect index of instruction execution; Step S580: Based on the quantitative scoring standards of the quality evaluation system, objectively score the execution effect of each control command.

8. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 7, characterized in that, Step S500 further includes: Step S590: Based on the communication protocol of the distributed control architecture, establish a real-time data exchange mechanism between the subsystems; Step S591: Based on the verification logic of the collaborative verification algorithm, ensure that the actions of each actuator are coordinated and consistent. Step S592: Based on the recording function of the instruction traceability system, record the execution process and timestamp information of each instruction in its entirety; Step S593: Based on the real-time monitoring function of the security monitoring system, prevent command conflicts and execution anomalies from occurring; Step S594: Based on the scheduling strategy of efficiency optimization algorithm, reduce the time delay in instruction transmission and execution; Step S595: Based on the fault-tolerant control mechanism, the backup scheme is automatically activated when some actuators fail.

9. A dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 8, characterized in that, Step S600 includes: Step S610: Based on the distributed architecture of the multi-dimensional feedback data acquisition system, collect feedback information from each execution stage in real time; Step S620: Based on the parameter adjustment strategy of the adaptive control algorithm, the operating parameters of the control system are optimized in real time; Step S630: Based on the forward-looking characteristics of the model predictive control algorithm, adjust the control strategy according to the dynamic characteristics of the system. Step S640: Based on the data analysis capabilities of machine learning algorithms, establish a system performance optimization model and continuously improve it; Step S650: Based on the adaptive characteristics of the online parameter tuning algorithm, automatically adjust the proportional, integral, and derivative parameters of the controller; Step S660: Based on the comprehensive evaluation function of the multi-dimensional evaluation index system, periodically evaluate the control effect of the system; Step S670: Based on the search capability of the optimization algorithm, continuously improve the control strategy to enhance system performance; Step S680: Based on the storage and retrieval functions of the knowledge base system, save the optimized control parameters and strategy scheme.

10. The dual-machine lifting dynamic balance control method based on real-time load feedback according to claim 9, characterized in that, Step S700 includes: Step S710: Based on the distributed architecture of the multi-level security monitoring system, monitor the security status of each link of the system in real time; Step S720: Based on the predictive function of the risk assessment model, a warning signal is generated in a timely manner when a potential risk is detected; Step S730: Based on the reasoning mechanism of the decision tree model, the optimal processing solution is automatically selected according to the anomaly type. Step S740: Based on the tiered response mechanism, the emergency response process initiates the corresponding emergency procedures according to different levels of urgency. Step S750, based on the protection function of the safety interlock mechanism, ensures the reliability and safety of the emergency response process; Step S760: Based on the statistical analysis of the emergency drill results, regularly update and optimize the emergency plan; Step S770: Based on the data collection function of the security event recording system, record all relevant information about security events in a complete manner; Step S780: Based on the evaluation criteria of the continuous improvement mechanism, regularly evaluate and optimize the safety measures.

Images