Crane operation safety early warning data processing method, risk assessment method and system

Abstract

The present application relates to the technical field of crane operation safety early warning and risk assessment, in particular to a crane operation safety early warning data processing method, a risk assessment method and a system. The data processing method first collects the original measurement values of multiple sensors at time t in real time, and converts them into unified dimensionless chord length through chord length mapping. Then, the drift increment is calculated based on the last optimal value as the reference, and the drift proportion and spectral entropy are combined to generate a credibility weight and normalize it. Finally, the chord length is weighted and fused based on the normalized weight, and the optimal measurement value is obtained through inversion, providing reliable input for subsequent risk assessment and solving the problems of non-uniform scale, drift and difficult quantification of credibility of multi-sensor data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of crane operation safety early warning and risk assessment, in particular to a crane operation safety early warning data processing method, a risk assessment method and a system. BACKGROUND

[0002] As the core heavy equipment in the fields of port loading and unloading, engineering construction, etc., the crane has a complex and changeable operation environment, and multiple factors such as meteorological conditions, ground state, spatial obstacles and personnel activities directly affect the safety and stability of hoisting operation. With the increasing requirements of industrial production on operation efficiency and safety standards, the traditional safety control mode relying on manual observation and experience judgment has been difficult to meet the needs of precise and global risk prevention and control, and it is urgent to build a full-process closed-loop safety early warning system through sensor perception, data processing and intelligent assessment technology to provide scientific and reliable decision support for crane operation.

[0003] At present, multiple types of sensors (such as anemometers, inclinometers, radars, etc.) have been gradually introduced in the field of crane safety monitoring to collect environmental and working condition data, and preliminary risk assessment is carried out based on the collected data. At present, simple mean fusion or fixed threshold screening is used for data processing, and the original measurement values of the sensors are directly used for risk determination; risk assessment mainly uses single-point coefficient determination, which compares the measurement value of a single dimension with a preset threshold to output a binary result of "safe / dangerous". Although some schemes introduce multi-dimensional risk coefficients, they do not achieve effective fusion and still rely on independent determination logic. These technologies have improved the safety of operation to some extent, but there are still obvious limitations in adaptability and accuracy in complex operation scenarios.

[0004] In terms of data processing, the scales of multi-source heterogeneous data are not unified at present, and the measurement ranges and physical dimensions of different types of sensors are different, which may lead to data distortion if directly fused, making it difficult to form a unified evaluation benchmark. At the same time, data drift and error are difficult to dynamically correct, and sensors may drift over time due to long-term operation, and instantaneous environmental interference may also cause measurement errors. The existing fixed threshold screening method cannot adaptively capture these dynamic deviations, resulting in insufficient data reliability. Secondly, the data reliability lacks quantitative evaluation. When multiple sensors are redundantly collected, equal weight fusion or subjective weighting is usually used, which cannot dynamically adjust the weight according to the data consistency and drift degree, making it difficult to highlight the role of high-reliability data and reducing the reliability of the fusion result.

[0005] In the risk analysis link based on data processing, there is a hard threshold step mutation, and the original risk coefficient is usually linearly mapped or binary processed. When the measured value approaches the threshold value, the risk level is prone to "one-size-fits-all" mutation, resulting in shaking or false triggering of the early warning instruction, affecting the stability of the operation; at the same time, the risk assessment lacks global coverage capability, and usually focuses on single-point or local area risk judgment, and fails to convert discrete risk coefficients into continuous spatial risk field, which cannot intuitively present the global risk distribution and is prone to miss local high-risk areas; secondly, the risk response lacks gradual adaptability, and the traditional evaluation result can only output binary instruction, which cannot realize speed limiting, early warning, shutdown and other graded responses according to the risk gradient, and it is difficult to balance between safety and efficiency, and cannot meet the operation needs in different risk scenarios. SUMMARY

[0006] In order to solve the problem that the original measurement values of multiple sensors in the crane operation are not uniform in scale, the data drift and the reliability are difficult to quantify, resulting in low data reliability, the present application provides a data processing method for crane operation safety warning. Based on this, in order to solve the technical problems of hard threshold mutation of original risk coefficient and lack of global coverage in risk judgment in crane operation risk assessment, on the basis of the data processing method, the present application further provides a crane operation risk assessment method. Combined with the data processing method and the risk assessment method, the present application further provides a crane operation safety warning and risk assessment system.

[0007] To achieve the above purpose, the present application provides the following technical scheme:

[0008] A data processing method for crane operation safety warning, comprising the following processing steps:

[0009] A1, real-time acquisition of original measurement value x of each sensor at t time in each position in the operation area i,t ;

[0010] A2, converting x i,t into a dimensionless chord length s at t time, R i is the system calibration radius of the i th sensor;

[0011] A3, based on s i,t , calculating the chord length drift increment of each sensor at t time , x best,i,t-1 is the optimal measurement value output by the i th sensor at t-1 time;

[0012] A4, based on Δs i,t , calculating the drift proportion of each sensor at t time , and calculating the spectral entropy at t time ; combined with the spectral entropy decay factor ε, the reliability weight of each sensor at time t is generated ; Δs j,t is the chord length drift increment of the original measurement value of the jth sensor at time t; ln is the natural logarithm; exp(·) is the exponential function with the natural constant e as the base;

[0013] A5, based on κ i,t , the normalized weight of each sensor at time t is calculated , κ j,t represents the reliability weight of the jth sensor at time t;

[0014] A6, based on w i,t , the chord lengths of all sensors are weighted and fused to obtain the fused chord length ; and the optimal measurement value x of each sensor at time t is calculated by the chord length inversion formula best,i,t as the input data of risk assessment.

[0015] As a further improvement of the above scheme: the sensors include six types for measuring wind speed, ground bearing capacity, ground slope, spatial obstacles, personnel intrusion and light intensity.

[0016] As a further improvement of the above scheme: the value of the spectral entropy decay factor ε is 1.

[0017] A crane operation risk assessment method, comprising the following evaluation steps:

[0018] B1, the optimal measurement value obtained by the above-mentioned data processing method for crane operation safety warning is converted into the corresponding original risk coefficient x;

[0019] B2, input each x into the S-type compression function to generate the corresponding elastic sub-coefficient; α and β are the bending position parameter and bending strength parameter respectively, and tanh is the hyperbolic tangent function;

[0020] B3, the working area is projected vertically into a regular hexagon, and six types of elastic sub-coefficients are assigned to the six vertices as fixed height values; the height between adjacent vertices is linearly interpolated, and the heights are connected in turn to form a closed boundary;

[0021] B4, an polar coordinate system is established inside the regular hexagon, and the risk height of any point inside the working area is calculated using the harmonic interpolation kernel formula;

[0022] B5, compare all risk heights at the current time with the preset safety plane threshold point by point, output green light, yellow light, red light three risk levels and corresponding operation instructions, wherein green light is normal operation, yellow light is speed limit operation, and red light is shutdown warning.

[0023] As a further improvement of the above scheme: the original risk coefficient of each type is calculated as follows:

[0024] The original risk coefficient of meteorological environment HJx is:

[0025] ;

[0026] In the formula, V best is the optimal wind speed value; V max is the maximum wind speed set;

[0027] The original risk coefficient of ground bearing CZx is:

[0028] ;

[0029] In the formula, σ best is the optimal ground bearing capacity value; σ need is the bearing capacity value required by the operation;

[0030] The original risk coefficient of ground flatness PZx is:

[0031] ;

[0032] In the formula, δ best is the optimal ground slope value; tan is the tangent function;

[0033] The original risk coefficient of space obstacles KJx adopts a binary quantization rule. If it is determined that the hoisting operation path is not blocked, it indicates no risk, and KJx=1; if there is a blocking obstacle, it indicates risk, and KJx=0;

[0034] The original risk coefficient of personnel intrusion RQx adopts a binary quantization rule. If it is determined that the operation path has no personnel intrusion, it indicates no risk, and RQx=1; if personnel intrusion is detected, it indicates risk, and RQx=0;

[0035] The original risk coefficient of light intensity ZMx is:

[0036] ;

[0037] In the formula, E best is the optimal light intensity value; min(·) is the minimum value operation; lx is the light intensity unit.

[0038] As a further improvement of the above scheme: the optimization interval of parameter α is [0.2, 3.0], and the optimization step is 0.2; the optimization interval of parameter β is [0.1, 1.0], and the optimization step is 0.1; both parameters are screened for the optimal value through field enumeration.

[0039] As a further improvement of the above scheme: the calculation formula of linear interpolation of the height between adjacent vertices is as follows:

[0040] ;

[0041] η is the transition ratio parameter; Z boundary (η) represents the boundary height when the ratio parameter is η; z k is the fixed height value at the position of the kth vertex; z k+1 is the fixed height value at the position of the k+1th vertex.

[0042] As a further improvement of the above scheme: the harmonic interpolation kernel formula is:

[0043] ;

[0044] K is the total number of boundaries, which is 6; R is the radius of the circumscribed circle of the regular hexagon; λ is the tension adjustment parameter, λ∈[0,0.5]; r is the radial distance from any point inside the regular hexagon to the center of the regular hexagon, and θ is the included angle between the radial direction of the arbitrary point and the positive direction of the X axis; is the polar angle of the midpoint of the kth boundary, is the boundary height at z (r,θ) corresponding position; S(r,θ) is the risk height at the position corresponding to (r,θ).

[0045] As a further improvement of the above scheme: when the risk height of all points in the work area is lower than the low-risk threshold, it is determined as a green light state, and the normal operation instruction is output;

[0046] When there is a point in the work area whose risk height is between the low-risk threshold and the medium-risk threshold, it is determined as a yellow light state, and the speed-limit operation instruction is output;

[0047] When the risk height of any point in the work area exceeds the medium-risk threshold, it is determined as a red light state, and the emergency stop instruction is output.

[0048] A crane operation safety warning and risk assessment system, comprising:

[0049] A sensor module comprising six types of sensors for measuring wind speed, ground bearing capacity, ground slope, spatial obstacles, personnel intrusion and light intensity, each type of sensor being deployed with multiple redundant links for real-time collection of original measurement values at each position in the work area at time t;

[0050] A data acquisition and transmission module for synchronously receiving the original measurement values collected by the sensor module, transmitting them to the core processing module, and ensuring data timing consistency;

[0051] The core processing module, pre-installed with data processing methods and risk assessment methods, is used for:

[0052] The above-described data processing method for crane operation safety early warning transforms the original measured values ​​into optimal measured values.

[0053] The above-described crane operation risk assessment method generates the original risk coefficient and elasticity coefficient based on the optimal measurement values, and completes the risk level determination.

[0054] The visualization module receives hazardous surface data output from the core processing module, generates a two-dimensional color risk map, and displays the overall risk distribution, risk level, and risk causes.

[0055] The output execution module, including a three-color warning light, a speed adjustment unit, and an emergency stop unit, is used to receive risk level instructions from the core processing module and execute normal operation, speed-limited operation, or emergency stop operation.

[0056] Compared with the prior art, the beneficial effects of the present invention are:

[0057] 1. This invention addresses the problems of inconsistent scales, data drift, and low data reliability caused by the difficulty in quantifying the reliability of original measurements from multiple sensors in crane operations through a closed-loop process of "unified scale → dynamic drift capture → quantification of reliability → weighted fusion → restoration of optimal value". The specific implementation logic is as follows:

[0058] First, to address the issue of inconsistent scales in raw measurements from multiple sensors, scale normalization of multi-source heterogeneous data is achieved through chord length mapping; then, an independent system calibration radius R is set for each sensor. i (Adapt the measurement range individually according to the sensor type), and use the formula The original measured values ​​x of different physical dimensions (such as wind speed in m / s, slope in arc) and different ranges are converted into the original measured values ​​x. i,t This is transformed into a geometric chord length without physical dimensions; and thus, through R... i Personalized adaptation to ensure x i,t / R i By keeping the data within a reasonable range and avoiding saturation of the sine function, different types of measurements such as wind speed, slope, and illumination are ultimately unified to the same geometric scale, effectively eliminating fusion distortion caused by dimensional differences and establishing a unified benchmark for subsequent data processing.

[0059] Secondly, to address the "data drift" problem, precise capture and quantification of drift are achieved through dynamic drift increment calculation. The optimal measurement value output at the previous time step (t-1) is used as the time series reference, rather than a fixed threshold, and is calculated using the formula... Calculate the drift increment Δs of the current chord length relative to the historical best value. i,tThis benchmark features real-time updates, adapting to both the aging drift caused by long-term sensor operation and capturing short-term fluctuations due to transient environmental disturbances, Δs. i,t The absolute value and positive / negative value quantify the degree and direction of drift, respectively, providing a dynamic data quality basis for subsequent credibility assessment and avoiding the impact of unidentified drift data on the reliability of the results.

[0060] Finally, to address the difficulty in quantifying credibility and the resulting unreasonable fusion weights, a credibility quantification system based on spectral entropy was constructed: firstly, through... Calculate the drift ratio of a single sensor, and then use the spectral entropy formula. Quantifying the divergence of multiple data sets: H t A larger value indicates worse data consistency, and vice versa; subsequently, this is combined with the spectral entropy decay factor ε, through... Generate credibility weights, giving higher weights to data with good consistency and a reasonable drift ratio. Further... Weights are normalized to ensure that the sum of the weights of all sensors is 1, thus avoiding distortion of results due to weight overflow during fusion.

[0061] Finally, based on the normalized weight w i,t The chord lengths of all sensors are weighted and fused, and then the inversion formula is used. The measurement is restored to the optimal value consistent with the original measurement dimension. The entire process systematically solves three core problems through a progressive design of "unifying the scale to eliminate dimensional differences, dynamically identifying data biases by drift detection, assigning reasonable weights by spectral entropy quantification, and highlighting highly reliable data by weighted fusion". In the end, it outputs reliable data with "minimum discrepancy and maximum credibility" to provide high-quality input for subsequent risk assessment.

[0062] 2. This invention systematically solves the technical problems of sudden changes in the hard threshold of the original risk coefficient and the lack of full coverage in risk determination in crane operation risk assessment through a progressive design of "coefficient smoothing transformation → global geometric modeling → continuous interpolation coverage". The specific implementation logic is as follows:

[0063] To address the issue of "abrupt hard threshold changes in the original risk coefficient," a two-stage process of "original risk coefficient standardization + S-curve compression function smoothing" is employed to achieve continuous and gradual changes in the risk coefficient. First, based on optimal measurements, multi-dimensional physical quantities such as meteorological, ground, and spatial parameters are transformed into original risk coefficients in the 0-1 range (e.g., the space obstacle coefficient KJx and personnel intrusion coefficient RQx are binary 0 / 1 values, while the meteorological environment coefficient HJx is a linearly mapped value), laying the foundation for standardization. Subsequently, an S-curve compression function is introduced. (x; a, b) = b*tanh(a x) + (1-b)x, the original risk coefficient is nonlinearly smoothed: the S-shaped characteristic of the hyperbolic tangent function tanh can convert binary discrete values (such as KJx=0 or 1), linear step values into continuous and gradually changing elastic sub-coefficients, avoiding the hard threshold effect of "jumping from 0 to 1 directly"; the parameters a and b are screened for optimal values through field enumeration, and the steepness of the curve can be adjusted according to the sensitivity requirements of different risk types, retaining the physical meaning of risk and completely eliminating the "one vote veto" type of step mutation, so that the risk coefficient change is more in line with the gradual change characteristics of actual operation.

[0064] In view of the problem of "lack of global coverage of risk judgment", through the three-dimensional design of "geometric carrier construction + boundary continuous + internal interpolation filling", the global continuous coverage of the operation area risk is realized. First, the operation area is projected as a regular hexagon geometric carrier, and the 6 vertices are bound to the 6 types of elastic sub-coefficients (corresponding to 6 types of risks such as weather, ground bearing, space obstacles, etc.) one by one, and a fixed height value is given, to establish the corresponding relationship between "risk coefficient-geometric position"; at the same time, through the linear interpolation formula , the risk height of adjacent vertices is smoothly changed with the transition proportion parameter η (0-1 continuous change), forming a closed boundary without breakpoints, avoiding risk mutation in the boundary area, and laying a foundation for global coverage. Secondly, the polar coordinate system (r, theta) is established inside the regular hexagon to adapt to the symmetrical structure of the operation area and ensure the uniformity of interpolation; through the harmonic interpolation kernel formula , the risk height of any point inside is calculated: the formula dynamically allocates weights through "geometric distance from point to boundary segment" and "tension parameter lambda", so that points near the boundary are closer to the risk value of the corresponding boundary, and the central region integrates all boundary risks, finally converting the 6 discrete elastic sub-coefficients into a continuous, spatially differentiable two-dimensional danger surface, realizing the risk quantification of every position in the operation area, and completely solving the defects of traditional single-point judgment and local coverage.

[0065] Finally, by comparing the risk height S(r, theta) of all points in the danger surface with the preset safety threshold point by point, the three risk levels of green light, yellow light and red light and the corresponding instructions are output, ensuring that the risks in the whole area can be accurately identified and responded. The whole process is designed through the core design of "S-shaped compression function to eliminate hard threshold mutation, regular hexagon + interpolation algorithm to realize global coverage", combined with parameter optimization, coordinate system adaptation and other details, systematically solves two technical problems, realizes the continuity, globality and accuracy of risk assessment. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 The data processing flowchart in the present application.

[0067] Figure 2 The risk assessment flowchart in the present application.

[0068] Figure 3 The projection diagram of the working area in the present application.

[0069] Figure 4 The dangerous surface graph in the normal state in the present application.

[0070] Figure 5 The dangerous surface graph in the abnormal state of the space obstacle in the present application.

[0071] Figure 6 The dangerous surface graph in the abnormal state of the personnel intrusion in the present application.

[0072] Figure 7 The dangerous surface graph in the abnormal state of both the space obstacle and the personnel intrusion in the present application. DETAILED DESCRIPTION

[0073] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0074] I. Implementation purpose

[0075] As shown in Figure 1 and Figure 2 , the present application realizes real-time early warning for the working process of the crane through the whole-process closed loop of "sensor data coupling + risk assessment", and improves the safety of the work. First, the original measurement values of multiple sensors (meteorological, ground, space, illumination, etc.) are converted into high-confidence optimal measurement values through the sensor data coupling method based on spectral entropy confidence weighted chord length inversion; then, multi-dimensional risk assessment is carried out based on the reliable data, a continuous and visual dangerous surface is constructed, and accurate risk level and operation instructions are output to ensure the safety, stability and accuracy of the hoisting operation.

[0076] II. Implementation premise

[0077] 1. Hardware configuration

[0078] Multi-type sensor module: ultrasonic anemometer (meteorological environment measurement), ground bearing capacity detection equipment (ground bearing measurement), laser leveling instrument (ground flatness measurement), 3D radar (360° space obstacle and personnel intrusion scanning), illuminance sensor (light intensity measurement), multiple sensors of each type are deployed to realize redundant collection.

[0079] Core processing modules: data acquisition and transmission system (acquiring measurements from multiple sensors in real time), core main control system (running data coupling algorithm and risk assessment algorithm);

[0080] Output and execution modules: Visualization module (displays the coupling results of dangerous surfaces and data), three-color warning light (green / yellow / red risk level indication), speed adjustment module (executes speed limit commands), and emergency shutdown module (responds to high-risk signals).

[0081] 2. Software Configuration

[0082] Data coupling related parameters: system calibration scale R (set according to the measurement range of various sensors), spectral entropy exponential decay factor ε=1.

[0083] Risk assessment related parameters: Wind speed safety threshold V max =10.8m / s (level 6 wind speed), ground slope safety threshold θ≤3° (≈0.052rad), light safety threshold 50lx, S-shaped compression function parameter range (α∈[0.2,3.0], step size 0.2, 15 points in total; β∈[0.1,1.0], step size 0.1, 10 points in total), radius R of the circumcircle of the regular hexagonal working area (set according to the actual working range), harmonic interpolation kernel tension parameter λ∈[0,0.5], safety plane threshold (dividing green / yellow / red risk zones);

[0084] Pre-installed algorithm modules: spectral entropy confidence weighted chord length inversion coupling algorithm, original risk coefficient calculation algorithm, sigmoid compression function algorithm, dangerous surface generation algorithm, and risk level determination algorithm.

[0085] III. Spectral Entropy Reliability-Weighted String Length Inversion Coupling Algorithm - Sensor Data Coupling

[0086] This section transforms the raw measurements from multiple sensors into optimal measurements with "minimum discrepancy and maximum reliability," providing reliable data input for risk assessment. The implementation process follows a closed-loop logic of "measurement acquisition - scale unification - drift capture - reliability assessment - weighted fusion - physical quantity restoration."

[0087] 1. Collect raw measurement values ​​from multiple sensors

[0088] All sensors are activated, and the data acquisition and transmission system collects raw measurement values ​​from multiple sensors at various measurement points within the work area in real time and synchronously.

[0089] Ultrasonic anemometer: Collects raw wind speed measurements (V) at various measurement points within the work area. best .

[0090] Ground bearing capacity testing equipment: collects raw measured values ​​σ of the actual bearing capacity of the ground at various measurement points within the work area.best .

[0091] Laser leveling instrument: Collect the original measurement value θ of the maximum ground slope at each measurement point in the work area.

[0092] 3D radar: Collect the binary decision original measurement value (0=obstacle / personnel, 1=no obstacle / personnel) of space obstacles and personnel intrusion at each measurement point in the work area.

[0093] Illuminance sensor: Collect the original measurement value E of ambient light intensity at each measurement point in the work area. best .

[0094] The multiple measurement values of each type of sensor are uniformly marked as x i,t (the original measurement value of the ith sensor at time t; each sensor measures the same type of physical quantity; i∈[1,n], n represents the total number of sensors), ensuring data time sequence synchronization and avoiding fusion deviation caused by collection delay.

[0095] 2. Chord length mapping

[0096] Convert the original measurement values x i of different types and ranges into chord lengths s i of uniform scale through geometric mapping, eliminating dimensional differences. After mapping, the measurement values of all sensors are converted into geometric chord lengths s i without physical dimensions, realizing the conversion of "multi-source heterogeneous data→uniform geometric quantity" and laying a foundation for subsequent fusion.

[0097] The mapping formula is as follows:

[0098] ;

[0099] The system calibration radius R i of the ith sensor is set separately according to the type of sensor, adapting to the range of measurement values, ensuring that the ratio x i,t / R i is within a reasonable interval, and avoiding saturation of the sin function.

[0100] 3. Calculate the drift increment

[0101] The optimal measurement value x best,i,t-1 of the previous frame (original measurement value of the ith sensor at time t-1) is used as the time sequence reference to calculate the chord length drift increment of the current chord length relative to the time sequence reference, and the calculation formula is as follows:

[0102] ;

[0103] The chord length drift increment Δsi,t The chord length drift increment of the original measurement value of the i-th sensor at the t time reflects the deviation degree of the current measurement value from the historical optimal value, and a positive value or a negative value reflects the deviation direction, thereby providing a dynamic basis for subsequent credibility assessment.

[0104] 4. Generating spectrum entropy credibility

[0105] The divergence of multiple measurement values is measured by Shannon spectrum entropy, and the credibility weight of each sensor is generated by combining the drift proportion.

[0106] The calculation formula of the drift proportion is as follows:

[0107]

[0108]

[0109] In the formula, p i,t represents the drift proportion of the i-th sensor at the t time, reflecting the proportion of the drift increment of a single sensor in the total drift increment. Δs j,t represents the drift increment of the original measurement value of the j-th sensor at the t time, j ∈ [1, n].

[0110] The calculation formula of the spectrum entropy is as follows:

[0111]

[0112] In the formula, H t is the entropy value at the t time. The greater the value, the greater the divergence of multiple measurement values and the worse the consistency, and the smaller the value, the better the consistency.

[0113] The calculation formula of the credibility weight is as follows:

[0114]

[0115] In the formula, ε is a spectrum entropy decay factor, and the value is 1; ln is a natural logarithm; exp(·) is an exponential function with a natural constant e as the base; κ i,t represents the credibility weight of the i-th sensor at the t time. The smaller the entropy value and the more reasonable the drift proportion, the greater the credibility weight, and vice versa.

[0116] 5. Credibility weight normalization

[0117] The credibility weight is normalized to obtain the final normalized weight, thereby ensuring the mathematical rationality of weighted fusion.

[0118] The calculation formula of the credibility weight normalization is as follows:

[0119] ;​​​​

[0120] where w i,t represents the normalized weight of the i-th sensor at time t. j,t represents the credibility weight of the j-th sensor at time t.

[0121] The normalized weight directly determines the contribution of individual sensor measurements in the fusion result. The weight of a sensor with high credibility is greater, effectively reducing the interference of abnormal data.

[0122] 6. Chord length inversion coupling

[0123] By weighted fusion and geometric inversion, the chord length of the same scale is restored to the optimal measurement value consistent with the original measurement value dimension.

[0124] The calculation formula of weighted fusion is specifically represented as follows:

[0125] ;

[0126] where s t represents the credibility weighted chord length of all sensors at time t.

[0127] The calculation formula of chord length inversion is specifically represented as follows:

[0128] ;

[0129] where x best,i,t represents the optimal measurement value of the i-th sensor at time t.

[0130] For each type of sensor, output 1 optimal measurement value (such as optimal wind speed value, optimal slope value, optimal light intensity value, etc.), as the core data input of risk assessment.

[0131] Four, risk assessment

[0132] This part designs a multi-dimensional spatial evaluation scheme for crane operation scenarios. Through the closed-loop logic of "reliable data input-continuous risk modeling-accurate decision output", it solves the defects of traditional evaluation "hard threshold mutation, single point judgment". The process is as follows: first, use multiple sensors to collect environmental data such as weather, ground, space, and light, and get the optimal measurement value through data coupling processing; then convert it into 6 types of standardized original risk coefficients, generate continuous and gradual elastic sub-coefficients through S-type compression function, and eliminate step mutations; then take regular hexagon as the geometric carrier, and take the elastic sub-coefficient as the vertex height to build a continuous boundary, and generate a globally visible dangerous surface through harmonic interpolation kernel in the internal polar coordinate system; finally, compare the surface with the preset safety plane point by point, output green (normal operation), yellow (limited speed operation), and red (shutdown warning) three risk levels and corresponding instructions, realize global coverage, gradual response, and accurate control of multi-dimensional risk.

[0133] (I) Original risk coefficient calculation algorithm

[0134] This step converts various optimal measurement values into standardized and quantifiable original risk coefficients (value range 0-1), realizing the conversion of "physical quantity → risk quantitative value".

[0135] 1. Meteorological environment original risk coefficient HJx

[0136] Input the optimal wind speed value V best , based on the influence law of wind speed on lifting operation, use the following formula for quantization, and strengthen the risk sensitivity of high wind speed interval through quadratic function mapping.

[0137] ;

[0138] Where, V max = 10.8 m / s (6 level wind speed index), is the maximum wind speed set.

[0139] The closer the wind speed is to or exceeds V max , the closer HJx is to 0, indicating higher weather risk; when the wind speed is 0, HJx = 1, indicating the best weather conditions.

[0140] 2. Ground bearing original risk coefficient CZx

[0141] Input the optimal ground bearing capacity value σ best , the operation personnel's preset operation demand bearing capacity value σ need , and the following formula for quantization:

[0142] ;

[0143] CZx directly reflects the matching degree of ground bearing capacity and operation demand. The closer the value is to 1, the higher the matching degree is, and the lower the ground risk is. When the value is less than 1, it indicates that the ground bearing capacity is insufficient.

[0144] 3. Ground flatness original risk coefficient PZx

[0145] Optimal ground slope value δ after input coupling best ( converted into radian units), which is quantified by the following formula.

[0146] ;

[0147] δ best ≤ 3° (≈0.052 rad), PZx≥0.95, indicating that the ground flatness meets the operation requirements. As δ best increases, PZx decreases, and the ground risk gradually increases. When tanδ best =1, PZx=0, indicating that the ground flatness completely does not meet the operation conditions.

[0148] 4. Spatial obstacle original risk coefficient KJx

[0149] Optimal spatial obstacle judgment value after input coupling, using binary quantization rules. If it is determined that the hoisting operation path is not blocked, KJx=1 (no risk); if there is a blocking obstacle, KJx=0 (high risk).

[0150] Prioritize the smoothness of the operation path to avoid collision risks caused by obstacles, and reserve gradual change space for subsequent elastic sub-coefficient conversion.

[0151] 5. Personnel intrusion original risk coefficient RQx

[0152] Optimal personnel intrusion judgment value after input coupling, using binary quantization rules. If it is determined that there is no personnel intrusion in the operation path and dangerous area, RQx=1 (no risk); if personnel intrusion is detected, RQx=0 (extremely high risk).

[0153] Follow the "personnel safety first" principle, and take personnel intrusion as the core high-risk item. At the same time, avoid the rigid determination of "one vote veto" through subsequent elastic sub-coefficient processing.

[0154] 6. Illumination intensity original risk coefficient ZMx

[0155] Optimal illumination intensity value E best after input coupling, which is quantified by the following formula:

[0156] ;

[0157] In the formula, min(·) is the minimum value operation.

[0158] When E best ≥ 50lx, ZMx = 1, representing sufficient light, no risk of visual field; when E best < 50lx, ZMx < 1, and the lower the light intensity, the smaller ZMx, the higher the risk of visual field limitation, which needs to be compensated by speed limit and other methods.

[0159] (II) Elasticity sub-coefficient algorithm

[0160] This step is used to convert the original risk coefficient from a hard threshold, a linear mapping with a step to a gradual, interpretable and data-driven nonlinear compression value, while preserving the physical meaning and eliminating the "one vote zero / full grid" mutation, providing continuous, non-step and differentiable input source for subsequent spatial surface construction, thereby improving the smoothness and robustness of the overall risk assessment.

[0161] 1. Algorithm principle

[0162] Design an S-shaped compression function:

[0163] ;

[0164] Where:

[0165] is the original risk coefficient of each risk, HJx, CZx, PZx, KJx, RQx, ZMx.

[0166] tanh is the hyperbolic tangent function.

[0167] α is the bending position parameter; the larger α, the earlier the curve bends; the smaller α, the closer to a straight line.

[0168] β is the bending strength parameter, the larger β, the more β components, the more obvious saturation; the smaller β, the closer to the original line.

[0169] α, β are selected by exhaustive search on site, α ∈ [0.2, 3.0], step 0.2, a total of 15 points. β ∈ [0.1, 1.0], step 0.1, a total of 10 points.

[0170] 2. Parameter optimization

[0171] For the six types of original risk coefficients, the optimal α and β values are selected by exhaustive search method in the preset parameter range; the optimization goal is to "preserve the physical meaning of the original risk", "eliminate the step mutation", "improve the smoothness of the subsequent surface", to ensure that each elasticity sub-coefficient reflects the true risk level and has continuous and differentiable mathematical properties.

[0172] 3. Elasticity sub-coefficient

[0173] The original risk coefficients of the six categories are respectively substituted into the optimized S-shaped compression function, to generate corresponding meteorological environment elasticity sub-coefficients Φ HJ , ground bearing elasticity sub-coefficients Φ CZ , ground flatness elasticity sub-coefficients Φ PZ , spatial obstacle elasticity sub-coefficients Φ KJ , personnel intrusion elasticity sub-coefficients Φ RQ , and light intensity elasticity sub-coefficients Φ ZM .

[0174] All the elasticity sub-coefficients are continuous and gradually changing values in the interval of 0-1, without breakpoints or steps.

[0175] (Three) Dangerous surface generation algorithm

[0176] This step converts the six discrete elasticity sub-coefficients into a continuous, spatially differentiable, and geometrically visible two-dimensional risk field through geometric modeling and interpolation algorithm, realizing the upgrade of risk from "single point judgment" to "global coverage".

[0177] 1. Continuous boundary construction

[0178] As shown in Figure 3 , the crane operation area is projected onto the XY plane (operation plane) as a regular hexagon, and the six vertices P1-P6 of the regular hexagon are respectively in one-to-one mapping relationship with the six types of elasticity sub-coefficients, forming a "risk-geometry" correspondence.

[0179] 2. Boundary height assignment

[0180] The six elasticity sub-coefficients Φ HJ , Φ CZ , Φ PZ , Φ KJ , Φ RQ , and Φ ZM are respectively assigned to the six vertices of the regular hexagon as the fixed height values z k (k∈[1,6]) to establish the basic constraints of the boundary height.

[0181] 3. Boundary continuity processing

[0182] The linear interpolation formula is used to smoothly transition the height between adjacent vertices:

[0183] ;

[0184] In the formula, Z boundary (η) represents the boundary height when the proportion parameter is η. η is a transition proportion parameter. When η continuously changes in the interval of [0,1], the boundary height linearly gradually changes from the height z k of the kth vertex to the height z k+1 of the k+1th vertex., forming a closed boundary without breakpoints and continuous smoothness, avoiding the risk of high mutation at the boundary.

[0185] 4. Calculate the internal surface interpolation

[0186] Coordinate system establishment: Establish a polar coordinate system (r, θ) in the internal hexagon working area, where r is the radial distance from any point in the internal hexagon to the center of the hexagon, and θ is the included angle between the radial direction of the arbitrary point and the positive direction of the X-axis; This coordinate system adapts to the symmetrical structure of the hexagon, ensuring the uniformity and directionality of the interpolation calculation.

[0187] Harmonic interpolation kernel operation: The risk height of any point in the working area is calculated using the harmonic interpolation kernel formula, and the calculation formula is as follows:

[0188] ;

[0189] Where R is the radius of the circumscribed circle of the hexagon, defining the effective range of the working area. is the polar angle of the midpoint of the kth boundary segment, is the boundary height at . λ is the tension adjustment parameter.

[0190] λ∈[0,0.5]; When λ=0, the surface is the smoothest and the risk changes the most gently; The larger λ is, the more obvious the surface fluctuation is, and the more sensitive the response to local risk is. It can be adjusted in real time according to the accuracy requirements of the working scene.

[0191] The risk height of the internal point in the working area is obtained by weighted summation of the height of the six continuous boundary segments, and the weight is determined by the "geometric distance from the point to the boundary segment" and the "tension parameter λ": the closer to the boundary segment, the greater the weight, and the risk height is closer to the boundary value; The central region is affected by all boundary segments, reflecting the overall comprehensive risk.

[0192] 5. Surface visualization

[0193] Convert the risk height value S(r, θ) of all points in the internal hexagon to a two-dimensional color risk map, and display it through a visual display terminal; Use color gradient to represent risk level (low risk→green, medium risk→yellow, high risk→red), intuitively present the overall risk distribution, and facilitate operators to real-time control the risk status of the working area.

[0194] (Four) Risk assessment and instruction output

[0195] This step converts the continuous spatial risk field into discrete risk levels and explicit operation instructions through point-by-point comparison and analysis of the danger surface and the preset safety plane, forming a closed loop of "risk modeling-decision output".

[0196] The risk height S(r, θ) of all points in the dangerous curved surface is compared with the preset safety plane threshold one by one; the safety plane threshold is divided into three intervals according to the operation risk level, corresponding to three risk levels of green light (low risk), yellow light (medium risk) and red light (high risk) respectively, to ensure that the determination standard is unified and quantifiable.

[0197] 1. Risk level determination

[0198] Green light state: the risk height S(r, θ) of all points in the operation area is lower than the low risk threshold, representing that the global risk is controllable and meeting the normal operation conditions.

[0199] Yellow light state: the risk height S(r, θ) of some points in the operation area is between the low risk threshold and the medium risk threshold, representing that there is medium risk in the local area, and the operation safety needs to be improved by reducing speed and other ways.

[0200] Red light state: the risk height S(r, θ) of any point in the operation area exceeds the medium risk threshold, representing that there is high risk in the local or global area, and the operation needs to be stopped immediately to avoid safety accidents.

[0201] 2. Instruction output and execution

[0202] Green light instruction: control the green light of the three-color warning light to turn on, send the "normal speed operation" instruction to the speed adjustment module, and the crane operates normally at the preset rated speed.

[0203] Yellow light instruction: control the yellow light of the three-color warning light to turn on, send the "speed limited operation" instruction (the speed limit ratio can be preset, such as 50%-80% of the rated speed) to the speed adjustment module, and mark the medium risk area on the visual terminal to prompt the operator to pay attention.

[0204] Red light instruction: control the red light of the three-color warning light to turn on and trigger the sound and light alarm, send the "stop immediately" instruction to the emergency stop module, and the crane stops all lifting actions; at the same time, highlight the high risk area and risk causes (such as personnel intrusion, wind speed exceeding the standard, etc.) on the visual terminal for the operator to check the risk.

[0205] Five, evaluation examples

[0206] As shown in Table 1, the coordinates of the vertices of the regular hexagon are shown in Table 1, which provides accurate reference for the subsequent construction of the boundary of the dangerous curved surface and the spatial basis for the internal interpolation calculation. Figure 3 Table 1 Vertex coordinates

[0207]

[0208]

[0209] ​As shown in Table 2, the normal state of the dangerous surface table is listed as the field elastic sub-coefficient corresponding to the six vertices of the regular hexagon and the physical meaning.

[0210] Table 2 Vertex elastic sub-coefficient in normal state

[0211]

[0212] From Table 2, it can be seen that P1 (meteorological environment) coefficient is 0.452, P2 (ground bearing capacity) is 0.778, P3 (ground flatness) is 0.841, P4 (space barrier) is 1.000, P5 (personnel intrusion) is 1.000, and P6 (light intensity) is 0.739. All the coefficients are in a reasonable interval, reflecting the normal state of each risk dimension, and providing vertex height input data for generating a globally continuous normal working condition dangerous surface.

[0213] Based on the field elastic sub-coefficient data of each vertex in the normal state shown in Table 2, a dangerous surface graph as shown in Figure 4 is constructed and generated. In the graph, the blue, yellow and red planes correspond to the low risk threshold 1.0, the medium risk threshold 1.2 and the high risk threshold 1.5 respectively, which intuitively divides different risk level intervals. Figure 4 It can be seen that the global dangerous surface at the current time is lower than the low risk threshold 1.0, and there is no local medium and high risk area. The system is determined as green light state, which fully meets the safety conditions of the crane normal operation.

[0214] As shown in Table 3, the dangerous surface table for abnormal state of space barrier is listed as the field elastic sub-coefficient corresponding to the six vertices of the regular hexagon and the physical meaning.

[0215] Table 3 Vertex coefficient in abnormal state of space barrier

[0216]

[0217] From the data in Table 3, it can be seen that in the dangerous surface in the abnormal state of space barrier, the field elastic sub-coefficient of P4 vertex (space barrier) is 0, and the rest of the vertex coefficients are consistent with the normal state (P1=0.452, P2=0.778, P3=0.841, P5=1.000, P6=0.739). The data is used to simulate the scene of abnormal space barrier, and provides input for verifying the capture and presentation ability of dangerous surface to local abnormal risk.

[0218] Based on the field elastic sub-coefficient data of each vertex in the abnormal state of space barrier in Table 3, a dangerous surface graph as shown in Figure 5 is constructed and generated. As shown in Figure 5It can be seen that the danger surface of part of the region in the work area is between the low risk (blue, threshold 1.0) and the medium risk (yellow, threshold 1.2) plane, which intuitively presents the local disturbance effect of the spatial obstacle anomaly on the global risk distribution. Combined with the risk level determination rule, it can be determined that the current operation state is yellow, and the speed-limiting operation instruction needs to be executed to avoid local medium risk and ensure the safety of the hoisting operation.

[0219] As shown in Table 4, the danger surface table under the personnel intrusion anomaly state is shown, which lists the field elastic sub-coefficients of the six vertices of the regular hexagon and the physical meanings.

[0220] Table 4 Vertex coefficients under the personnel intrusion anomaly state

[0221]

[0222] From the data in Table 4, it can be known that in the danger surface under the personnel intrusion anomaly state, the field elastic sub-coefficient of P5 vertex (spatial obstacle) is 0, and the rest of the vertex coefficients are consistent with the normal state (P1=0.452, P2=0.778, P3=0.841, P4=1.000, P6=0.739). The data is used to simulate the scene of personnel intrusion anomaly, and provides input for verifying the capturing and presenting ability of the danger surface to local abnormal risk.

[0223] Based on the field elastic sub-coefficients of each vertex under the personnel intrusion anomaly state in Table 4, the danger surface graph as shown in Figure 6 can be constructed and generated. As shown in Figure 6 It can be seen that the danger surface of part of the region in the work area is between the low risk (blue, threshold 1.0) and the medium risk (yellow, threshold 1.2) plane, which intuitively presents the local disturbance effect of the personnel intrusion anomaly on the global risk distribution. Combined with the risk level determination rule, it can be determined that the current operation state is yellow, and the speed-limiting operation instruction needs to be executed to avoid local medium risk and ensure the safety of the hoisting operation.

[0224] As shown in Table 5, the danger surface table under the two index anomaly state is shown, which lists the field elastic sub-coefficients of the six vertices of the regular hexagon and the physical meanings.

[0225] Table 5 Vertex coefficients under the two index anomaly state

[0226]

[0227] As can be seen from the data in Table 5, in the dangerous surfaces under the two abnormal index states, except for the field elastomer coefficients of vertex P4 (spatial obstacle) and vertex P5 (personnel intrusion) which are 0, the coefficients of the other vertices are consistent with those under the normal state (P1=0.452, P2=0.778, P3=0.841, P5=1.000, P6=0.739). This data is used to simulate the scenario of abnormality in the two risk dimensions, and provides input for verifying the ability of dangerous surfaces to capture and present local abnormal risks.

[0228] Based on the field elastic coefficient data of each vertex under the two abnormal index states shown in Table 5, a system was constructed and generated as follows: Figure 7 The danger surface diagram is shown. (From...) Figure 7 As can be seen, the hazardous surfaces in some areas of the work area lie between the medium-risk (yellow, threshold 1.2) and high-risk (red, threshold 1.5) planes, visually demonstrating the superimposed disturbance effect of concurrent anomalies in two types of risk indicators—spatial obstacles and personnel intrusion—on the overall risk distribution. Based on the risk level determination rules, the current work status is clearly red, requiring immediate triggering of an emergency stop command and activation of audible and visual alarms. High-risk areas and their causes should be highlighted to prevent compound high-risk operational hazards.

[0229] The data analysis above reveals two key points: First, it clarifies the precise coordinates of the six vertices of the hexagonal area where the crane is operating, laying a spatial foundation for the geometric modeling of the hazardous surface. Second, it provides the on-site elastic coefficients of the vertices and the corresponding hazardous surface diagrams for each operating condition. These data correspond to three risk level scenarios: green, yellow, and red, providing crucial experimental support for verifying the hazardous surface's ability to capture different risk states, its effectiveness in presenting the overall risk distribution, and the accuracy of risk level determination.

[0230] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A data processing method for crane operation safety early warning, characterized in that, The method comprises the following processing steps: A1, real-time collection of raw measurement values x of each sensor at each position in the work area at time t i,t ; A2, x i,t Converts the chord length at time t to a uniform chord length , R i System calibration radius for the ith sensor A3, based on s i,t , calculate the chord length drift increment of each sensor at time t , x best,i,t-1 is the optimal measurement value of the ith sensor at time t-1; A4、 based on Δs i,t , calculate the drift ratio of each sensor at time t , and calculate the spectral entropy at time t ; combined with the spectral entropy decay factor ε, generate the credibility weight of each sensor at time t ; Δs j,t is the chord length drift increment of the jth sensor at time t; ln is the natural logarithm; exp(·) is the exponential function with the natural constant e as the base; A5, based on K i,t , calculate the normalized weight of each sensor at time t , K j,t represents the credibility weight of the jth sensor at time t; A6, based on w i,t The chord lengths of all sensors are weighted and fused to obtain the fused chord length at time t. ,pass Calculate the optimal measurement value x of each sensor at time t. best,i,t And use it as input data for risk assessment.

2. The data processing method for crane operation safety early warning according to claim 1, characterized in that, The sensors include six types for measuring wind speed, ground bearing capacity, ground slope, spatial obstacles, personnel intrusion and light intensity.

3. The data processing method for crane operation safety early warning according to claim 2, characterized in that, The spectral entropy decay factor ε is 1.

4. A method of crane operation risk assessment, characterized by, The method comprises the following evaluation steps: B1. Each optimal measurement value obtained by the crane operation safety early warning data processing method according to any one of claims 1-3 is converted into a corresponding original risk coefficient x; B2, input each x into the S-type compression function where the corresponding elastic sub-coefficient is generated; α and β are bending position parameters and bending strength parameters, respectively, and tanh is the hyperbolic tangent function. B3. The operation area is vertically projected into a regular hexagon, and six types of elastic sub-coefficients are assigned to the six vertices of the regular hexagon as fixed height values; the height between adjacent vertices is linearly interpolated, and the heights are sequentially connected to form a closed boundary; B4. An polar coordinate system is established inside the regular hexagon, and a harmonic interpolation kernel formula is used to calculate the risk height of any point inside the operation area; B5. All risk heights at the current time are compared with the preset safety plane threshold point by point, and three risk levels of green light, yellow light and red light and corresponding operation instructions are output, wherein the green light is normal operation, the yellow light is speed-limited operation, and the red light is shutdown warning.

5. A method of risk assessment of crane operations according to claim 4, characterized in that, The calculation of each type of original risk coefficient is as follows: The meteorological environment original risk coefficient HJx is: ； In the formula, V best is the optimal wind speed value; V max is the set maximum wind speed; The ground bearing original risk coefficient CZx is: ； In the formula, σ best is the optimal ground bearing capacity value; σ need is the work demand bearing capacity value; The ground flatness original risk coefficient PZx is: ； where δ best is the optimal ground slope value; tan is the tangent function; The spatial obstacle original risk coefficient KJx adopts a binary quantization rule, if it is determined that the hoisting operation path is not blocked, it indicates no risk, KJx=1; If there is a blocking obstacle, it indicates risk, KJx=0; The personnel intrusion original risk coefficient RQx adopts a binary quantization rule, if it is determined that the operation path is not invaded by personnel, it indicates no risk, RQx=1; if personnel intrusion is detected, it indicates risk, RQx=0; The light intensity original risk coefficient ZMx is: ； In the formula, E best is the optimal light intensity value; min(·) is the minimum value taking operation; lx is the light intensity unit.

6. A method of risk assessment of crane operations according to claim 4, characterized in that, The optimization interval of the parameter α is [0.2, 3.0], and the optimization step is 0.2; the optimization interval of the parameter β is [0.1, 1.0], and the optimization step is 0.1; both parameters are screened for the optimal value through field enumeration.

7. A method of risk assessment of crane operations according to claim 4, characterized in that, The calculation formula for linear interpolation of the height between adjacent vertices is as follows: ； where η is a transition ratio parameter; Z boundary (η) represents the boundary height when the ratio parameter is η; z k is a fixed height value at the position of the kth vertex; z k+1 is a fixed height value at the position of the k+1th vertex.

8. A method of risk assessment of crane operations according to claim 4, characterized in that, The harmonic interpolation kernel formula is: When the risk height of all points in the operation area is lower than the low risk threshold, it is determined as a green light state, and a normal operation instruction is output; ； In the formula, K is the total number of boundary segments, which takes the value of 6; R is the radius of the circumcircle of the regular hexagon; λ is the tension adjustment parameter, λ∈[0,0.5]; r is the radial distance from any point inside the regular hexagon to the center of the regular hexagon; θ is the angle between the radial direction of the point and the positive X-axis. Let the polar angle be the midpoint of the k-th boundary segment. for The boundary height at (r,θ) is the risk height at the location corresponding to (r,θ).

9. A method of risk assessment of crane operations according to claim 4, characterized in that, When there is a point in the operation area whose risk height is between the low risk threshold and the medium risk threshold, it is determined as a yellow light state, and a speed-limited operation instruction is output; When the risk height of any point in the operation area exceeds the medium risk threshold, it is determined as a red light state, and an emergency shutdown instruction is output. It comprises:

10. A crane operation safety early warning and risk assessment system, characterized in that, A sensor module comprising six types of sensors for measuring wind speed, ground bearing capacity, ground slope, spatial obstacles, personnel intrusion and light intensity, each type of sensor is deployed with multiple redundant links for real-time collection of original measurement values at time t at each position in the operation area; A data acquisition and transmission module for synchronously receiving the original measurement values collected by the sensor module, transmitting them to the core processing module, and ensuring data time sequence consistency; A core processing module preloaded with a data processing method and a risk evaluation method, for ​ The data processing method for crane operation safety warning of any one of claims 1-3, which converts the original measurement value into an optimal measurement value; The crane operation risk assessment method of any one of claims 4-9, which generates an original risk coefficient, an elasticity sub-coefficient based on the optimal measurement value, and completes risk level determination; A visual display module for receiving the dangerous surface data output by the core processing module, generating a two-dimensional color risk map, and displaying the global risk distribution, risk level and risk inducement; An output execution module including a three-color warning light, a speed adjustment unit and an emergency shutdown unit for receiving the risk level instruction output by the core processing module and executing normal operation, speed-limited operation or emergency shutdown operation.

Images