A load control system and method based on crane hoisting

Abstract

The present invention relates to the field of wind power generation control technology, and in particular to a load control system and method based on crane hoisting. First, the parameter data collected in real time is processed, the physical load of the crane is calculated by using a load judgment algorithm, and a potential correlation model is introduced; then, the physical load of the crane, the working physical parameters of the crane, the hoisting environmental parameters and the potential correlation model are analyzed, and the potential correlation model is updated to obtain the real-time working state of the crane; a deep learning neural network model is introduced to obtain the prediction results of the hoisting environmental parameters; the decision group of the working state of the crane is obtained by a decision tree algorithm, a network topology is constructed, and matching processing is performed to obtain the best response decision; finally, based on the operation instructions generated by the best response decision, the operating parameters of the crane are automatically adjusted. The technical problems of inaccurate and untimely control in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation control, and in particular to a load control system and method based on crane hoisting. Background Art

[0002] A crane load control system is a technology system used to monitor and manage loads during lifting operations. In heavy industry and construction, cranes are often used to move and lift heavy objects, and the emergence of load control systems aims to improve the safety, efficiency and precision of operations.

[0003] Wind farms are usually composed of large wind turbines, which need to be hoisted by cranes during installation and maintenance. Load control systems play a key role here. Load control systems can be integrated with the hoisting plan to ensure that the load is always within a safe range during the hoisting process to prevent accidents.

[0004] There are many load control methods. my country's patent "A power load control device, control method, terminal, medium and application", application number: "CN202110909033.8", publication date: "2021.11.2", mainly includes: using the preprocessed data set to train the LSTM model, improving the activation function in the LSTM unit, and obtaining an improved LSTM model; adding an attention mechanism to the improved LSTM model to obtain an improved AM-LSTM short-term load forecasting model; combining the improved AM-LSTM short-term load forecasting model with the hardware equipment power load control device to achieve power load control. The power load control device based on the improved short-term power load forecasting technology provided by the invention improves the AM-LSTM model, combines the deep learning model with the traditional power load control device, and introduces the attention mechanism into the field of power load control for the first time. It has a high prediction accuracy, can adapt to complex factors, and has more accurate control.

[0005] However, the above technology has at least the following technical problems: the control is not accurate enough and not timely enough. Summary of the invention

[0006] The embodiments of the present application provide a load control system and method based on crane hoisting, thereby solving the technical problems of inaccurate and untimely control in the prior art and achieving the technical effect of high-speed and accurate adaptive control.

[0007] The present application provides a load control system and method based on crane hoisting, which specifically includes the following technical solutions:

[0008] A load control system based on crane hoisting includes the following parts:

[0009] Parameter acquisition module, load detection module, environment perception module, mining module, data analysis module, intelligent prediction module, decision support module, decision module, control module, user interface, feedback module;

[0010] The parameter acquisition module is used to obtain the physical parameters of the crane, including the lifting weight, boom angle, and boom length, to provide a data basis for the load detection module;

[0011] The load detection module is used to calculate the physical load of the crane using a load judgment algorithm, monitor the physical load of the crane in real time, and output the real-time physical load data of the crane;

[0012] The environmental perception module is used to obtain and process real-time parameters of the hoisting environment, obtain environmental parameter data, including wind speed, temperature, humidity, and monitor changes in the hoisting environment in real time;

[0013] The mining module is used to obtain a potential correlation model between the working physical parameters and the environmental parameters based on the historical working physical parameters of the crane and the historical parameters of the hoisting environment, so as to provide a basis for the data analysis module;

[0014] The data analysis module is used to analyze the real-time data based on the load detection module and the environment perception module, and combine the potential association model of the mining module to obtain the real-time working status of the crane;

[0015] The intelligent prediction module is used to predict the environment through the environmental parameter prediction function based on the real-time environmental parameters to obtain the environmental parameter prediction results;

[0016] The decision-making auxiliary module is used to train the historical working data and historical response decisions of the crane through a decision tree algorithm to obtain a decision group of the crane's working status and provide a matching basis for the decision-making module; the historical working data includes the historical working physical parameter data of the crane and the historical parameter data of the hoisting environment;

[0017] The decision module is used to construct a network topology structure based on the real-time working state of the crane obtained by the data analysis module, the environmental parameter prediction results obtained by the intelligent prediction module, and the decision group of the working state of the crane obtained by the decision assistance module, perform matching processing to obtain the best response decision, and generate an operation instruction according to the best response decision;

[0018] The control module is used to automatically adjust the operating parameters of the crane according to the operating instructions of the decision module;

[0019] The feedback module is used to detect the output results of each module, and when an unsafe state is detected, a safety warning is fed back to the operator through the user interface;

[0020] The user interface is used to provide real-time feedback and guidance to the operator based on the operating instructions of the decision-making module and the actual operating status of the crane.

[0021] A load control method based on crane hoisting comprises the following steps:

[0022] S1. Collect parameter data in real time, process the collected crane physical parameter data and hoisting environment parameters, calculate the physical load of the crane using a load judgment algorithm, and mine and analyze the crane's historical working physical parameter data and hoisting environment historical parameter data, and introduce a potential correlation model between the crane's working physical parameters and hoisting environment parameters;

[0023] S2. Perform data analysis on the physical load of the crane, the working physical parameters of the crane, the hoisting environment parameter data and the potential correlation model, and at the same time, update the potential correlation model in real time to obtain the real-time working status of the crane;

[0024] S3. Predict future environmental changes based on historical parameter data of the hoisting environment, introduce a deep learning neural network model, and obtain the prediction results of the hoisting environment parameters;

[0025] S4. Obtain the decision group of the crane working state through the decision tree algorithm, build a network topology structure based on the decision group of the crane working state, and perform matching processing based on the real-time working state of the crane and the prediction results of environmental parameters combined with the decision group of the crane working state to obtain the best response decision;

[0026] S5. Automatically adjust the operating parameters of the crane based on the operating instructions generated by the best response decision;

[0027] Applied to a load control system based on crane hoisting.

[0028] Preferably, the S1 specifically includes:

[0029] The parameter acquisition module is used to collect the physical parameters of the crane in real time. The collected physical parameters of the crane are processed through sliding average filtering. The load judgment algorithm is introduced to obtain the physical load of the crane and then judge the physical load status of the crane.

[0030] Preferably, in said S1, it further includes:

[0031] Analyze the historical working physical parameters of the crane and the historical parameters of the lifting environment, select the features most relevant to the physical load of the crane, form a feature set, expand the feature set through feature construction to obtain an extended feature set, organize the extended feature set, label encode the organized extended feature set, train the encoded feature set using linear regression, decision tree and random forest algorithms respectively, and perform weighted averaging or voting on the model results of the linear regression, decision tree and random forest algorithms to obtain a potential association model.

[0032] Preferably, the S2 specifically includes:

[0033] The potential association model is used to predict the physical load of the crane. The weight coefficient α is introduced to balance the actual physical load data and the load data predicted by the model. At the same time, the potential association model is updated in real time to obtain the real-time working status of the crane.

[0034] Preferably, the S4 specifically includes:

[0035] A network topology is constructed based on the decision groups, and the network topology is viewed as a graph, where each node represents a working state and the edges represent potential transitions from one state to another.

[0036] Preferably, in said S4, it further includes:

[0037] Calculate the correlation between the real-time working state of the crane and each working state in the decision group, and based on the correlation, select the working states with correlation greater than a threshold μ to form a first matching set; μ is obtained by an empirical method; based on the first matching set, obtain the first decision matching set by searching the grid; combine the environmental parameter prediction results, calculate the probability of matching between the real-time working state of the crane and the first decision matching set, and select the decision with the highest probability of matching as the best response decision, and finally, generate an operation instruction according to the best response decision.

[0038] Beneficial effects:

[0039] The multiple technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0040] 1. This application collects the physical parameters and environmental parameters of the crane in real time, and combines it with the mining and analysis of historical data to obtain the working status and environmental conditions of the crane in real time and accurately, which helps to improve the safety and efficiency of crane operation. Based on the potential association model, the system can predict the future working status and environmental conditions of the crane, so as to make decisions in advance and avoid possible risks.

[0041] 2. This application can provide the real-time working status of the crane by combining the actual physical load data and the load data predicted by the model, thereby ensuring the safety and efficiency of the lifting operation. The introduction of weight coefficients enables the system to automatically adjust according to the real-time changes in environmental parameters, thereby improving the robustness and adaptability of the model.

[0042] 3. This application utilizes decision tree algorithms and network topology structures to better understand the transitions between the working states of the crane under different conditions, thereby providing a more accurate decision-making basis for the decision-making module. By combining real-time working status, environmental parameter prediction results and historical decision groups for matching processing, it can quickly and accurately provide the crane with the best response decision, thereby ensuring the safety and efficiency of the lifting operation.

[0043] 4. The technical solution of the present application can effectively solve the technical problems of inaccurate and untimely control. By collecting the physical parameters and environmental parameters of the crane in real time and combining the mining and analysis of historical data, the working status and environmental conditions of the crane can be obtained in real time and accurately, which helps to improve the safety and efficiency of the crane operation. Based on the potential association model, the system can predict the future working status and environmental conditions of the crane, so as to make decisions in advance and avoid possible risks. By combining the actual physical load data and the load data predicted by the model, the technology can provide the real-time working status of the crane, so as to ensure the safety and efficiency in the lifting operation. The introduction of weight coefficients enables the system to automatically adjust according to the real-time changes of environmental parameters, thereby improving the robustness and adaptability of the model. By using the decision tree algorithm and network topology, the technology can better understand the transfer between the working states of the crane under different conditions, thereby providing a more accurate decision-making basis for the decision-making module. By matching the real-time working status, environmental parameter prediction results and historical decision groups, it can quickly and accurately provide the best response decision for the crane, thereby ensuring the safety and efficiency of the lifting operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a module diagram of a load control system based on crane hoisting described in this application;

[0045] Figure 2 This is a flow chart of a load control method based on crane hoisting described in this application. DETAILED DESCRIPTION

[0046] The embodiment of the present application provides a load control system and method based on crane hoisting, which solves the technical problems of inaccurate and untimely control in the prior art. The overall idea is as follows:

[0047] First, the parameter data is collected and processed, and the historical parameter data is mined and analyzed to obtain the potential correlation model between the parameter data; then, data analysis is performed based on the processed parameter data results and the potential correlation model between the parameter data to obtain the real-time working status of the crane; future environmental changes are predicted based on the historical parameter data of the hoisting environment to obtain the prediction results of the hoisting environment parameters, and provide environmental data support for the decision-making module; a training model is constructed to train the historical working data and historical response decisions of the crane to generate a decision group for the working status of the crane, and the decision group obtained based on the real-time working status of the crane and the environmental parameter prediction results combined with the decision support module is matched to obtain the best response decision, and an operation instruction is generated according to the best response decision and transmitted to the control module; based on the operation instruction, the operating parameters of the crane are automatically adjusted to achieve adaptive control; by real-time collection of the physical parameters and environmental parameters of the crane, combined with historical data Mining analysis can obtain the working status and environmental conditions of the crane in real time and accurately, which helps to improve the safety and efficiency of crane operation. Based on the potential association model, the system can predict the future working status and environmental conditions of the crane, so as to make decisions in advance and avoid possible risks. By combining the actual physical load data and the load data predicted by the model, this technology can provide the real-time working status of the crane, thereby ensuring the safety and efficiency of the lifting operation. The introduction of weight coefficients enables the system to automatically adjust according to the real-time changes in environmental parameters, thereby improving the robustness and adaptability of the model. Using the decision tree algorithm and network topology, this technology can better understand the transition between the working states of the crane under different conditions, thereby providing a more accurate decision-making basis for the decision-making module. By combining the real-time working status, environmental parameter prediction results and historical decision groups for matching processing, it can quickly and accurately provide the best response decision for the crane, thereby ensuring the safety and efficiency of the lifting operation.

[0048] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0049] See attached Figure 1 The load control system based on crane hoisting described in this application includes the following parts:

[0050] Parameter acquisition module, load detection module, environment perception module, mining module, data analysis module, intelligent prediction module, decision support module, decision module, control module, user interface, feedback module;

[0051] The parameter acquisition module acquires the physical parameters of the crane to provide a data basis for the load detection module; the physical parameters include lifting weight, boom angle, and boom length;

[0052] The load detection module calculates the physical load of the crane using a load judgment algorithm based on the physical parameters obtained by the parameter acquisition module, monitors the physical load of the crane in real time, and outputs the real-time physical load data of the crane;

[0053] The environmental perception module obtains the real-time parameters of the hoisting environment, and performs data processing on the obtained real-time parameters to obtain accurate environmental parameter data to monitor the changes in the hoisting environment in real time and provide key environmental parameter data for data processing; the environmental parameters are such as wind speed, temperature, and humidity;

[0054] The mining module mines and obtains a potential correlation model between the working physical parameters and the environmental parameters based on the historical working physical parameters of the crane and the historical parameters of the hoisting environment, thereby providing a basis for the data analysis module;

[0055] The data analysis module performs data analysis based on the real-time data of the load detection module and the environment perception module in combination with the potential association model obtained by the mining module to obtain the real-time working status of the crane;

[0056] The intelligent prediction module predicts the environment based on historical environmental data to obtain a prediction function, and predicts the environment based on real-time environmental parameters to obtain environmental parameter prediction results;

[0057] The decision-making auxiliary module trains and studies the crane's historical working data and historical response decisions by building a model to obtain a decision group for all working states, providing a matching basis for the decision-making module; the historical working data includes the crane's historical working physical parameter data and the hoisting environment historical parameter data;

[0058] The decision module matches the best response decision based on the real-time working status of the crane obtained by the data analysis module, the environmental parameter prediction results obtained by the intelligent prediction module, and the decision group obtained by the decision support module, and generates an operation instruction according to the best response decision;

[0059] The control module automatically adjusts the operating parameters of the crane according to the operating instructions of the decision module to achieve adaptive control of the crane wind power wall load;

[0060] The user interface provides real-time feedback and guidance to the operator based on the operation instructions of the decision module and the actual operation status of the crane;

[0061] The feedback module performs detection based on the output results of each module, and when an unsafe state is detected, a safety warning is quickly fed back to the operator through the user interface.

[0062] Refer to the attached Figure 2The load control method based on crane hoisting described in this application comprises the following steps:

[0063] S1. Collect parameter data in real time, process the collected crane physical parameter data and hoisting environment parameters, calculate the physical load of the crane using a load judgment algorithm, and mine and analyze the crane's historical working physical parameter data and hoisting environment historical parameter data, and introduce a potential correlation model between the crane's working physical parameters and hoisting environment parameters;

[0064] The parameter acquisition module is used to collect the physical parameters of the crane in real time, and the physical parameters are transmitted to the load detection module, and the physical load of the crane is calculated using the load judgment algorithm to provide the data analysis module with real-time physical load data support of the crane;

[0065] Due to the measurement error, the collected crane physical parameters are processed by sliding average filtering to obtain accurate physical parameters, which provide accurate physical parameter data for calculating the crane load;

[0066] The load judgment algorithm is further used to analyze the physical parameters of the crane, obtain the physical load of the crane, and then judge the physical load state of the crane, specifically:

[0067] First, the crane is divided into m parts, such as the top of the boom, the middle of the boom, and the bottom of the boom, using the empirical method to calculate the dynamic response coefficient D of any part. i and friction coefficient M i The dynamic response coefficient and friction coefficient can be obtained from the technical manual of the crane or experimental measurement; the physical load state of the crane is calculated by combining the physical parameter data of the crane, and the formula is as follows:

[0068]

[0069] Where F(i) represents the physical load state of the i-th part of the crane; L represents the maximum length of the boom; W(x) represents the load when the boom length is x; A(x) represents the angle between the boom and the vertical direction when the boom length is x; M i represents the i-th element in the friction coefficient set, which represents the friction coefficient of a certain part of the crane; D i represents the i-th element of the dynamic response coefficient set, which indicates the response of a certain part of the crane to the dynamic load; R represents the radius of the crane; T represents the influence coefficient of temperature on the effective working radius, obtained through experiments; Θ represents the ambient temperature; ∈ represents the influence coefficient of operating efficiency and working radius; η represents the operating efficiency of the crane; It indicates the influence of the hanging weight on the hanging rod; D i M iIt represents the load generated by the dynamic response and friction of a certain part of the crane, taking into account the dynamic characteristics and friction of the crane, and is another important factor affecting the physical load of the crane; (R+T·Θ-(1-η·∈)) represents the effective working radius of the crane, taking into account the radius of the crane, the influence of temperature on the radius and the influence of operating efficiency, and is an adjustment factor used to adjust the value of the physical load; in the above formula, after considering the influence of the hoisting weight on the boom and the dynamic response and friction load of a certain part of the crane, the physical load state of a certain part of the crane is obtained after adjusting the effective working radius; for other parts other than the boom, the physical load state of the crane is obtained by consulting relevant information, and finally the physical load state F of the crane is obtained;

[0070] Using the environmental perception module, the hoisting environmental parameters are collected in real time, and the hoisting environmental parameters are processed to obtain accurate data, providing environmental data support for the data analysis module;

[0071] First, using the environmental perception module, first collect the hoisting environmental parameters in real time, such as wind speed, temperature, humidity, air pressure, etc., and perform preliminary data cleaning on the collected raw data, such as removing obvious outliers. In the process of data processing, data loss or incomplete data may be encountered. In order to solve the problem of data loss, data interpolation technology, such as linear interpolation, polynomial interpolation or time series-based prediction model, is used to estimate the missing data points; in order to improve the accuracy of the data, filtering technology, such as Kalman filtering, median filtering and other methods are used to reduce the noise in the data; when applying the above interpolation and filtering technologies, the time delay problem of data will be introduced. In order to solve the time delay problem of data, parallel computing or optimization algorithms are used to accelerate the speed of data processing. In addition, consider using more efficient hardware or performing partial preprocessing at the data acquisition end to reduce the computing burden of the control system; after the above processing and optimization, accurate and cleaned environmental data are obtained. These data will provide key environmental data support for the data analysis module, thereby helping the crane wind power wall load control system to make more accurate decisions.

[0072] The mining module is used to analyze the historical working physical parameters of the crane and the historical parameters of the hoisting environment, identify the potential correlation model between the working physical parameters and the environmental parameters, and provide an analysis basis for the data analysis module;

[0073] Clean the historical working physical parameters of the crane and the historical parameters of the hoisting environment, process missing values ​​and outliers, obtain accurate historical data, use correlation coefficients, mutual information or other statistical methods to select the features most relevant to the crane load, and form a feature set; expand the feature set based on feature construction, such as: polynomial features, if the original features are x and y, then x can be constructed 2 ,y 2, xy and other features; interactive features, such as considering the interaction between two or more features, for example, the interaction between temperature and humidity may affect the load of the crane; feature construction based on domain knowledge: new features can be constructed based on the understanding of crane operation and physical principles; time series features, if the data is time series data, past data can be considered as new features, such as considering the historical load data of the crane, using the average load of the past 3 hours, the maximum load of the past 24 hours, etc. as new features; transformation features, performing some mathematical transformation on the original features, such as logarithmic, square root or exponential transformation, such as considering the load data of the crane, if its distribution is skewed, consider using logarithmic transformation log(·) to make the data closer to normal distribution; obtain an extended feature set, use principal component analysis to reduce the dimension of features based on the extended feature set, obtain the sorted feature set, and then perform label encoding on the sorted feature set to facilitate model training, use linear regression, decision tree and random forest algorithms to train single models respectively on the encoded feature set, and then perform weighted average or voting on their model results to obtain the final association model, the formula is as follows:

[0074] In order to facilitate understanding and highlight the importance of certain variables, some parameters are represented separately, including w for hoisting weight, θ for boom angle, l for boom length, v for wind speed, t for temperature, h for humidity, p for air pressure, d for boom diameter, s for crane speed, and other identified working physical parameters and environmental parameters can be represented as sets X and Y;

[0075] Further, using

[0076]

[0077] Where x represents other working physical parameters, y represents other environmental parameters, P(t|h) represents the conditional probability of temperature t at a given humidity h; w×l represents the product of the lifting weight and the length of the boom, describing the load correlation coefficient of the crane; e (θ×v) The index representing the product of boom angle and wind speed describes the effect of wind speed on boom angle, that is, the effect of wind force on the crane; p+d represents the combined effect of air pressure and boom diameter on crane stability; e (d×s) An exponent representing the product of boom diameter and crane speed, describing the dynamic response coefficient of the crane; Indicates the relationship between other working physical parameters and temperature and humidity and the influence of crane speed;∑ y∈Y y×e (d×s) Indicates the relationship between other environmental parameters and boom diameter and crane speed;

[0078] Finally, the potential correlation model between the working physical parameters and the environmental parameters is obtained, which provides an analysis basis for the data analysis module;

[0079] The latent association model Model describes the relationship between the working physical parameters and the environmental parameters, and predicts the physical load of the crane;

[0080] This application collects the physical parameters and environmental parameters of the crane in real time, and combines it with the mining and analysis of historical data to obtain the working status and environmental conditions of the crane in real time and accurately, which helps to improve the safety and efficiency of crane operation. Based on the potential association model, the system can predict the future working status and environmental conditions of the crane, so as to make decisions in advance and avoid possible risks.

[0081] S2. Perform data analysis on the physical load of the crane, the working physical parameters of the crane, the hoisting environment parameter data and the potential correlation model, and at the same time, update the potential correlation model in real time to obtain the real-time working status of the crane;

[0082] Based on the physical load of the crane, the working physical parameters of the crane, the hoisting environmental parameter data, and the potential correlation model between the working physical parameters and the environmental parameters, data analysis is performed to obtain the real-time working status of the crane. The specific process is as follows:

[0083] The potential association model Model is used to predict the physical load of the crane. In order to cope with the situation that the actual physical load data of the crane may be inconsistent with the load data predicted by the model and the variability of the lifting environment parameters, a weight coefficient α is introduced to balance the actual physical load data and the load data predicted by the model. The formula is as follows:

[0084]

[0085] Where n represents the dimension of the environmental parameter set; i represents the i-th environmental parameter; d i Indicates that i The relevant weight coefficient is obtained by fitting the test data; α represents the weight coefficient, which is used to balance the actual physical load and the load predicted by the model, and its value is between 0 and 1;

[0086] At the same time, the potential correlation model Model is updated in real time to adapt to the rapid changes in environmental parameters and obtain the real-time working status F of the crane. realtime , the formula is as follows:

[0087] F realtime =α×F+(1-α)×Model

[0088] Where F represents the physical load state of the crane;

[0089] The real-time working status is sent to the decision-making module to provide data basis for the decision-making module.

[0090] By combining actual physical load data with load data predicted by the model, this application can provide the real-time working status of the crane, thereby ensuring safety and efficiency in lifting operations. The introduction of weight coefficients enables the system to automatically adjust according to real-time changes in environmental parameters, thereby improving the robustness and adaptability of the model.

[0091] S3. Predict future environmental changes based on historical parameter data of the hoisting environment, introduce a deep learning neural network model, and obtain the prediction results of the hoisting environment parameters;

[0092] First, the lifting environment historical parameter data is cleaned and standardized to obtain accurate lifting environment historical parameter data D. The accurate lifting environment historical parameter data set D is trained and the preprocessed data set is divided into a training set and a validation set. For example, 70% of the data is used as the training set D. train , 30% as validation set D valid , use the deep learning neural network model to train the training set, specifically:

[0093] Input layer: input node number and training set D train The number of elements is the same as that of the training set, and the training set is used as the input layer, and then linear combination is performed.

[0094] Z=WD train +b

[0095] Passed to the hidden layer, the weight W and bias b can be initialized using random numbers generated by normal distribution;

[0096] Hidden layer: There can be multiple hidden layers. The number of nodes in each layer can be defined according to empirical requirements. Z is used as input, and the output R is obtained through linear transformation and activation function processing, which is passed to the next layer. Common activation functions include ReLU, Tanh and Sigmoid.

[0097] Dropout layer: Take the output R of the hidden layer as input, randomly turn off some neurons, and obtain the data D(R) after some neurons are turned off, where D(·) represents the Dropout operation, and transmit the data D(R) to the output layer and loss function;

[0098] Output layer: The output of the Dropout layer is used as input, and after linear transformation and activation function processing, the prediction results and prediction model are obtained. The number of output nodes is the same as the target number of the prediction task. For regression tasks, there is usually only one output node.

[0099] Define the loss function to calculate the difference between the predicted and true labels, using the cross entropy loss function, specifically:

[0100]

[0101] Among them, y represents the true value, represents the predicted value, λ represents the regularization coefficient, and W represents the weight; the gradient of the model parameters is calculated according to the loss value Loss, and the parameters are updated;

[0102] Dropout is only used during training. During validation or testing, all neurons are active, but the output of each neuron is multiplied by (1-Dropout ratio) to keep the active part of the network unchanged;

[0103] Finally, the environmental parameter prediction function f(D) is obtained to realize the prediction of the hoisting environmental parameters and obtain the hoisting environmental parameter prediction result E;

[0104] After the training is completed, the model is validated based on the validation data set D valid , the existing model validation method is used to evaluate the performance of the model using the validation set to verify the accuracy and practicality of the model.

[0105] S4. Obtain the decision group of the crane working state through the decision tree algorithm, build a network topology structure based on the decision group of the crane working state, and perform matching processing based on the real-time working state of the crane and the prediction results of environmental parameters combined with the decision group of the crane working state to obtain the best response decision;

[0106] The decision tree algorithm is used to train the crane's historical working data and historical response decisions to obtain the decision group of the crane's working status;

[0107] Furthermore, the decision group obtained by the decision support module is matched based on the real-time working status of the crane and the prediction results of environmental parameters to obtain the best response decision, which includes:

[0108] A network topology is constructed based on the decision group. The network topology can be viewed as a graph, where each node represents a working state and an edge represents a possible transition from one state to another. The network topology can better understand the transition between the working states of the crane under different conditions;

[0109] Based on the historical working status records of the crane, for each working status s i , calculate its frequency of occurrence in historical working status data:

[0110]

[0111] Based on the above calculated frequencies, a weight is assigned to each working state, and the weight is calculated as follows:

[0112] w(s i)=log(1+f(s i ))

[0113] Among them, w(s i ) indicates working status s i The weight of f(s i ) indicates working status s i The frequency of occurrence;

[0114] Update the calculated weights to the corresponding nodes of the network. Each node in the network has a weight corresponding to its importance or frequency of occurrence.

[0115] First matching process:

[0116] Based on the real-time working status of the crane realtime and each work state s in the decision group i , calculate the correlation between them, the formula is as follows:

[0117]

[0118] Among them, N represents the dimension of working status, F realtime,j The jth element representing the real-time working status, s i,j represents the jth element of the i-th working state in the decision group, and represents the mean;

[0119] Based on the correlation calculated above, the working states with correlation greater than a threshold μ are selected to form a first matching set Match1; μ is obtained by an expert through an empirical method;

[0120] Based on the first matching set obtained by the correlation between the real-time working state and each working state in the decision group, the first decision matching set Match is obtained by searching the grid. d ;

[0121] Second matching process:

[0122] Combined with the environmental parameter prediction results, calculate the real-time working status F of the crane realtime Match the first decision set d The probability of matching:

[0123]

[0124] Among them, d i Represents the first decision matching set Match d Any element in P(F realtime ,E|d i ) indicates that after choosing decision d iUnder the condition of realtime and environmental parameters predict the probability of outcome E by estimating from historical data, looking at the past choices of d i How many times have F been observed? realtime and E; P(d i ) represents decision d i The prior probability of choosing d i The probability of choosing d in all cases is calculated by estimating it from historical data. i frequency; P(F realtime ,E) indicates that the real-time working status F is observed realtime and environmental parameters predict the unconditional probability of the result E. This can be done by statistically analyzing the historical data F realtime It is estimated by the frequency of co-occurrence with E;

[0125] Through all possible decisions d i Calculate the probability of fit P(d i |F realtime ,E), and select the decision with the highest probability of matching as the best response decision:

[0126]

[0127] d best The decision with the highest probability of matching is selected as the best response decision;

[0128] Finally, the operation instructions are generated according to the best response decision and transmitted to the control module;

[0129] This application utilizes decision tree algorithms and network topology structures to better understand the transitions between the working states of the crane under different conditions, thereby providing a more accurate decision-making basis for the decision-making module. By combining real-time working status, environmental parameter prediction results and historical decision groups for matching processing, it can quickly and accurately provide the crane with the best response decision, thereby ensuring the safety and efficiency of the lifting operation.

[0130] S5. Automatically adjust the operating parameters of the crane based on the operating instructions generated by the best response decision;

[0131] Based on the operation instructions generated by the above optimal response decision, the operation parameters of the crane are automatically adjusted to achieve adaptive control of the wind power wall load of the crane;

[0132] During system operation, each module is tested, and when an unsafe state is detected, a safety warning is quickly fed back to the operator through the user interface.

[0133] In summary, a load control system and method based on crane hoisting described in this application are completed.

[0134] The technical solutions in the above embodiments of the present application have at least the following technical effects or advantages:

[0135] 1. This application collects the physical parameters and environmental parameters of the crane in real time, and combines it with the mining and analysis of historical data to obtain the working status and environmental conditions of the crane in real time and accurately, which helps to improve the safety and efficiency of crane operation. Based on the potential association model, the system can predict the future working status and environmental conditions of the crane, so as to make decisions in advance and avoid possible risks.

[0136] 2. This application can provide the real-time working status of the crane by combining the actual physical load data and the load data predicted by the model, thereby ensuring the safety and efficiency of the lifting operation. The introduction of weight coefficients enables the system to automatically adjust according to the real-time changes in environmental parameters, thereby improving the robustness and adaptability of the model.

[0137] 3. This application utilizes decision tree algorithms and network topology structures to better understand the transitions between the working states of the crane under different conditions, thereby providing a more accurate decision-making basis for the decision-making module. By combining real-time working status, environmental parameter prediction results and historical decision groups for matching processing, it can quickly and accurately provide the crane with the best response decision, thereby ensuring the safety and efficiency of the lifting operation.

[0138] Effect research:

[0139] The technical solution of the present application can effectively solve the technical problems of inaccurate and untimely control. In addition, the above-mentioned system or method has undergone a series of effect surveys. By collecting the physical parameters and environmental parameters of the crane in real time and combining the mining and analysis of historical data, the working status and environmental conditions of the crane can be obtained in real time and accurately, which helps to improve the safety and efficiency of the crane operation. Based on the potential association model, the system can predict the future working status and environmental conditions of the crane, so as to make decisions in advance and avoid possible risks; by combining the actual physical load data and the load data predicted by the model, the technology can provide the real-time working status of the crane, so as to ensure the safety and efficiency in the lifting operation. The introduction of weight coefficients enables the system to automatically adjust according to the real-time changes in environmental parameters, thereby improving the robustness and adaptability of the model; using the decision tree algorithm and network topology, the technology can better understand the transfer between the working states of the crane under different conditions, thereby providing a more accurate decision basis for the decision module, and by matching the real-time working status, the environmental parameter prediction results and the historical decision group for matching, it can quickly and accurately provide the best response decision for the crane, thereby ensuring the safety and efficiency of the lifting operation.

[0140] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0141] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0142] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0143] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A load control system based on crane hoisting, characterized in that: Includes the following parts: Parameter acquisition module, load detection module, environment perception module, mining module, data analysis module, intelligent prediction module, decision support module, decision module, control module, user interface, feedback module; The parameter acquisition module is used to obtain the physical parameters of the crane, including the lifting weight, boom angle, and boom length, to provide a data basis for the load detection module; The load detection module is used to calculate the physical load of the crane using the load judgment algorithm, monitor the physical load of the crane in real time, and output the real-time physical load data of the crane; analyze the physical parameters of the crane using the load judgment algorithm, calculate the physical load of the crane and then judge the physical load state of the crane. The specific calculation is as follows: Divide the crane into m parts using the expert experience method and calculate the dynamic response coefficient of any part and friction coefficient , combined with the crane physical parameter data to calculate the physical load state of the crane, the formula is as follows: ， in, Indicates the crane's The physical load status of the part; Indicates the maximum length of the boom; Indicates that the boom length is The lifting weight at the time; Indicates that the boom length is The angle between the boom and the vertical direction when ; represents the first elements; The first elements; Indicates the radius of the crane; Indicates the influence coefficient of temperature on the effective working radius; Indicates the ambient temperature; Indicates the influence coefficient of operating efficiency and working radius; Indicates the operating efficiency of the crane; It represents the effect of the hanging weight on the boom; Indicates the dynamic response of a part of the crane and the load caused by friction; Indicates the effective working radius of the crane; The environmental perception module is used to obtain and process real-time parameters of the hoisting environment, obtain environmental parameter data, including wind speed, temperature, humidity, and monitor changes in the hoisting environment in real time; The mining module is used to obtain a potential correlation model between the working physical parameters and the environmental parameters based on the historical working physical parameters of the crane and the historical parameters of the hoisting environment, so as to provide a basis for the data analysis module; The data analysis module is used to analyze the real-time data based on the load detection module and the environment perception module, and combine the potential association model of the mining module to obtain the real-time working status of the crane; The intelligent prediction module is used to predict the environment through the environmental parameter prediction function based on the real-time environmental parameters to obtain the environmental parameter prediction results; The decision-making auxiliary module is used to train the historical working data and historical response decisions of the crane through a decision tree algorithm to obtain a decision group of the crane's working status and provide a matching basis for the decision-making module; the historical working data includes the historical working physical parameter data of the crane and the historical parameter data of the hoisting environment; The decision module is used to construct a network topology structure based on the real-time working state of the crane obtained by the data analysis module, the environmental parameter prediction results obtained by the intelligent prediction module, and the decision group of the working state of the crane obtained by the decision assistance module, perform matching processing to obtain the best response decision, and generate an operation instruction according to the best response decision; The control module is used to automatically adjust the operating parameters of the crane according to the operating instructions of the decision module; The feedback module is used to detect the output results of each module, and when an unsafe state is detected, a safety warning is fed back to the operator through the user interface; The user interface is used to provide real-time feedback and guidance to the operator based on the operating instructions of the decision-making module and the actual operating status of the crane.

2. A load control method based on crane hoisting, characterized in that: The following steps are involved: S1. Collect parameter data in real time, process the collected physical parameter data of the crane and the hoisting environment parameters, calculate the physical load of the crane using a load judgment algorithm, and mine and analyze the historical working physical parameter data of the crane and the historical parameter data of the hoisting environment, and introduce a potential correlation model between the working physical parameters of the crane and the hoisting environment parameters; S2. Perform data analysis on the physical load of the crane, the working physical parameters of the crane, the lifting environment parameter data and the potential correlation model, and at the same time, update the potential correlation model in real time to obtain the real-time working status of the crane; S3. Predict future environmental changes based on historical parameter data of the hoisting environment, introduce a deep learning neural network model, and obtain the prediction results of the hoisting environment parameters; S4. Obtain the decision group of the crane working state through the decision tree algorithm, build a network topology structure based on the decision group of the crane working state, and perform matching processing based on the real-time working state of the crane and the prediction results of environmental parameters combined with the decision group of the crane working state to obtain the best response decision; S5. Automatically adjust the operating parameters of the crane based on the operating instructions generated by the best response decision; Applicable to a load control system based on crane hoisting as described in claim 1.

3. A load control method based on crane hoisting according to claim 2, characterized in that: The S1 specifically includes: The parameter acquisition module is used to collect the physical parameters of the crane in real time. The collected physical parameters of the crane are processed through sliding average filtering. The load judgment algorithm is introduced to obtain the physical load of the crane and then judge the physical load status of the crane.

4. A load control method based on crane hoisting according to claim 2, characterized in that: In the S1, it also includes: Analyze the historical working physical parameters of the crane and the historical parameters of the lifting environment, select the features most relevant to the physical load of the crane, form a feature set, expand the feature set through feature construction to obtain an extended feature set, organize the extended feature set, label encode the organized extended feature set, train the encoded feature set using linear regression, decision tree and random forest algorithms respectively, and perform weighted averaging or voting on the model results of the linear regression, decision tree and random forest algorithms to obtain a potential association model.

5. The load control method based on crane hoisting according to claim 2, characterized in that: The S2 specifically includes: Using the latent association model to predict the physical load of the crane, introducing the weight coefficient To balance the actual physical load data and the load data predicted by the model, and at the same time update the potential associated model in real time to obtain the real-time working status of the crane.

6. A method for controlling loads based on crane hoisting according to claim 2, characterized in that: The S4 specifically includes: A network topology is constructed based on the decision groups, and the network topology is viewed as a graph, where each node represents a working state and the edges represent potential transitions from one state to another.

7. A method for controlling loads based on crane hoisting according to claim 6, characterized in that: In said S4, it also includes: Calculate the correlation between the real-time working state of the crane and each working state in the decision group, and based on the correlation, select the working state with a correlation greater than a threshold. The working status constitutes a first matching set; Obtained by empirical method; based on the first matching set, the first decision matching set is obtained by searching the grid; combined with the environmental parameter prediction results, the probability of matching between the real-time working status of the crane and the first decision matching set is calculated, and the decision with the highest probability of matching is selected as the best response decision, and finally, an operation instruction is generated according to the best response decision.

Images