CN111439681A - Intelligent identification method and system for unsafe operation based on tower crane - Google Patents

Intelligent identification method and system for unsafe operation based on tower crane Download PDF

Info

Publication number
CN111439681A
CN111439681A CN202010055629.1A CN202010055629A CN111439681A CN 111439681 A CN111439681 A CN 111439681A CN 202010055629 A CN202010055629 A CN 202010055629A CN 111439681 A CN111439681 A CN 111439681A
Authority
CN
China
Prior art keywords
hoisting
unsafe
behaviors
intelligent
response data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010055629.1A
Other languages
Chinese (zh)
Other versions
CN111439681B (en
Inventor
蒋伟光
丁烈云
骆汉宾
周诚
刘振华
袁硕
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN202010055629.1A priority Critical patent/CN111439681B/en
Publication of CN111439681A publication Critical patent/CN111439681A/en
Application granted granted Critical
Publication of CN111439681B publication Critical patent/CN111439681B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B66—HOISTING; LIFTING; HAULING
    • B66C—CRANES; LOAD-ENGAGING ELEMENTS OR DEVICES FOR CRANES, CAPSTANS, WINCHES, OR TACKLES
    • B66C15/00—Safety gear
    • B66C15/06—Arrangements or use of warning devices
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B66—HOISTING; LIFTING; HAULING
    • B66C—CRANES; LOAD-ENGAGING ELEMENTS OR DEVICES FOR CRANES, CAPSTANS, WINCHES, OR TACKLES
    • B66C13/00—Other constructional features or details
    • B66C13/16—Applications of indicating, registering, or weighing devices
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B66—HOISTING; LIFTING; HAULING
    • B66C—CRANES; LOAD-ENGAGING ELEMENTS OR DEVICES FOR CRANES, CAPSTANS, WINCHES, OR TACKLES
    • B66C15/00—Safety gear
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mechanical Engineering (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Jib Cranes (AREA)
  • Control And Safety Of Cranes (AREA)

Abstract

The invention discloses an intelligent identification system for unsafe operation behaviors of a tower crane and a working method thereof. The system mainly comprises a hoisting real-time response data acquisition module, a hoisting unsafe behavior intelligent identification module and an identification result visualization module. The data acquisition module realizes real-time acquisition of the mechanical response data of the tower crane structure. The unsafe behavior identification module identifies unsafe operation behaviors through the collected data, wherein the key point is the construction of an intelligent identification model of the unsafe operation behaviors, and the mapping relation between mechanical response data of the hoisting structure and the unsafe operation behaviors is established through an intelligent learning algorithm. And the hoisting identification result visualization module is used for visually displaying and early warning the final identification result. The intelligent identification system and the working method provided by the invention can accurately early warn the unsafe operation behavior of the tower crane and correct the violation at the root of the accident, thereby optimizing the field safety management process of the tower crane.

Description

Intelligent identification method and system for unsafe operation based on tower crane
Technical Field
The invention belongs to the field of hoisting operation safety control, and particularly relates to an unsafe operation intelligent identification method and system based on a tower crane.
Background
The tower crane is a key device for engineering construction and transportation activities, and simultaneously, the market scale of the tower crane industry is continuously expanded along with the popularization of a new process of an assembly type building. However, in the engineering construction process, the production accidents of the tower crane are not effectively controlled all the time, and the problem which puzzles the development of the construction industry at home and abroad is solved. The reason is that the unsafe operation behavior of people is a fire fuse of a hoisting accident, and once illegal operations such as overload, inclined hoisting, hoisting under a strong wind condition and the like occur, the safety accident that the whole tower crane is unstable is easily caused.
At present, most researches and inventions for controlling hoisting risks concentrate on early warning of blind areas and collision risks, and risk identification and control of instability and overturning of the tower crane caused by unsafe operation behaviors of people are neglected. On one hand, the traditional method of carrying out on-site safety management by means of a small number of sensors and safety personnel has large errors, and risks cannot be accurately pre-warned. On the other hand, the existing risk control thinking is mainly concentrated in the hoisting operation process, and identification of unsafe operation behaviors in the whole process including hoisting is lacked, so that the whole early warning process is not comprehensive and safe to manage in time, and the real hoisting risk prevention and control effect is difficult to achieve.
Therefore, a method for identifying unsafe hoisting operation behavior quickly and accurately is needed to improve the active control capability of field safety management.
Disclosure of Invention
Aiming at the defects or improvement requirements in the prior art, the invention provides an unsafe operation intelligent identification method and system based on a tower crane, and aims to establish a mapping relation between mechanical response data generated by hoisting behaviors and hoisting behavior categories based on an intelligent learning model, so that unsafe operation behaviors in the hoisting process of the tower crane are quickly and intelligently identified, and the occurrence of hoisting operation risk accidents is prevented.
In order to achieve the above object, according to one aspect of the present invention, there is provided a method for intelligently identifying unsafe operation based on a tower crane, comprising an offline training phase and an online detection phase, wherein:
the off-line training phase comprises:
s1, constructing a historical response state database of hoisting operation to acquire and record mechanical response data corresponding to the hoisting behavior of the tower crane; establishing a hoisting unsafe behavior classification library, and marking the mechanical response data according to the classification of safe behaviors and unsafe behaviors;
s2, constructing an intelligent identification training model for hoisting unsafe operation behaviors based on an intelligent learning algorithm, and training the intelligent identification training model for hoisting unsafe operation behaviors by using the hoisting historical response state database and the unsafe behavior classification database in the S1 to obtain the intelligent identification model for hoisting unsafe operation;
and (3) an online detection stage:
and S3, collecting real-time mechanical response data of the hoisting operation structure and inputting the real-time mechanical response data into the S2 intelligent identification model for unsafe hoisting operation, and obtaining an identification result of unsafe hoisting operation under the current hoisting working condition.
Further, the step S1 includes the following sub-steps:
s11, in a hoisting operation historical response database, normal hoisting operation comprises composite operation of hoisting, rotation, amplitude variation and/or the hoisting behaviors, through a normal hoisting experiment, sensor mechanical response data under various normal hoisting behaviors are obtained, a fluctuation diagram of the sensor mechanical response data is drawn, data of a fluctuation section are extracted from the fluctuation diagram and are added into a tower crane hoisting operation mechanical state response database as effective mechanical response data, and a corresponding relation between the mechanical response data and the normal hoisting operation behaviors is established;
s12, hoisting operation historical response database and hoisting unsafe behavior classification library, wherein the unsafe hoisting operation behaviors comprise oblique hoisting, overload hoisting, hoisting operation in strong wind environment, hoisting sudden stop, sudden start and/or composite hoisting behaviors when approaching full load; and performing unsafe hoisting operation experiments, obtaining sensor mechanical response data under various unsafe hoisting operation behaviors, drawing a fluctuation graph of the sensor mechanical response data, extracting data of a fluctuation section from the fluctuation graph as effective mechanical response data, adding the effective mechanical response data into a tower crane hoisting operation mechanical state response database, establishing one-to-one correspondence relation between the mechanical response data and various unsafe hoisting operation behaviors, and constructing a hoisting unsafe behavior classification database.
Further, in step S2, the intelligent recognition training model of unsafe operation behavior of hoisting based on intelligent learning is represented as:
[X]{θd}＝[H]
wherein, [ X ]]The method comprises the following steps of (1) mechanically responding data of a hoisting structure under a certain operation condition; [ H ]]Hoisting a behavior reaction matrix under the working condition; { theta ]dThe method comprises the steps that a parameter set to be identified in an intelligent identification training model for hoisting unsafe operation behaviors is obtained;
the intelligent identification model of unsafe operation behavior of data-driven hoisting obtained after training is expressed as follows:
[X]{θh}＝[H]
wherein, [ X ]]The mechanical response state of the hoisting structure under the real-time working condition is obtained; [ H ]]A hoisting behavior reaction matrix; { theta ]hIs { theta }dAnd (6) obtaining a stable and reliable parameter set after training.
Further, the step S2 includes the following sub-steps:
s21, establishing an intelligent identification training model for unsafe hoisting operation behaviors based on the BP neural network;
and S22, respectively taking the mechanical response data and the hoisting unsafe operation behavior types corresponding to the mechanical response data as input and output, and training the intelligent identification training model of the hoisting unsafe operation behaviors in the step S21 by adopting a gradient descent method to obtain the intelligent identification model of the hoisting unsafe operation behaviors.
Further, the step S22 includes the following sub-steps:
s221, dividing the mechanical response data into unsafe hoisting operation behaviors and safe hoisting operation behaviors, and giving labels corresponding to the mechanical response data and the hoisting behaviors one by one;
s222, initializing the BP neural network, and constructing an input layer, a hidden layer and an output layer; the weight of the ith neuron of the input layer to the jth neuron of the hidden layer is wijThe weight from the jth neuron of the hidden layer to the kth neuron of the output layer is wjkThe bias of the input layer to the hidden layer is set to ajBias of output layer set as bkThe learning rate is η, and the excitation function g (x) is a Sigmoid function in the form:
Figure BDA0002372696720000041
wherein x is input data of a neuron;
s223, inputting the mechanical response data into a BP neural network to obtain the output of each neuron, wherein the hidden layer neuron output calculation formula is as follows:
Figure BDA0002372696720000042
wherein x isiResponding data for the input layer structure;
Hjthe output of each neuron node of the hidden layer;
the output calculation formula of the output layer is as follows:
Figure BDA0002372696720000043
wherein, OkIs the output of the output layer neuron node;
s224, the inverse transmission error function is taken as:
Figure BDA0002372696720000044
wherein, YkFor the desired output of the output layer, OkThe actual output of the output layer is the neuron error of the output layer;
s225, calculating the error between the output of the last layer and the real result, and reversely transmitting the error to the preceding neuron to obtain the error of each neuron, so that the error is reversely transmitted to the first layer, wherein the hidden layer error calculation formula is as follows:
Figure BDA0002372696720000045
wherein the content of the first and second substances,jthe jth neuron error of the hidden layer;
koutput layer neuron errors;
the input layer error calculation formula is:
Figure BDA0002372696720000051
wherein the content of the first and second substances,iis the error value of the ith neuron of the input layer;
s226, after the error of each neuron is obtained, correcting the weight by adopting a gradient descent method to obtain more accurate transfer weight, wherein the calculation formula of the correction weight is as follows:
Figure BDA0002372696720000052
wherein, w'ijModifying weights for input layer neurons to hidden layer neurons;
w′jk(k ═ 1) is the revised weight from hidden layer neurons to output layer neurons;
and continuously training by adopting the corrected model until the training result is converged.
Further, in step S3, when an unsafe operation behavior is identified, an early warning message is issued and recorded.
To achieve the above object, the present invention also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method as described in any of the preceding claims.
In order to achieve the above object, the present invention further provides an intelligent identification system for unsafe operation based on a tower crane, which includes the computer readable storage medium and a processor, wherein the processor is used for calling and processing the computer program stored in the computer readable storage medium.
In order to achieve the above object, the present invention further provides an unsafe operation intelligent identification system based on a tower crane, including: the system comprises a hoisting real-time response data acquisition module, a hoisting unsafe behavior intelligent identification module and a hoisting identification result visualization module;
the hoisting real-time response data acquisition module comprises a real-time data acquisition sensor and a transmission submodule and is used for acquiring and uploading mechanical state response data of the hoisting machinery in hoisting operation in real time so as to be analyzed by the hoisting unsafe behavior identification module;
the intelligent identification module for the unsafe hoisting behavior comprises an intelligent identification model for the unsafe hoisting operation behavior obtained by training according to the offline training stage in the intelligent identification method for the unsafe hoisting operation, wherein the intelligent identification model is used for quickly identifying the unsafe hoisting operation behavior, and the identification result is transmitted to a visual module for the identification result;
the hoisting identification result visualization module comprises a display device and an early warning device, is used for displaying the identification result of unsafe operation behaviors in real time, and carries out early warning and/or recording when the identification result is unsafe operation.
In general, compared with the prior art, the above technical solution contemplated by the present invention can obtain the following beneficial effects:
1. according to the invention, an identification model based on data driving is constructed by adopting an intelligent learning method, the model is trained through hoisting historical response data and corresponding hoisting behaviors, and the intelligent identification model for unsafe behaviors of hoisting operation is obtained, so that illegal operations in the hoisting operation process can be accurately early warned, and a more reliable method is provided for field safety management. The corresponding system can complete monitoring and early warning on unsafe operation behaviors based on the real-time data acquisition module, the behavior identification module and the data visualization module, and realize visualization early warning on unsafe hoisting behaviors.
2. The fluctuation data of the mechanical response of the sensor can effectively and visually reflect whether the tower crane is in hoisting behavior or not and what hoisting behavior is in hoisting behavior, and the redundancy and the repeatability of the sensor data are greatly reduced by selecting the fluctuation data, so that the accuracy and the robustness are higher when the fluctuation data is selected as the effective data of the mechanical response to perform model training and behavior identification.
Drawings
FIG. 1 is a schematic diagram of an intelligent identification system for unsafe operation in hoisting of a tower crane according to the present invention;
FIG. 2 is a functional schematic of FIG. 1;
FIG. 3 is a tower crane model and corresponding sensor layout for an embodiment;
fig. 4 (a) to (d) are schematic diagrams of the classification fluctuation data of the hoisting operation mechanical state response database in the specific embodiment;
fig. 5 is a schematic flowchart of a data-driven unsafe operation behavior recognition model constructed by an intelligent learning method according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
As shown in fig. 1 and fig. 2, a schematic diagram of an intelligent identification system for unsafe operation of a tower crane provided by the present invention includes: the system comprises a hoisting real-time data acquisition module, an unsafe behavior identification module and an identification data visualization module.
The hoisting real-time response data acquisition module acquires hoisting state response data in real time through the sensor. The method comprises the steps that tension and compression data are measured by a tension and compression sensor, acceleration data are measured by an acceleration sensor, load data are measured by a tension sensor, the tension and compression data measured by the tension and compression sensor reflect the stress state of a structure, and the acceleration data measured by the acceleration sensor reflect the vibration frequency of the structure;
the hoisting instability unsafe behavior identification module adopts a BP neural network to construct an identification model through a machine learning intelligent algorithm, inputs historical hoisting mechanical state response data and corresponding unsafe hoisting behaviors for training, forms a hoisting operation unsafe behavior intelligent identification model, realizes identification of real-time hoisting mechanical state data, and obtains identification results of the hoisting operation unsafe behaviors;
the hoisting identification result visualization module visualizes the identification result through tower crane terminal display equipment, and reminds field management and operators to correct unsafe hoisting behaviors through interface early warning.
The method comprises the following steps of:
s1, as shown in figure 3, tension and compression sensors are arranged at four vertexes of a tower crane base, tension sensors are arranged at lifting ropes and used for collecting experimental data and lifting real-time data, the tension and compression sensors read the data, redundancy and repeatability of the obtained data are reduced according to screening of sensor fluctuation data, safe lifting behavior data and unsafe lifting behavior data are screened, the safe lifting behavior data and the unsafe lifting behavior data are input into a database to construct a lifting operation mechanical state response database, and a lifting unsafe behavior classification database is established according to lifting behavior records.
As shown in fig. 4, a sensor data fluctuation graph is drawn, and effective sensor data are screened according to different fluctuation rules of sensor response data corresponding to various hoisting behaviors to construct a hoisting operation mechanical state response database.
S2, constructing an intelligent identification model for identifying unsafe behaviors of tower crane hoisting operation based on machine learning, wherein the model reflects the corresponding relation between structural dynamic response data and unsafe hoisting behaviors. The method comprises the following specific steps:
s21, obtaining a model for identifying and training unsafe behaviors in tower crane hoisting operation based on a BP neural network method, carrying out data test after model training, and enabling the stabilized model to be an intelligent identification model for the unsafe behaviors in tower crane hoisting operation.
And S22, combining the lifting operation unsafe behavior identification training model obtained in the step S21 with the response state data collected by the sensor and the lifting instability unsafe operation behavior corresponding to the response state data, training the model by adopting a gradient descent method, solving a parameter set of the intelligent identification model, and completing a model parameter identification training process. The method comprises the following specific steps:
and S221, respectively arranging the historical response databases in the mechanical state response database into a training set, a verification set and a test set, further dividing the sensor data into unsafe hoisting operation behaviors and safe hoisting operation behaviors in the training set, and endowing the sensor data with labels corresponding to the hoisting behaviors one by one.
S222, initializing the BP neural network, constructing an input layer, a hidden layer and an output layer, setting the number of nodes of the input layer to be 82 according to the response data matrix of the input structure body, setting the hidden layer to be 1 layer, setting the hidden layer to be 16 nodes, and then outputting the result to the output layer; the weight of the ith neuron of the input layer to the jth neuron of the hidden layer is wijThe weight from the jth neuron of the hidden layer to the kth neuron of the output layer is wjkThe bias of the input layer to the hidden layer is set to ajBias of output layer set as bkLearning rate η, excitation function g (x), wherein excitation function g (x) takes the form of a Sigmoid function:
Figure BDA0002372696720000091
s223, inputting the structural body response data of the training set into a BP neural network to obtain the output of each neuron, wherein the hidden layer neuron output calculation formula is as follows:
Figure BDA0002372696720000092
wherein x isi(i ═ 1 … 82) is input layer structure response data;
Hj(j ═ 1 … 16) is the output of each neuron node of the hidden layer.
The output calculation formula of the output layer is as follows:
Figure BDA0002372696720000093
wherein, Ok(k ═ 1) is the output of the output layer neuron node.
S224, the inverse transmission error function is taken as:
Figure BDA0002372696720000094
wherein, Yk(k 1) is the desired output of the output layer, OkThe actual output of the output layer and the neuron error of the output layer.
S225, after the error value of the last layer is obtained through calculation, the error is reversely propagated to the preceding neuron, the error of each neuron is obtained, the error is reversely propagated to the first layer all the time, and the hidden layer error calculation formula is as follows:
Figure BDA0002372696720000095
wherein the content of the first and second substances,j(j-1 … 16) is hiddenContaining the jth neuron error;
k(k ═ 1) is the output layer neuron error.
The input layer error calculation formula is:
Figure BDA0002372696720000101
wherein the content of the first and second substances,i(i-1 … 82) is the error value for the ith neuron when input to the layer.
S226, after the error of each neuron is obtained, correcting the weight by adopting a gradient descent method to obtain more accurate transfer weight, wherein the correction calculation formula is as follows:
Figure BDA0002372696720000102
wherein, w'ij(i-1 … 82; j-1 … 16) is the revised weight of the input layer neurons to the hidden layer neurons;
w′jk(k ═ 1) is the revised weight of the hidden layer neurons to the output layer neurons.
And continuously training by adopting the corrected model until the training result is converged.
And S23, based on the lifting operation unsafe behavior intelligent identification training model obtained in the S21, bringing the data concentrated in the test into the training model for testing, comparing the error between the test result and the actual result, and curing the test result into an intelligent identification model based on a BP (back propagation) neural network when the training model achieves a stable and reliable result.
And S3, inputting the real-time response data of the local structural body collected by the sensor by using the intelligent identification model established in the S2, and identifying unsafe hoisting behaviors. By the aid of a machine learning method, the hoisting behaviors are identified in real time according to structural body mechanical response data fed back under different hoisting behaviors, and whether unsafe hoisting behaviors are performed on the tower crane is judged.
As shown in fig. 5, a flow chart of the unsafe operation behavior identification method based on machine learning according to the embodiment of the present invention is shown, and the specific steps are as follows:
s1, acquiring a large amount of tension and compression sensor data through multiple experiments, preprocessing the sensor data, selecting effective sensor data aiming at various hoisting behaviors, and forming a hoisting operation response state database;
s2, giving labels to the obtained effective sensor data, giving a '0' label to the normal hoisting behavior data, marking the unsafe hoisting behavior data in a layered mode, giving a '1' label to all the unsafe behavior data, and giving labels to different types of unsafe behavior data respectively to form a hoisting unsafe behavior classification library;
s3, dividing the data endowed with the labels in the second step into a training set, a verification set and a test set, inputting a large number of test sets into an intelligent identification training model, calculating output and output errors of a hidden layer and an output layer, adjusting the weights of the input layer, the hidden layer and the output layer according to a back propagation algorithm, bringing the test sets and the verification sets into the model for training, and ending the training when the errors reach preset requirements to form an unsafe hoisting operation behavior intelligent identification module;
and S4, the test set data is brought into the model for testing, and the accuracy of the intelligent identification module for identifying different hoisting behaviors is checked.
And S5, inputting the real-time sensor data of the tower crane into the intelligent identification system for unsafe operation behaviors of the tower crane, and monitoring the operation behaviors of the tower crane.
According to the method, unsafe hoisting behavior identification is carried out through the response state data of the local structural body in the hoisting process of the tower crane, namely, when abnormal behaviors such as overload hoisting, inclined hoisting, sudden stop and sudden start occur, the response of the structural key data reflected by the method is inconsistent with the normal state. A large amount of hoisting operation basic historical data are collected to be used as training samples, a data-driven machine training model is built, and a large amount of marked structural state data of unsafe operation behaviors are input into the built machine training model for training to obtain an intelligent identification model of the unsafe behaviors of the hoisting operation. And carrying out supervision and early warning on unsafe hoisting behaviors in subsequent hoisting operation activities.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (9)

1. An unsafe operation intelligent identification method based on a tower crane is characterized by comprising the following off-line training stage and on-line detection stage, wherein:
the off-line training phase comprises:
s1, constructing a historical response state database of hoisting operation to acquire and record mechanical response data corresponding to the hoisting behavior of the tower crane; establishing a hoisting unsafe behavior classification library, and marking the mechanical response data according to the classification of safe behaviors and unsafe behaviors;
s2, constructing an intelligent identification training model for hoisting unsafe operation behaviors based on an intelligent learning algorithm, and training the intelligent identification training model for hoisting unsafe operation behaviors by using the hoisting historical response state database and the unsafe behavior classification database in the S1 to obtain the intelligent identification model for hoisting unsafe operation;
and (3) an online detection stage:
and S3, collecting real-time mechanical response data of the hoisting operation structure and inputting the real-time mechanical response data into the S2 intelligent identification model for unsafe hoisting operation, and obtaining an identification result of unsafe hoisting operation under the current hoisting working condition.
2. The intelligent unsafe operation identification method of claim 1, wherein the step S1 comprises the following sub-steps:
s11, in a hoisting operation historical response database, normal hoisting operation comprises composite operation of hoisting, rotation, amplitude variation and/or the hoisting behaviors, through a normal hoisting experiment, sensor mechanical response data under various normal hoisting behaviors are obtained, a fluctuation diagram of the sensor mechanical response data is drawn, data of a fluctuation section are extracted from the fluctuation diagram and are added into a tower crane hoisting operation mechanical state response database as effective mechanical response data, and a corresponding relation between the mechanical response data and the normal hoisting operation behaviors is established;
s12, hoisting operation historical response database and hoisting unsafe behavior classification library, wherein the unsafe hoisting operation behaviors comprise oblique hoisting, overload hoisting, hoisting operation in strong wind environment, hoisting sudden stop, sudden start and/or composite hoisting behaviors when approaching full load; and performing unsafe hoisting operation experiments, obtaining sensor mechanical response data under various unsafe hoisting operation behaviors, drawing a fluctuation graph of the sensor mechanical response data, extracting data of a fluctuation section from the fluctuation graph as effective mechanical response data, adding the effective mechanical response data into a tower crane hoisting operation mechanical state response database, establishing one-to-one correspondence relation between the mechanical response data and various unsafe hoisting operation behaviors, and constructing a hoisting unsafe behavior classification database.
3. The intelligent unsafe operation identification method of claim 1 or 2, wherein in step S2, the intelligent identification training model for lifting unsafe operation behavior formed based on intelligent learning is represented as:
[X]{θd}＝[H]
wherein, [ X ]]The method comprises the following steps of (1) mechanically responding data of a hoisting structure under a certain operation condition; [ H ]]Hoisting a behavior reaction matrix under the working condition; { theta ]dThe method comprises the steps that a parameter set to be identified in an intelligent identification training model for hoisting unsafe operation behaviors is obtained;
the intelligent identification model of unsafe operation behavior of data-driven hoisting obtained after training is expressed as follows:
[X]{θh}＝[H]
wherein, [ X ]]The mechanical response state of the hoisting structure under the real-time working condition is obtained; [ H ]]A hoisting behavior reaction matrix; { theta ]hIs { theta }dAnd (6) obtaining a stable and reliable parameter set after training.
4. The intelligent unsafe operation identification method of claim 1 or 2, wherein the step S2 comprises the following sub-steps:
s21, establishing an intelligent identification training model for unsafe hoisting operation behaviors based on the BP neural network;
and S22, respectively taking the mechanical response data and the hoisting unsafe operation behavior types corresponding to the mechanical response data as input and output, and training the intelligent identification training model of the hoisting unsafe operation behaviors in the step S21 by adopting a gradient descent method to obtain the intelligent identification model of the hoisting unsafe operation behaviors.
5. The intelligent unsafe operation identification method of claim 4, wherein the step S22 comprises the following sub-steps:
s221, dividing the mechanical response data into unsafe hoisting operation behaviors and safe hoisting operation behaviors, and giving labels corresponding to the mechanical response data and the hoisting behaviors one by one;
s222, initializing the BP neural network, and constructing an input layer, a hidden layer and an output layer; the weight of the ith neuron of the input layer to the jth neuron of the hidden layer is wijThe weight from the jth neuron of the hidden layer to the kth neuron of the output layer is wjkThe bias of the input layer to the hidden layer is set to ajBias of output layer set as bkThe learning rate is η, and the excitation function g (x) is a Sigmoid function in the form:
Figure FDA0002372696710000031
wherein x is input data of a neuron;
s223, inputting the mechanical response data into a BP neural network to obtain the output of each neuron, wherein the hidden layer neuron output calculation formula is as follows:
Figure FDA0002372696710000032
wherein x isiResponding data for the input layer structure;
Hjthe output of each neuron node of the hidden layer;
the output calculation formula of the output layer is as follows:
Figure FDA0002372696710000033
wherein, OkIs the output of the output layer neuron node;
s224, the inverse transmission error function is taken as:
Figure FDA0002372696710000034
wherein, YkFor the desired output of the output layer, OkThe actual output of the output layer is the neuron error of the output layer;
s225, calculating the error between the output of the last layer and the real result, and reversely transmitting the error to the preceding neuron to obtain the error of each neuron, so that the error is reversely transmitted to the first layer, wherein the hidden layer error calculation formula is as follows:
Figure FDA0002372696710000035
wherein the content of the first and second substances,jthe jth neuron error of the hidden layer;
koutput layer neuron errors;
the input layer error calculation formula is:
Figure FDA0002372696710000041
wherein the content of the first and second substances,iis the error value of the ith neuron of the input layer;
s226, after the error of each neuron is obtained, correcting the weight by adopting a gradient descent method to obtain more accurate transfer weight, wherein the calculation formula of the correction weight is as follows:
Figure FDA0002372696710000042
wherein, w'ijModifying weights for input layer neurons to hidden layer neurons;
w′jk(k ═ 1) is the revised weight from hidden layer neurons to output layer neurons;
and continuously training by adopting the corrected model until the training result is converged.
6. The intelligent unsafe operation identification method of claim 1 or 2, wherein in step S3, when unsafe operation behavior is identified, an early warning message is issued and recorded.
7. A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor, implements the method of any one of claims 1 to 6.
8. An intelligent identification system for unsafe operation based on a tower crane, which is characterized by comprising the computer readable storage medium of claim 7 and a processor, wherein the processor is used for calling and processing the computer program stored in the computer readable storage medium.
9. The utility model provides an unsafe operation intelligent recognition system based on tower crane which characterized in that includes: the system comprises a hoisting real-time response data acquisition module, a hoisting unsafe behavior intelligent identification module and a hoisting identification result visualization module;
the hoisting real-time response data acquisition module comprises a real-time data acquisition sensor and a transmission submodule and is used for acquiring and uploading mechanical state response data of the hoisting machinery in hoisting operation in real time so as to be analyzed by the hoisting unsafe behavior identification module;
the intelligent identification module for the unsafe hoisting behavior comprises an intelligent identification model for the unsafe hoisting operation behavior obtained by training in the offline training stage of the intelligent identification method for the unsafe hoisting operation according to any one of claims 1 to 6, wherein the intelligent identification model is used for rapidly identifying the unsafe hoisting operation behavior, and the identification result is transmitted to a visualization module for the identification result;
the hoisting identification result visualization module comprises a display device and an early warning device, is used for displaying the identification result of unsafe operation behaviors in real time, and carries out early warning and/or recording when the identification result is unsafe operation.
CN202010055629.1A 2020-01-17 2020-01-17 Intelligent identification method and system for unsafe operation based on tower crane Active CN111439681B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010055629.1A CN111439681B (en) 2020-01-17 2020-01-17 Intelligent identification method and system for unsafe operation based on tower crane

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010055629.1A CN111439681B (en) 2020-01-17 2020-01-17 Intelligent identification method and system for unsafe operation based on tower crane

Publications (2)

Publication Number Publication Date
CN111439681A true CN111439681A (en) 2020-07-24
CN111439681B CN111439681B (en) 2021-11-02

Family

ID=71627026

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010055629.1A Active CN111439681B (en) 2020-01-17 2020-01-17 Intelligent identification method and system for unsafe operation based on tower crane

Country Status (1)

Country Link
CN (1) CN111439681B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114314347A (en) * 2022-01-21 2022-04-12 大连科润重工起重机有限公司 Safety monitoring and management system for hoisting machinery
WO2022121923A1 (en) * 2020-12-10 2022-06-16 东北大学 Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium
CN114715806A (en) * 2022-06-08 2022-07-08 杭州未名信科科技有限公司 Emergency control method, device and medium for abnormal state of tower crane and tower crane
CN116449725A (en) * 2023-06-19 2023-07-18 深蓝(天津)智能制造有限责任公司 Visual calandria intelligent control method and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104085789A (en) * 2014-05-06 2014-10-08 新乡市起重机厂有限公司 Intelligent monitoring method of running state of crane
WO2017212115A1 (en) * 2016-06-10 2017-12-14 Lifthanger Finland Oy Load handling device for a transportation unit, and a transportation unit with said device
US20190197644A1 (en) * 2016-12-27 2019-06-27 Pusan National University Industry-University Cooperation Foundation System and method for planning yard crane work
CN110135648A (en) * 2019-05-21 2019-08-16 北京起重运输机械设计研究院有限公司 A kind of safe condition prediction technique, device and the electronic equipment of operating equipment

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104085789A (en) * 2014-05-06 2014-10-08 新乡市起重机厂有限公司 Intelligent monitoring method of running state of crane
WO2017212115A1 (en) * 2016-06-10 2017-12-14 Lifthanger Finland Oy Load handling device for a transportation unit, and a transportation unit with said device
US20190197644A1 (en) * 2016-12-27 2019-06-27 Pusan National University Industry-University Cooperation Foundation System and method for planning yard crane work
CN110135648A (en) * 2019-05-21 2019-08-16 北京起重运输机械设计研究院有限公司 A kind of safe condition prediction technique, device and the electronic equipment of operating equipment

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022121923A1 (en) * 2020-12-10 2022-06-16 东北大学 Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium
CN114314347A (en) * 2022-01-21 2022-04-12 大连科润重工起重机有限公司 Safety monitoring and management system for hoisting machinery
CN114715806A (en) * 2022-06-08 2022-07-08 杭州未名信科科技有限公司 Emergency control method, device and medium for abnormal state of tower crane and tower crane
CN116449725A (en) * 2023-06-19 2023-07-18 深蓝(天津)智能制造有限责任公司 Visual calandria intelligent control method and system
CN116449725B (en) * 2023-06-19 2023-09-12 深蓝(天津)智能制造有限责任公司 Visual calandria intelligent control method and system

Also Published As

Publication number Publication date
CN111439681B (en) 2021-11-02

Similar Documents

Publication Publication Date Title
CN111439681B (en) Intelligent identification method and system for unsafe operation based on tower crane
CN111274737A (en) Method and system for predicting remaining service life of mechanical equipment
CN107862331A (en) It is a kind of based on time series and CNN unsafe acts recognition methods and system
CN109292567A (en) A kind of elevator faults prediction technique based on BP neural network
CN110348752B (en) Large industrial system structure safety assessment method considering environmental interference
CN108764601A (en) A kind of monitoring structural health conditions abnormal data diagnostic method based on computer vision and depth learning technology
CN108921230A (en) Method for diagnosing faults based on class mean value core pivot element analysis and BP neural network
CN109681391B (en) Blade root bolt fracture fault detection method and medium
CN109973331B (en) Wind turbine generator system wind turbine blade fault diagnosis algorithm based on bp neural network
CN110703712B (en) Industrial control system information security attack risk assessment method and system
Singh et al. EYE-on-HMI: A Framework for monitoring human machine interfaces in control rooms
CN104085789B (en) A kind of intelligent monitoring method of crane running status
CN102297767A (en) Health monitoring method of rope system based on angle monitoring when support displaces angularly
CN107491058A (en) A kind of industrial control system sequence attack detection method and equipment
CN108596364B (en) Dynamic early warning method for major hazard source in chemical industry park
CN113593605A (en) Industrial audio fault monitoring system and method based on deep neural network
CN101894214A (en) Mine ventilation system fault judging method based on hereditary neural network
CN111158338A (en) Chemical risk monitoring method based on principal component analysis
CN114897084A (en) Tower crane structure safety monitoring method based on graph convolution neural network
CN115187026A (en) Industrial risk monitoring method and system and readable storage medium
CN114417729A (en) Mining area environment safety early warning method based on BP neural network
CN113538997A (en) Equipment predictive maintenance learning system
CN112510699A (en) Transformer substation secondary equipment state analysis method and device based on big data
CN117163835B (en) Logistics control method and system based on virtual reality
CN106081911B (en) A kind of derrick crane on-line monitoring system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant