CN113277388B - Data acquisition control method for electric hanging basket - Google Patents
Data acquisition control method for electric hanging basket Download PDFInfo
- Publication number
- CN113277388B CN113277388B CN202110362590.2A CN202110362590A CN113277388B CN 113277388 B CN113277388 B CN 113277388B CN 202110362590 A CN202110362590 A CN 202110362590A CN 113277388 B CN113277388 B CN 113277388B
- Authority
- CN
- China
- Prior art keywords
- weight
- motor
- hanging basket
- data
- network
- 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.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B1/00—Control systems of elevators in general
- B66B1/24—Control systems with regulation, i.e. with retroactive action, for influencing travelling speed, acceleration, or deceleration
- B66B1/28—Control systems with regulation, i.e. with retroactive action, for influencing travelling speed, acceleration, or deceleration electrical
- B66B1/30—Control systems with regulation, i.e. with retroactive action, for influencing travelling speed, acceleration, or deceleration electrical effective on driving gear, e.g. acting on power electronics, on inverter or rectifier controlled motor
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B1/00—Control systems of elevators in general
- B66B1/24—Control systems with regulation, i.e. with retroactive action, for influencing travelling speed, acceleration, or deceleration
- B66B1/28—Control systems with regulation, i.e. with retroactive action, for influencing travelling speed, acceleration, or deceleration electrical
- B66B1/30—Control systems with regulation, i.e. with retroactive action, for influencing travelling speed, acceleration, or deceleration electrical effective on driving gear, e.g. acting on power electronics, on inverter or rectifier controlled motor
- B66B1/304—Control systems with regulation, i.e. with retroactive action, for influencing travelling speed, acceleration, or deceleration electrical effective on driving gear, e.g. acting on power electronics, on inverter or rectifier controlled motor with starting torque control
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B1/00—Control systems of elevators in general
- B66B1/34—Details, e.g. call counting devices, data transmission from car to control system, devices giving information to the control system
- B66B1/3415—Control system configuration and the data transmission or communication within the control system
- B66B1/3446—Data transmission or communication within the control system
- B66B1/3453—Procedure or protocol for the data transmission or communication
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B1/00—Control systems of elevators in general
- B66B1/34—Details, e.g. call counting devices, data transmission from car to control system, devices giving information to the control system
- B66B1/3476—Load weighing or car passenger counting devices
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/0006—Monitoring devices or performance analysers
- B66B5/0012—Devices monitoring the users of the elevator system
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/0006—Monitoring devices or performance analysers
- B66B5/0018—Devices monitoring the operating condition of the elevator system
- B66B5/0031—Devices monitoring the operating condition of the elevator system for safety reasons
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/02—Applications of checking, fault-correcting, or safety devices in elevators responsive to abnormal operating conditions
- B66B5/14—Applications of checking, fault-correcting, or safety devices in elevators responsive to abnormal operating conditions in case of excessive loads
- B66B5/145—Applications of checking, fault-correcting, or safety devices in elevators responsive to abnormal operating conditions in case of excessive loads electrical
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention relates to a data acquisition control method of an electric hanging basket, which comprises the following steps: step 1: video acquisition, picture analysis and processing, specifically as follows: step 1.1: video acquisition, storage and data preprocessing; step 1.2: face recognition analysis processing; step 1.3: the neural network predicts the weight of staff; step 1.4: network packet loss self-adaptive repair; step 2: the basket parameter real-time data analysis and processing are as follows: step 2.1: estimating the total load of the motor; step 2.2: a self-adaptive starting strategy of the motor; step 2.3: according to the scheme, an LSTM visual neural network is mainly adopted for transmitting and processing hanging basket parameters, the weight is estimated according to the length and the width of the human body outline, and the self-adaptive estimation of the complicated weight from the length and the width of the human body in the photo is realized; the load evaluation function is achieved, the motor starting strategy is adjusted adaptively, the use of an expensive gravity sensor is avoided, and the cost in industrial production is reduced.
Description
Technical Field
The invention relates to a control method, in particular to a data acquisition control method for an electric hanging basket, and belongs to the technical field of data acquisition control for electric hanging baskets.
Background
The electric hanging basket is manned and carried lifting equipment commonly used for construction sites, and can be used for assisting high-altitude operators to finish the work of installing glass curtain walls and the like outside the construction sites. A set of electric hanging basket data acquisition system is developed to upgrade and reform the traditional hanging basket, has the function of uploading data in a networking way, and adopts a 4G module networking way.
The internet of things equipment often uses an embedded system as a terminal, the embedded terminal adopts an embedded microprocessor as a main control chip, data of the equipment are acquired by accessing different sensors, and the network module is accessed for networking. Some embedded microprocessors are also provided with special image processing chips, so that video and audio acquired by the cameras and the microphones can be compression-coded. And finally, the data are sent to the cloud through a network interface of the embedded operating system. Compared with the traditional acquisition system, the embedded data acquisition system has the advantages of low cost, small volume, low power consumption, convenient use and the like.
Safety information such as a main rope, a safety rope and an inclination angle of the hanging basket is obtained in real time through a series of sensors arranged on the hanging basket. The information is processed in a centralized way by the embedded system, closed loop adjustment can be carried out, and an alarm can be sent out rapidly locally, so that accidents are prevented. Meanwhile, the safety information and the additional information such as video and sound are uploaded to the cloud end, so that management personnel can conveniently and comprehensively know the running condition of the hanging basket, and the high-rise service functions such as hanging basket renting and the like can be expanded. Therefore, the running state of the hanging basket can be controlled in real time, potential safety hazards can be found out in time, evidence can be reserved for responsibility identification, and upper-layer services are richer.
Disclosure of Invention
The invention provides an electric hanging basket data acquisition control system, which aims at the problems existing in the prior art, and the technical scheme provides safety measures such as serial data, network data receiving and transmitting processing, timing task service program, video automatic export program, audio and video acquisition coding program, streaming media service program running environment configuration script, GPIO and network monitoring program for attendance check-in management of workers, epidemic prevention and control management, perfecting safety belt, safety helmet wearing, hanging basket inclination monitoring, hanging rope tension test and the like.
In order to achieve the above purpose, the technical scheme of the invention is as follows, and the method for controlling the data acquisition of the electric hanging basket comprises the following steps:
step 1: video acquisition, picture analysis and processing, specifically as follows:
step 1.1: video acquisition, storage and data preprocessing;
step 1.2: face recognition analysis processing;
step 1.3: the neural network predicts the weight of staff;
step 1.4: network packet loss self-adaptive repair;
step 2: the basket parameter real-time data analysis and processing are as follows:
step 2.1: estimating the total load of the motor;
step 2.2: a self-adaptive starting strategy of the motor;
step 2.3: transmitting and processing parameters of the hanging basket;
the video acquisition, storage and data preprocessing in step 1.1 are specifically as follows:
step 1.1.1: setting an Internet protocol version 4 (TCP/IPv 4) attribute, matching an ip address with an ip address of an external camera, accessing the external camera to a CMOS camera by using an rtsp push-flow protocol, encoding a video signal into an analog signal, and accessing the analog signal into a camera interface of an MCU after the analog signal is transmitted into an analog video encoding chip of a circuit board through a coaxial line, wherein if the industrial field intensity has more electric interference, a twisted pair can be used for shielding electromagnetic induction transmission signals; meanwhile, a low-code-rate, high-quality and Jiang Rongcuo-force and strong-network-adaptability H.264 mode is adopted to encode the video;
step 1.1.2: for the coded video and audio streams, the RTMP protocol with good real-time performance in the streaming media protocol and supporting video streams and audio streams in various coding modes is adopted; accessing an NB-IoT module for networking, and storing the video stream to an SD card or uploading the video stream to a background real-time monitoring interface;
step 1.1.3: the method comprises the steps of carrying out cross compiling on a PC end, downloading the cross compiling data into an embedded system, then transmitting the cross compiling data to a cloud end through an embedded operation system network serial port, and collecting face photos, figure photos and weight data of workers as training data in the step 2.3.2; converting the face photo into a pgm format, unifying the photos into 10 pieces of uniform size, respectively placing the photos in a specified sequence number folder to generate a csv label file, marking the figure photo by an OpenCV function getOrientation and a PCA method, outputting the length and the width of the figure outline, and generating a label file for storage;
the face recognition analysis processing in step 1.2 is specifically as follows:
the method comprises the steps of judging the names of workers through face recognition, calculating the total number of workers in a hanging basket, training a model by using Facerecognizer-like Eigenfaces in OpenCV, extracting face csv label files processed in step 1.1.3, training to generate face xml model files, designing a monitoring platform based on an MFC, firstly opening a camera, loading the face model generated above, carrying out face detection, carrying out the signing of the names of the workers, and identifying a plurality of people in the hanging basket.
The neural network predicts the weight of staff in step 1.3, and is specifically as follows:
step 1.3.1: the worker stands at a specified position away from the camera, the system automatically shoots a figure photo, the figure photo needs to be marked with a figure outline photo through an OpenCV function getOrientation and a PCA method, the length and the width of the figure outline are output, and for the mobility of the items, the LSTM neural network is trained to be convenient for directly estimating the weight according to the length and the width of the figure of the person;
step 1.3.2: inputting a figure training data set generated in the LSTM network step 1.1.3, wherein the figure training data set contains a label file with a figure outline length and a figure outline width, and the corresponding weight data is used as supervision data to train the neural network to obtain a result. The body weight of the person is directly predicted from the length and the width of the figure photo;
step 1.3.3: inputting the body length and width acquired in real time into a trained LSTM neural network, predicting the approximate weight range of the person by the network, identifying the total number of people in the hanging basket through the step 1.2, shooting for a plurality of times by the system, and accumulating the weight of the worker predicted by the network.
The self-adaptive repairing method of the network packet loss in the step 1.4 comprises the following steps:
step 1.4.1: the bandwidth of the blocked link should be increased appropriately in order to reduce network packet loss caused by network link blocking;
step 1.4.2: the use of Qos (traffic priority and resource reservation control mechanism) prioritizes real-time applications, although this approach does not alleviate network link blocking, it can prioritize voice and video to reduce the likelihood of wire breaks.
The step 2.1 is to estimate the total load of the motor, and the specific method is as follows:
the worker predicted in the step 1.2 is sent to weight to accumulate the total weight of hanging basket equipment, the hanging basket equipment comprises the weight of the hanging basket, the total weight of a rope connected with a motor, a speed sensor, a frequency converter, an inclinometer, an accelerometer, a hanging rope force sensor and the like, the wind speed sensor measures and calculates the wind speed, the force exerted on the hanging basket by the external environment is calculated through buoyancy and a pressure formula, and therefore the total load on the motor is calculated.
The self-adaptive starting strategy of the motor in the step 2.2 comprises the following specific steps:
step 2.2.1: a closed-loop control algorithm is adopted to rapidly and stably start the motor. And when the total load is greater than 200kg, adopting a PI control algorithm. When the load increases, namely the armature circuit string resistance R1 increases, the armature voltage Ua and the magnetic flux phi are kept unchanged, and the load torque T 1 When the electromagnetic torque T and the armature current Ia are unchanged, the mechanical characteristic equation after speed regulation is as follows:
wherein n is the rotating speed, ua is the armature voltage, R1 is the armature loop series resistance, phi is the magnetic flux, T 1 Is load torque, T is electromagnetic torque, ia is armature current, ce is electromotive force constant, C T As torque constant, ra is armature resistance;
it is known that after R1 increases, the motor rotation speed n decreases, and when Ua and Ia are unchanged, the total power p1=uaia input from the power supply is unchanged. As the varistor R1 increases, the winding copper loss Pcu =ia 2 (Ra+R1) also increases the electromagnetic power P thus converted into mechanical power M = (P1-Pcu) will decrease, while P M When the load torque is unchanged, the motor rotation speed is reduced along with the increase of the armature circuit series resistor R1, PI parameters are adjusted after the motor rotation speed is reduced, and the rotation speed can be increased by increasing the armature power supply voltage table and reducing the magnetic flux phi, so that the motor is adjusted to run at different powers;
step 2.2.2: the total weight of the load is less than 200kg, the system is easy to be interfered, the fuzzy PI control algorithm is adopted, the fuzzy PI control algorithm combines the advantages of a fuzzy controller and a traditional PI controller, the system has good dynamic, steady-state performance and anti-interference capacity, the fuzzy controller takes errors and error changes as input, the PI parameters kp and ki are modified in real time, and the requirement of self-tuning of the PI parameters at different moments can be met;
step 2.2.3: the fuzzy PI algorithm of the system takes the feedback speed omega and the speed set value omega ref of a motor encoder as input quantities, outputs output quantities to a current loop through input quantity fuzzification, a data expert database, fuzzy reasoning and anti-fuzzification, and in a control system, three-purpose resistance method is used for sampling three-phase current to obtain phase A current ia and phase B current ib, and according to the three-phase current characteristics of ia+ib+ic=0, mutually orthogonal time-varying current values ialpha and ibeta are obtained through Clarke transformation; iα and iβ are subjected to Park transformation to obtain orthogonal current constants id and iq under a rotating coordinate system; and comparing id and iq with a speed variation quantity i x d and a target reference value i x q to be used as error input quantity of a current loop PI algorithm, converting the error input quantity into a static reference coordinate system after inverse Park conversion to obtain valpha and vbeta, obtaining va, vb and vc after inverse Clarke conversion, calculating a new PWM duty ratio value according to a three-phase voltage value, and generating a desired voltage vector, thereby completing speed and current double closed loop control of the system.
The transmission and processing of the basket parameters in the step 2.3 are specifically as follows:
step 2.3.1: the file is transmitted by adopting the FTP protocol, a set of FTP server is built, and the file uploading function can be conveniently realized by adopting a corresponding FTP tool at the equipment end;
step 2.3.2: considering that the FTP protocol is only responsible for file transmission, and a server cannot know whether a newly uploaded file exists or not, a report flow is designed for the data needing to be replied for voice intercom, and after the file is successfully uploaded, the newly uploaded file name is reported to the background in the form of common data and an instruction is sent to the background;
step 2.3.3: the background receives the instructions and responds.
Compared with the prior art, the invention has the following advantages,
1) According to the invention, an LSTM visual neural network is mainly adopted, and the weight is estimated according to the length and the width of the human body outline, so that the self-adaptive estimation of the complicated weight of the length and the width of the human body in the photo is realized; the effect of evaluating the load is achieved, so that the motor starting strategy is adaptively adjusted, the use of an expensive gravity sensor is avoided, and the cost in industrial production is reduced;
2) The Fuzzy PID algorithm is adopted to improve the response speed and control precision of the control algorithm, and is convenient for mass production, so that preparation is made for the design of a high-precision servo control system, in a linear control theory, the integral control function can eliminate steady-state errors, but the dynamic response is slow, the proportional control function is quick in dynamic response, and the proportional integral control function can obtain higher steady-state precision and also has faster dynamic response, therefore, a P control strategy is introduced into the Fuzzy controller to form Fuzzy-PI composite control, and the steady-state performance of the Fuzzy controller is improved;
3) The invention adopts a video monitoring processing scheme of 'video acquisition module-image compression module + DE storage module + network module + embedded processor', mainly realizes an MPEG-4 hardware encoded embedded network video terminal with video acquisition function, and accesses the cloud for management, can utilize NPU module of Hai Si processor for pattern recognition, and completes advanced functions such as video analysis;
4) According to the invention, a self-adaptive motor abnormality processing strategy is adopted, when the overweight of the hanging basket, the overlarge inclination angle and the overlarge acceleration are detected, a circuit is disconnected, the rotation of the motor is stopped, the cable is tensioned and locked, and at the moment, multiple detection and confirmation can be carried out; if the inclination angle is confirmed to be too large, the acceleration is too large and other circuit faults are confirmed, restarting the main circuit, and reporting abnormal faults to the background; if the hanging basket is confirmed to be overweight, the hanging basket loudspeaker prompts personnel to be overweight through voice, an abnormal fault is reported to the background, and the main circuit is restarted after the tension of the main hanging rope is restored to a normal value;
5) Since the device requires control of the adaptive motor, the invention is equipped with a script auto-configuration operating environment, such as: dynamic link library, connection network, synchronization time and system upgrade; because the system keeps running stably for a long time, a series of protection programs such as GPIO related programs, network monitoring programs and other system environment monitoring repair programs are set to automatically detect and repair system faults or indicate the working state of the system so as to recover the normal working state as soon as possible. Such as: a network monitoring program is needed to monitor the on-off state of the network and restart the network connection module when the network is disconnected; when the motor is restarted, the environment detection program automatically configures the running environment, so that the motor is ensured to be started in a more stable state.
Drawings
FIG. 1 is a flow chart of the working principle of the invention;
fig. 2 is a schematic diagram of an adaptive start strategy procedure for an electric machine.
The specific embodiment is as follows:
in order to enhance the understanding of the present invention, the present embodiment will be described in detail with reference to the accompanying drawings.
Example 1: referring to fig. 1 and 2, a method for controlling data acquisition of an electric hanging basket includes the following steps:
step 1: video acquisition, picture analysis and processing, specifically as follows:
step 1.1: video acquisition, storage and data preprocessing;
step 1.2: face recognition analysis processing;
step 1.3: the neural network predicts the weight of staff;
step 1.4: network packet loss self-adaptive repair;
step 2: the basket parameter real-time data analysis and processing are as follows:
step 2.1: estimating the total load of the motor;
step 2.2: a self-adaptive starting strategy of the motor;
step 2.3: transmitting and processing parameters of the hanging basket;
the video acquisition, storage and data preprocessing in step 1.1 are specifically as follows:
step 1.1.1: setting an Internet protocol version 4 (TCP/IPv 4) attribute, matching an ip address with an ip address of an external camera, accessing the external camera to a CMOS camera by using an rtsp push-flow protocol, encoding a video signal into an analog signal, and accessing the analog signal into a camera interface of an MCU after the analog signal is transmitted into an analog video encoding chip of a circuit board through a coaxial line, wherein if the industrial field intensity has more electric interference, a twisted pair can be used for shielding electromagnetic induction transmission signals; meanwhile, a low-code-rate, high-quality and Jiang Rongcuo-force and strong-network-adaptability H.264 mode is adopted to encode the video;
step 1.1.2: for the coded video and audio streams, the RTMP protocol with good real-time performance in the streaming media protocol and supporting video streams and audio streams in various coding modes is adopted; accessing an NB-IoT module for networking, and storing the video stream to an SD card or uploading the video stream to a background real-time monitoring interface;
step 1.1.3: the method comprises the steps of carrying out cross compiling on a PC end, downloading the cross compiling data into an embedded system, then transmitting the cross compiling data to a cloud end through an embedded operation system network serial port, and collecting face photos, figure photos and weight data of workers as training data in the step 2.3.2; converting the face photo into a pgm format, unifying the photos into 10 pieces, respectively placing the 10 pieces in a folder with a specified serial number to generate a csv label file, marking the figure photo by an OpenCV function getOrientation and a PCA method, outputting the length and the width of the figure outline, and generating a label file for storage.
The face recognition analysis processing in step 1.2 is specifically as follows:
the method comprises the steps of judging the names of workers through face recognition, calculating the total number of workers in a hanging basket, training a model by using Facerecognizer-like Eigenfaces in OpenCV, extracting face csv label files processed in step 1.1.3, training to generate face xml model files, designing a monitoring platform based on an MFC, firstly opening a camera, loading the face model generated above, carrying out face detection, carrying out the signing of the names of the workers, and identifying a plurality of people in the hanging basket.
The neural network predicts the weight of staff in step 1.3, and is specifically as follows:
step 1.3.1: the worker stands at a specified position away from the camera, the system automatically shoots a figure photo, the figure photo needs to be marked with a figure outline photo through an OpenCV function getOrientation and a PCA method, the length and the width of the figure outline are output, and for the mobility of the items, the LSTM neural network is trained to be convenient for directly estimating the weight according to the length and the width of the figure of the person;
step 1.3.2: inputting a figure training data set generated in the LSTM network step 1.1.3, wherein the figure training data set contains a label file with a figure outline length and a figure outline width, and the corresponding weight data is used as supervision data to train the neural network to obtain a result. The body weight of the person is directly predicted from the length and the width of the figure photo;
step 1.3.3: inputting the body length and width acquired in real time into a trained LSTM neural network, predicting the approximate weight range of the person by the network, identifying the total number of people in the hanging basket through the step 1.2, shooting for a plurality of times by the system, and accumulating the weight of the worker predicted by the network.
The self-adaptive repairing method of the network packet loss in the step 1.4 comprises the following steps:
step 1.4.1: the bandwidth of the blocked link should be increased appropriately in order to reduce network packet loss caused by network link blocking;
step 1.4.2: the use of Qos (traffic priority and resource reservation control mechanism) prioritizes real-time applications, although this approach does not alleviate network link blocking, it can prioritize voice and video to reduce the likelihood of wire breaks.
The step 2.1 is to estimate the total load of the motor, and the specific method is as follows:
the worker predicted in the step 1.2 is sent to weight to accumulate the total weight of hanging basket equipment, the hanging basket equipment comprises the weight of the hanging basket, the total weight of a rope connected with a motor, a speed sensor, a frequency converter, an inclinometer, an accelerometer, a hanging rope force sensor and the like, the wind speed sensor measures and calculates the wind speed, the force exerted on the hanging basket by the external environment is calculated through buoyancy and a pressure formula, and therefore the total load on the motor is calculated.
The self-adaptive starting strategy of the motor in the step 2.2 comprises the following specific steps:
step 2.2.1: a closed-loop control algorithm is adopted to rapidly and stably start the motor. And when the total load is greater than 200kg, adopting a PI control algorithm. When the load increases, namely the armature circuit string resistance R1 increases, the armature voltage Ua and the magnetic flux phi are kept unchanged, and the load torque T 1 When the electromagnetic torque T and the armature current Ia are unchanged, the mechanical characteristic equation after speed regulation is as follows:
wherein n is the rotating speed, ua is the armature voltage, R1 is the armature loop series resistance, phi is the magnetic flux, T 1 Is load torque, T is electromagnetic torque, ia is armature current, ce is electromotive force constant, C T As torque constant, ra is armature resistance;
it is known that after R1 increases, the motor rotation speed n decreases, and when Ua and Ia are unchanged, the total power p1=uaia input from the power supply is unchanged. As the varistor R1 increases, the winding copper loss Pcu =ia 2 (Ra+R1) also increases the electromagnetic power P thus converted into mechanical power M = (P1-Pcu) will decrease, while P M When the load torque is unchanged, the motor rotation speed is reduced along with the increase of the armature circuit series resistor R1, PI parameters are adjusted after the motor rotation speed is reduced, and the rotation speed can be increased by increasing the armature power supply voltage table and reducing the magnetic flux phi, so that the motor is adjusted to run at different powers;
step 2.2.2: the total weight of the load is less than 200kg, the system is easy to be interfered, the fuzzy PI control algorithm is adopted, the fuzzy PI control algorithm combines the advantages of a fuzzy controller and a traditional PI controller, the system has good dynamic, steady-state performance and anti-interference capacity, the fuzzy controller takes errors and error changes as input, the PI parameters kp and ki are modified in real time, and the requirement of self-tuning of the PI parameters at different moments can be met;
step 2.2.3: the fuzzy PI algorithm of the system takes the feedback speed omega and the speed set value omega ref of a motor encoder as input quantities, outputs output quantities to a current loop through input quantity fuzzification, a data expert database, fuzzy reasoning and anti-fuzzification, and in a control system, three-purpose resistance method is used for sampling three-phase current to obtain phase A current ia and phase B current ib, and according to the three-phase current characteristics of ia+ib+ic=0, mutually orthogonal time-varying current values ialpha and ibeta are obtained through Clarke transformation; iα and iβ are subjected to Park transformation to obtain orthogonal current constants id and iq under a rotating coordinate system; and comparing id and iq with a speed variation quantity i x d and a target reference value i x q to be used as error input quantity of a current loop PI algorithm, converting the error input quantity into a static reference coordinate system after inverse Park conversion to obtain valpha and vbeta, obtaining va, vb and vc after inverse Clarke conversion, calculating a new PWM duty ratio value according to a three-phase voltage value, and generating a desired voltage vector, thereby completing speed and current double closed loop control of the system.
The transmission and processing of the basket parameters in the step 2.3 are specifically as follows:
step 2.3.1: the file is transmitted by adopting the FTP protocol, a set of FTP server is built, and the file uploading function can be conveniently realized by adopting a corresponding FTP tool at the equipment end;
step 2.3.2: considering that the FTP protocol is only responsible for file transmission, and a server cannot know whether a newly uploaded file exists or not, a report flow is designed for the data needing to be replied for voice intercom, and after the file is successfully uploaded, the newly uploaded file name is reported to the background in the form of common data and an instruction is sent to the background;
step 2.3.3: the background receives the instructions and responds.
Specific examples: in the embodiment, the Hai Si processor Hi3516 is used as an embedded system main control chip, and the Linux kernel version is 2.6.35. A method for controlling data acquisition of an electric hanging basket, the method comprising the following steps:
step 1: video acquisition, picture analysis and processing, specifically as follows:
step 1.1: video acquisition, storage and data preprocessing;
step 1.1.1: setting an Internet protocol version 4 (TCP/IPv 4) attribute, matching an ip address with an ip address of an external camera, accessing the external camera to a CMOS camera by using an rtsp push-flow protocol, encoding a video signal into an analog signal, and accessing the analog signal into a camera interface of an MCU after the analog signal is transmitted into an analog video encoding chip of a circuit board through a coaxial line, wherein if the industrial field intensity has more electric interference, a twisted pair can be used for shielding electromagnetic induction transmission signals; meanwhile, a low-code-rate, high-quality and Jiang Rongcuo-force and strong-network-adaptability H.264 mode is adopted to encode the video;
step 1.1.2: for the coded video and audio streams, the RTMP protocol with good real-time performance in the streaming media protocol and supporting video streams and audio streams in various coding modes is adopted; and accessing an NB-IoT module for networking, and storing the video stream to an SD card or uploading the video stream to a background real-time monitoring interface.
Step 1.1.3: and (4) after cross compiling at the PC end, downloading the data into an embedded system, transmitting the data to a cloud end through an embedded operating system network serial port, and collecting face photos, figure photos and weight data of workers as training data in the step (4). And converting the facial photos into a pgm format, unifying the photos into a uniform size, and placing 10 photos in a specified serial number folder respectively. Generating a csv label file. The body shape photo is required to be marked by an OpenCV function getOrientation and a PCA method, the length and the width of the body shape outline are output, and a label file is generated and stored.
Step 1.2: face recognition analysis processing;
and judging the name of the workers through face recognition, checking in, and calculating the total number of workers in the hanging basket. Training a model by using a Facerecognizer Eigenfaces function in OpenCV, extracting the face csv label file processed in the step 1.1.3, and training to generate a face xml model file. Based on an MFC design monitoring platform, a camera is firstly opened, a face model generated above is loaded, face detection is carried out, name sign-in of workers is carried out, and a plurality of people in the hanging basket are identified.
Step 1.3: the neural network predicts the weight of staff;
step 1.3.1: the worker stands at a designated position away from the camera, the system automatically shoots a figure photo, the figure photo is required to be marked with a figure outline photo through an OpenCV function getOrientation and a PCA method, and the length and the width of the figure outline are output. Training the LSTM neural network facilitates direct estimation of body weight based on length and width of a person's figure for item mobility.
Step 1.3.2: inputting a figure training data set generated in the LSTM network step 1.1.3, wherein the figure training data set contains a label file with a figure outline length and a figure outline width, and the corresponding weight data is used as supervision data to train the neural network to obtain a result. The body weight of the person is directly predicted from the length and width of the figure photo.
Step 1.3.3: the length and width of the body acquired in real time are input into a trained LSTM neural network, which predicts the approximate weight range of the person. And (3) taking multiple shots by the system through the total number of people in the hanging basket identified in the step (1.3), and accumulating the weight of workers predicted by the network.
Step 1.4: network packet loss self-adaptive repair;
step 1.4.1: the bandwidth of the blocked link should be increased appropriately in order to reduce network packet loss caused by network link blocking.
Step 1.4.2: the use of Qos (traffic priority and resource reservation control mechanism) prioritizes real-time applications, although this approach does not alleviate network link blocking, it can prioritize voice and video to reduce the likelihood of wire breaks.
Step 2: the basket parameter real-time data analysis and processing are as follows:
step 2.1: estimating the total load of the motor;
step 2.1.1: and (2) feeding the weight of the worker predicted in the step (1.2) to accumulate the total weight of hanging basket equipment, wherein the total weight of the hanging basket equipment comprises the weight of the hanging basket, the total weight of a rope connected with a motor, a speed sensor, a frequency converter, an inclinometer, an accelerometer, a hanging rope force sensor and the like. The wind speed sensor measures and calculates the wind speed, and the force exerted on the hanging basket by the external environment at the moment is calculated through the buoyancy and pressure formula. From this, the total load on the motor is calculated.
Step 2.2: a self-adaptive starting strategy of the motor;
step 2.2.1: a closed-loop control algorithm is adopted to rapidly and stably start the motor. And when the total load is greater than 200kg, adopting a PI control algorithm. When the load increases, namely the armature circuit string resistance R1 increases, the armature voltage Ua and the magnetic flux phi are kept unchanged, and the load torque T 1 When the electromagnetic torque T and the armature current Ia are unchanged, the mechanical characteristic equation after speed regulation is as follows:
wherein n is the rotating speed, ua is the armature voltage, R1 is the armature loop series resistance, phi is the magnetic flux, T 1 Is load torque, T is electromagnetic torque, ia is armature current, ce is electricPotential constant, C T As torque constant, ra is armature resistance;
it is known that after R1 increases, the motor rotation speed n decreases, and when Ua and Ia are unchanged, the total power p1=uaia input from the power supply is unchanged. As the varistor R1 increases, the winding copper loss Pcu =ia 2 (Ra+R1) also increases the electromagnetic power P thus converted into mechanical power M = (P1-Pcu) will decrease, while P M When the load torque is unchanged, the motor rotation speed is reduced along with the increase of the armature circuit series resistor R1, PI parameters are adjusted after the motor rotation speed is reduced, and the rotation speed can be increased by increasing the armature power supply voltage table and reducing the magnetic flux phi, so that the motor is adjusted to run at different powers;
step 2.2.2: the total weight of the load is less than 200kg, the system is more easily disturbed, and a fuzzy PI control algorithm is adopted. The fuzzy PI control algorithm combines the advantages of a fuzzy controller and a traditional PI controller, and can enable the system to have good dynamic and steady-state performance and anti-interference capability. The fuzzy controller takes errors and error changes as input, modifies PI parameters kp and ki in real time, and can meet the requirements of self-tuning of the PI parameters at different moments.
Step 2.2.3: the fuzzy PI algorithm of the system takes the feedback speed omega and the speed set value omega ref of the motor encoder as input quantity, and outputs output quantity to a current loop through input quantity fuzzification, a data expert database, fuzzy reasoning and anti-fuzzification. In a control system, three-purpose resistance method is used for sampling three-phase current to obtain ia and ib, and according to the three-phase current characteristics of ia+ib+ic=0, mutually orthogonal time-varying current values ialpha and ibeta are obtained through Clarke transformation; iα and iβ are subjected to Park transformation to obtain orthogonal current constants id and iq under a rotating coordinate system; and comparing id and iq with a speed variation quantity i x d and a target reference value i x q to be used as error input quantity of a current loop PI algorithm, converting the error input quantity into a static reference coordinate system after inverse Park conversion to obtain valpha and vbeta, obtaining va, vb and vc after inverse Clarke conversion, calculating a new PWM duty ratio value according to a three-phase voltage value, and generating a desired voltage vector, thereby completing speed and current double closed loop control of the system.
Step 2.3: transmitting and processing parameters of the hanging basket;
step 2.3.1: and the file is transmitted by adopting an FTP protocol, a set of FTP server is built, and the file uploading function can be conveniently realized by adopting a corresponding FTP tool at the equipment end.
Step 2.3.2: considering that the FTP protocol is only responsible for file transmission, and the server cannot know whether a newly uploaded file exists, a report flow is designed for the data needing to be replied for voice intercom, and after the file is successfully uploaded, the newly uploaded file name is reported to the background in the form of common data, and an instruction is sent to the background.
Step 2.3.3: the background receives the instructions and responds.
It should be noted that the above-mentioned embodiments are not intended to limit the scope of the present invention, and equivalent changes or substitutions made on the basis of the above-mentioned technical solutions fall within the scope of the present invention as defined in the claims.
Claims (2)
1. The data acquisition control method for the electric hanging basket is characterized by comprising the following steps of:
step 1: video acquisition, picture analysis and processing, specifically as follows:
step 1.1: video acquisition, storage and data preprocessing;
step 1.2: face recognition analysis processing;
step 1.3: the neural network predicts the weight of staff;
step 1.4: network packet loss self-adaptive repair;
step 2: the basket parameter real-time data analysis and processing are as follows:
step 2.1: estimating the total load of the motor;
step 2.2: a self-adaptive starting strategy of the motor;
step 2.3: the transmission and processing of basket parameters, the video acquisition, storage and data preprocessing in step 1.1 are as follows:
step 1.1.1: setting an Internet protocol version 4, namely TCP/IPv4 attribute, matching an ip address with an ip address of an external camera, accessing the external camera to a CMOS camera by using an rtsp push flow protocol, encoding a video signal into an analog signal, and accessing the analog signal into a camera interface of an MCU after the analog signal is transmitted into an analog video encoding chip of a circuit board through a coaxial line, wherein if the industrial field has more electric interference, a twisted pair is adopted to shield electromagnetic induction transmission signals; meanwhile, a low-code-rate, high-quality and Jiang Rongcuo-force and strong-network-adaptability H.264 mode is adopted to encode the video;
step 1.1.2: for the coded video and audio streams, the RTMP protocol with good real-time performance in the streaming media protocol and supporting video streams and audio streams in various coding modes is adopted; accessing an NB-IoT module for networking, and storing the video stream to an SD card or uploading the video stream to a background real-time monitoring interface;
step 1.1.3: the method comprises the steps of carrying out cross compiling on a PC end, downloading the cross compiling data into an embedded system, then transmitting the cross compiling data to a cloud end through an embedded operation system network serial port, and collecting face photos, figure photos and weight data of workers as training data in the step 2.3; converting the face photo into a pgm format, unifying the photos into 10 pieces, respectively placing the 10 pieces in a folder with a specified serial number to generate a csv label file, labeling the figure photo by an OpenCV function getOrientation and a PCA method to obtain a figure outline photo, outputting the length and width of the figure outline, generating a label file, and storing the label file, wherein the step 1.2 of face recognition analysis processing is specifically as follows:
face recognition judges the name of a worker to check in, calculates the total number of workers in a hanging basket, trains a model by using a Facerecognizer-like Eigenfaces function in OpenCV, extracts face csv label files processed in the step 1.1.3, trains and generates face xml model files, designs a monitoring platform based on an MFC, firstly opens a camera, loads the face model generated above, carries out face detection, carries out the check-in of the name of the worker, and identifies a plurality of people in the hanging basket, and predicts the weight of the staff by using a neural network in the step 1.3, and the method comprises the following steps of:
step 1.3.1: the worker stands at a specified position away from the camera, the system automatically shoots a figure photo, the figure photo is required to be marked with a figure outline photo through an OpenCV function getOrientation and a PCA method, the length and the width of the figure outline are output, and the LSTM neural network is trained to be convenient for directly estimating the weight according to the length and the width of the figure of the person;
step 1.3.2: inputting a figure training data set generated in the step 1.1.3 of the LSTM network, wherein the figure training data set contains a label file of figure outline length and width, corresponding weight data is used as supervision data, and a neural network is trained to obtain a result, so that the weight of the person can be directly predicted from the length and width of a figure photo;
step 1.3.3: inputting the body length and width acquired in real time into a trained LSTM neural network, predicting the approximate weight range of the person by the network, shooting the total number of people in the hanging basket identified in the step 1.2 for a plurality of times by the system, accumulating the weight of workers predicted by the network, and adaptively repairing the lost packets of the network in the step 1.4, wherein the method comprises the following steps of:
step 1.4.1: increasing bandwidth of the blocking link for reducing network packet loss caused by network link blocking;
step 1.4.2: the real-time application is preferentially processed by using Qos, namely the flow priority and the resource reservation control mechanism, and the total load of the motor is estimated in step 2.1, and the specific method is as follows:
the total weight of the hanging basket equipment is accumulated by the worker weight sending predicted in the step 1.2, the hanging basket total weight comprises the weight of the hanging basket, the total weight of a rope connected with a motor, a speed sensor, a frequency converter, an inclinometer, an accelerometer, a hanging rope force sensor equipment, a wind speed sensor, a wind speed measuring and calculating device, the force exerted on the hanging basket by the external environment at the moment is calculated through a buoyancy and pressure formula, and therefore the total load on the motor is calculated, and the self-adaptive starting strategy of the motor in the step 2.2 is specifically implemented by the following steps:
step 2.2.1: the closed-loop control algorithm is adopted, so that the motor is started quickly and stably, the total load weight is larger than 200kg, the PI control algorithm is adopted, when the load is increased, namely the armature loop string resistance R1 is increased, the armature voltage Ua and the magnetic flux phi are kept unchanged, and the load torque T is kept unchanged 1 When the electromagnetic torque T and the armature current Ia are unchanged, the mechanical characteristic equation after speed regulation is as follows:
wherein n is the rotation speed, ua is the armature voltage, R1 is the armature loop series resistance, phi is the magnetic flux, T1 is the load torque, T is the electromagnetic torque, ia is the armature current, ce is the electromotive force constant, C T As torque constant, ra is armature resistance;
it can be seen that after R1 increases, the motor speed n will decrease, and when Ua and Ia are unchanged, the total power p1=uaia input from the power supply will be unchanged, and as the varistor R1 increases, the winding copper loss Pcu =ia 2 (Ra+R1) also increases the electromagnetic power P thus converted into mechanical power M = (P1-Pcu) will decrease, while P M When the load torque is unchanged, the motor rotation speed is reduced along with the increase of the armature circuit series resistor R1, PI parameters are adjusted after the motor rotation speed is reduced, the rotation speed is increased by increasing the armature power supply voltage table to reduce the magnetic flux phi, and then the motor is adjusted to run at different powers;
step 2.2.2: the total weight of the load is less than 200kg, the system is easy to be interfered, a fuzzy PI control algorithm is adopted, the fuzzy PI control algorithm combines the advantages of a fuzzy controller and a traditional PI controller, the system has good dynamic, steady-state performance and anti-interference capacity, the fuzzy controller takes errors and error changes as input, PI parameters kp and ki are modified in real time, and the requirement of self-setting of the PI parameters at different moments is met;
step 2.2.3: the fuzzy PI algorithm of the system takes the feedback speed omega and the speed set value omega ref of a motor encoder as input quantities, outputs output quantities to a current loop through input quantity fuzzification, a data expert database, fuzzy reasoning and anti-fuzzification, and in a control system, three-purpose resistance method is used for sampling three-phase current to obtain phase A current ia and phase B current ib, and according to the three-phase current characteristics of ia+ib+ic=0, mutually orthogonal time-varying current values ialpha and ibeta are obtained through Clarke transformation; iα and iβ are subjected to Park transformation to obtain orthogonal current constants id and iq under a rotating coordinate system; and comparing id and iq with a speed variation quantity i x d and a target reference value i x q to be used as error input quantity of a current loop PI algorithm, converting the error input quantity into a static reference coordinate system after inverse Park conversion to obtain valpha and vbeta, obtaining va, vb and vc after inverse Clarke conversion, calculating a new PWM duty ratio value according to a three-phase voltage value, and generating a desired voltage vector, thereby completing speed and current double closed loop control of the system.
2. The method for controlling data acquisition of an electric hanging basket according to claim 1, wherein the transmission and processing of the hanging basket parameters in step 2.3 are as follows:
step 2.3.1: adopting an FTP protocol to carry out file transmission, constructing a set of FTP server, and adopting a corresponding FTP tool at the equipment end to conveniently realize the file uploading function;
step 2.3.2: considering that the FTP protocol is only responsible for file transmission, and a server cannot know whether a newly uploaded file exists or not, a report flow is designed for the data needing to be replied for voice intercom, and after the file is successfully uploaded, the newly uploaded file name is reported to the background in the form of common data and an instruction is sent to the background;
step 2.3.3: the background receives the instructions and responds.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110362590.2A CN113277388B (en) | 2021-04-02 | 2021-04-02 | Data acquisition control method for electric hanging basket |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110362590.2A CN113277388B (en) | 2021-04-02 | 2021-04-02 | Data acquisition control method for electric hanging basket |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113277388A CN113277388A (en) | 2021-08-20 |
| CN113277388B true CN113277388B (en) | 2023-05-02 |
Family
ID=77276228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110362590.2A Active CN113277388B (en) | 2021-04-02 | 2021-04-02 | Data acquisition control method for electric hanging basket |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113277388B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114590668B (en) * | 2022-05-10 | 2022-08-12 | 大汉科技股份有限公司 | Operation management system for unmanned elevator based on big data |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103701396B (en) * | 2013-12-13 | 2016-01-20 | 天津大学 | A kind of motor speed tracking and controlling method based on Adaptive Fuzzy Neural-network |
| CN105577058B (en) * | 2015-12-28 | 2018-08-21 | 江苏大学 | The five mutually fault-tolerant magneto method for control speed based on fuzzy automatic disturbance rejection controller |
| GB201613138D0 (en) * | 2016-07-29 | 2016-09-14 | Unifai Holdings Ltd | Computer vision systems |
| CN107678340A (en) * | 2017-10-10 | 2018-02-09 | 中建八局第二建设有限公司 | A kind of intelligent hanging basket monitoring system, comprehensive monitoring system and monitoring method |
| CN109507869A (en) * | 2018-11-15 | 2019-03-22 | 南京越博电驱动系统有限公司 | A kind of optimization method of the motor control PI parameter suitable for permanent magnet synchronous motor |
| CN109829390A (en) * | 2019-01-09 | 2019-05-31 | 浙江新再灵科技股份有限公司 | A kind of elevator intelligent scheduling system and method based on deep learning |
| CN109761118A (en) * | 2019-01-15 | 2019-05-17 | 福建天眼视讯网络科技有限公司 | Wisdom ladder networking control method and system based on machine vision |
| CN110311807B (en) * | 2019-06-06 | 2022-04-08 | 东南大学 | Network condition self-adaptive monitoring method for electric hanging basket data acquisition system |
| CN110443126A (en) * | 2019-06-27 | 2019-11-12 | 平安科技（深圳）有限公司 | Model hyper parameter adjusts control method, device, computer equipment and storage medium |
| CN110807434B (en) * | 2019-11-06 | 2023-08-15 | 威海若维信息科技有限公司 | Pedestrian re-recognition system and method based on human body analysis coarse-fine granularity combination |
| CN111476195A (en) * | 2020-04-20 | 2020-07-31 | 安徽中科首脑智能医疗研究院有限公司 | Face detection method, face detection device, robot and computer-readable storage medium |
-
2021
- 2021-04-02 CN CN202110362590.2A patent/CN113277388B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113277388A (en) | 2021-08-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113277388B (en) | Data acquisition control method for electric hanging basket | |
| CN1145577C (en) | Automatic calibration of current regulator control compensation for elevator motor drive with locked rotor | |
| CN107341489B (en) | Image recognition system for machine room monitoring | |
| CN107168294B (en) | Unmanned inspection monitoring method for thermal power water system equipment | |
| CN105829984A (en) | Machine diagnostic encoder | |
| US20070078614A1 (en) | Integration of intelligent motor with power management device | |
| CN109413188A (en) | A kind of industrial equipment management system for internet of things and method | |
| CN109142904B (en) | Image recognition-based fault monitoring method and device for direct-current transmission converter valve | |
| CN106483902B (en) | A kind of automatically controlled control system of barotor Internet-based | |
| CN113759791A (en) | Monitoring system, method and device based on intelligent gateway and intelligent gateway | |
| CN114110406B (en) | Intelligent lubrication management system | |
| CN110161952A (en) | A kind of method for managing and monitoring and device carrying out inspection for fire-fighting equipment | |
| CN110941228A (en) | Intelligent railway traction substation management and control system | |
| CN112910746A (en) | Home robot and intelligent home safety monitoring system intercommunicated with home equipment | |
| US20220407945A1 (en) | Method and system for processing sensor data for transmission | |
| CN108961169A (en) | Monitor grasp shoot method and device | |
| JP2022526350A (en) | A system for providing and / or monitoring operational data for a compressed air supply device, a method for providing operational data for a compressed air supply device, and / or a method for monitoring functionality. | |
| JP6935504B2 (en) | Data acquisition method, inverter, and rotary electric machine | |
| CN105867266B (en) | A kind of wisdom household operation device and management method | |
| CN108861930A (en) | The Internet of Things monitoring method and system of elevator operation | |
| CN112526945B (en) | Full-process monitoring and early warning system with feedback and closed-loop control functions | |
| CN109917734A (en) | Building equipment intelligent monitor system | |
| CN112083693B (en) | Paint production line information management system based on internet of things technology | |
| CN107087169A (en) | A kind of compression method, UVC video equipments and electronic equipment | |
| CN117250900B (en) | Artificial intelligent motor control system based on automatic identification operation mode |
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 |