CN115478574B - Excavator load prediction method based on radial basis function neural network - Google Patents
Excavator load prediction method based on radial basis function neural network Download PDFInfo
- Publication number
- CN115478574B CN115478574B CN202211347698.5A CN202211347698A CN115478574B CN 115478574 B CN115478574 B CN 115478574B CN 202211347698 A CN202211347698 A CN 202211347698A CN 115478574 B CN115478574 B CN 115478574B
- Authority
- CN
- China
- Prior art keywords
- hydraulic actuator
- excavator
- load
- bucket
- neural 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
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 17
- 238000003062 neural network model Methods 0.000 claims abstract description 19
- 230000003993 interaction Effects 0.000 claims abstract description 10
- 239000000463 material Substances 0.000 claims abstract description 10
- 230000008569 process Effects 0.000 claims abstract description 8
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 44
- 238000012549 training Methods 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 11
- 230000009471 action Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 9
- 210000002569 neuron Anatomy 0.000 claims description 9
- 230000005484 gravity Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 101100420683 Drosophila melanogaster Sarm gene Proteins 0.000 claims description 3
- 230000001133 acceleration Effects 0.000 claims description 3
- 230000004913 activation Effects 0.000 claims description 3
- 238000013016 damping Methods 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 239000010720 hydraulic oil Substances 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000003064 k means clustering Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 230000033772 system development Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000011217 control strategy Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
Classifications
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F3/00—Dredgers; Soil-shifting machines
- E02F3/04—Dredgers; Soil-shifting machines mechanically-driven
- E02F3/28—Dredgers; Soil-shifting machines mechanically-driven with digging tools mounted on a dipper- or bucket-arm, i.e. there is either one arm or a pair of arms, e.g. dippers, buckets
- E02F3/36—Component parts
- E02F3/42—Drives for dippers, buckets, dipper-arms or bucket-arms
- E02F3/43—Control of dipper or bucket position; Control of sequence of drive operations
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F3/00—Dredgers; Soil-shifting machines
- E02F3/04—Dredgers; Soil-shifting machines mechanically-driven
- E02F3/28—Dredgers; Soil-shifting machines mechanically-driven with digging tools mounted on a dipper- or bucket-arm, i.e. there is either one arm or a pair of arms, e.g. dippers, buckets
- E02F3/36—Component parts
- E02F3/42—Drives for dippers, buckets, dipper-arms or bucket-arms
- E02F3/43—Control of dipper or bucket position; Control of sequence of drive operations
- E02F3/435—Control of dipper or bucket position; Control of sequence of drive operations for dipper-arms, backhoes or the like
- E02F3/438—Memorising movements for repetition, e.g. play-back capability
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F9/00—Component parts of dredgers or soil-shifting machines, not restricted to one of the kinds covered by groups E02F3/00 - E02F7/00
- E02F9/20—Drives; Control devices
- E02F9/22—Hydraulic or pneumatic drives
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mechanical Engineering (AREA)
- Theoretical Computer Science (AREA)
- Structural Engineering (AREA)
- Civil Engineering (AREA)
- Mining & Mineral Resources (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Operation Control Of Excavators (AREA)
Abstract
The invention is suitable for the technical field of engineering machinery, and provides an excavator load prediction method based on a radial basis function neural network, which comprises the following steps: combining the structural load, the passive load and the interaction load in the working process of the excavator into the overall uncertainty of the excavator load; establishing an excavator load prediction model; pressure and flow signals of hydraulic actuators on the movable arm, the bucket rod and the bucket of the excavator are collected and input into a control unit; preprocessing the signal; constructing a feature vector by using the processed signals; the feature vector is input into a radial basis function neural network model, and the load of the excavator at the current moment is output. The method predicts the load of the excavator based on the radial basis function neural network model, does not depend on an accurate physical model and priori knowledge of material property parameters, accords with the complex and changeable actual working conditions of the excavator working environment, can realize accurate prediction of the real-time load of the excavator, and can be applied to intelligent excavating system development.
Description
Technical Field
The invention belongs to the technical field of engineering machinery, and particularly relates to an excavator load prediction method based on a radial basis function neural network.
Background
In order to improve the excavating efficiency of the hydraulic excavator and improve the fuel economy of the hydraulic excavator, the technology such as excavating track planning, intelligent excavating system development and the like become a research hot spot in recent years. However, due to the complex and variable working environments of the excavator, the load of the excavator has outstanding nonlinearity and uncertainty, and the implementation of the overall machine control strategy and the stability of the excavating process are seriously affected. The real-time and accurate prediction of the load of the hydraulic excavator is a core scientific problem behind the technologies of evaluating the instantaneous power output of the whole excavator, constructing an energy-saving and efficient excavating track, designing an automatic excavating control strategy in the full time domain and the like.
Due to the influence of sensor technology and complex working environments, predicting the load of an excavator is a complex task from the viewpoint of modeling and measurement. At present, three main methods are used for the research of the load of the excavator: discrete unit methods, finite element methods, and analytical expression methods. Although the discrete unit method and the finite element method have high simulation precision, the calculation cost of the discrete unit method and the finite element method involves a large number of numerical iterations, and the application of the discrete unit method and the finite element method in a real-time controller is severely limited. The analytical model has high calculation efficiency, but the accuracy of the analytical model is excessively dependent on material property parameters such as density, hardness, moisture, components and the like. The physical properties of materials vary widely and are not uniform, and parameter calibration and identification are required for each type of material. All of these have hampered engineering applications of traditional methods for predicting the load of an excavator, and there is a need for a method for predicting the load of an excavator with high accuracy that meets the real-time requirements of on-line prediction.
Disclosure of Invention
The embodiment of the invention aims to provide an excavator load prediction method based on a radial basis function neural network, which aims to solve the problems that the traditional excavator load prediction method based on a numerical simulation and analysis expression method is high in calculation cost and highly depends on priori knowledge of material property parameters.
The invention is realized in such a way that the excavator load prediction method based on the radial basis function neural network comprises the following steps:
step one, combining structural load, passive load and interaction load in the working process of the excavator into the overall uncertainty of the load of the excavator;
step two, establishing an excavator load prediction radial basis function neural network model;
step three, collecting pressure and flow signals of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator of the excavator, and inputting the collected signals into a control unit;
step four, preprocessing the signals acquired in the step three as original signals;
fifthly, constructing a feature vector according to the preprocessed signals;
and step six, inputting the constructed feature vector into an excavator load prediction radial basis function neural network model, and outputting the excavator load at the current moment.
In a further technical solution, in the first step, the structural load is generated by gravity and inertia of the working device, the passive load is generated by joint friction and viscous damping of hydraulic oil, and the interaction load is generated by the interaction of the bucket and the material.
According to a further technical scheme, the second step specifically comprises the steps of creating a radial basis function neural network model and training the model;
creating a radial basis function neural network model:
the radial basis function neural network adopts a three-layer network structure, and the input layer vector comprises the flow of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator at the last two moments, and also comprises the digging track of the tip of the bucket, the driving force of the three groups of hydraulic actuators under the action of structural load and the driving force of the three groups of hydraulic actuators which are actually measured; the hidden layer adopts 22 neurons, and a Gaussian function is used as an activation function; since the excavator load is difficult to directly measure and characterize, but is reflected in the driving forces of the three sets of hydraulic actuators, the driving forces of the boom hydraulic actuator, the arm hydraulic actuator and the bucket hydraulic actuator at the current moment are selected as network outputs;
model training:
acquiring pressure and flow signals of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator in the actual excavating process of the excavator, and constructing a training set; the training of the radial basis function neural network model consists of a self-organizing learning phase and a supervised learning phase: the self-organizing learning stage utilizes a k-means clustering algorithm and a p nearest neighbor algorithm to determine the center c of hidden layer neurons j And width sigma j The method comprises the steps of carrying out a first treatment on the surface of the In the supervised learning stage, the connection weight w between the hidden layer and the output layer is adaptively updated on line by using a recursive least square algorithm with forgetting factors, and the calculation formula is as follows:
wherein P (k) and L (k) are intermediate variables of the algorithm, phi (k) is an output vector of a hidden layer at the moment k, lambda is a forgetting factor, and y (k) is a system output value at the moment k.
According to a further technical scheme, the specific steps of preprocessing the original acquisition signals in the fourth step are as follows:
step A, denoising an original acquisition signal by utilizing wavelet filtering in a control unit, so as to reduce background noise and interference of a vibration high-frequency signal; selecting Daubechies 5 wavelet as wavelet basis function, adopting five-level wavelet decomposition and soft threshold function;
step B, analyzing the structural load dynamics characteristics of the excavator by using the filtered data; converting translational motion of a hydraulic actuator into relative rotational motion of a working device by utilizing an excavator kinematics model, and obtaining an excavating track of the bucket tip; the excavator dynamic model is utilized to calculate the load of the excavator structure, the influence of load uncertainty is reduced, the model prediction precision is improved, and the method is as follows:
wherein, theta,The rotation angle, the angular speed and the angular acceleration of the three working devices; m (θ) i ) Is an inertial matrix;characterizing the effects of centrifugal force and coriolis force as square terms of velocity; g (θ) i ) Is a gravity influence item; τ i Driving torque for each working device; f (F) sboom 、F sarm 、F sbucket Thrust forces of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator are respectively; e (E) 1 、E 2 、E 3 The hydraulic arm is an acting arm of force of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator to the joint;
step C, normalization; the signal is normalized as follows
Wherein x is the original acquired data, x min 、x max For the minimum value and the maximum value of a certain characteristic signal in the original acquired data, x scale Is the normalized signal value.
In a further technical scheme, in the fifth step, the specific steps of constructing the feature vector according to the preprocessed signal are as follows:
constructing eigenvectors of the preprocessed signals according to the sequence of the flow of the hydraulic actuator, the excavating track of the bucket tip, the driving force of the hydraulic actuator of the working device under the action of structural load and the driving force of the actually measured hydraulic actuator, wherein the eigenvectors are as follows:
X(t)＝[q i (t-1),q i (t-2),J y (t-1),J y (t-2),F si (t-1),F si (t-1),F i (t-1),F i (t-2)]
wherein i=1, 2,3 represent the hydraulic actuators of the boom, arm, bucket, respectively; q is the normalized discrete data of the flow signal of the hydraulic actuator; j (J) y Digging track discrete data for the normalized bucket tip; f (F) si Discrete driving force data of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator under the action of the normalized structural load; f (F) i For normalized actual measurementThe driving force signals of the boom hydraulic actuator, the arm hydraulic actuator and the bucket hydraulic actuator are discrete data.
In a further technical scheme, in the step six, a calculation formula of the excavator load at the current moment is output as follows:
wherein phi is j (t) is the output of the jth hidden layer neuron at time t;the predicted thrust of the hydraulic actuator of the movable arm, the bucket rod and the bucket at the t moment.
The excavator load prediction method based on the radial basis function neural network provided by the embodiment of the invention does not depend on a precise physical model and priori knowledge of material property parameters, is more in line with the complex and changeable actual working conditions of the excavator working environment, and meets the real-time requirement of the excavator load prediction; the load dynamic model of the excavator structure is fused with the radial basis function neural network model, so that the influence of load uncertainty can be reduced, and the load prediction of the excavator can be realized more accurately.
Drawings
Fig. 1 is a specific schematic diagram of an excavator load prediction method based on a radial basis function neural network according to an embodiment of the present invention;
FIG. 2 is a schematic working diagram of an excavator load prediction method based on a radial basis function neural network according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a radial basis function neural network prediction model of an excavator load prediction method based on a radial basis function neural network according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Specific implementations of the invention are described in detail below in connection with specific embodiments.
Example 1
An excavator load prediction method based on a radial basis function neural network comprises the following steps:
step one, combining structural load, passive load and interaction load in the working process of the excavator into the overall uncertainty of the load of the excavator;
structural loads are generated by the gravity and inertia of the working device, passive loads are generated by joint friction and viscous damping of hydraulic oil, and interaction loads are generated by the interaction of the bucket and materials;
step two, establishing an excavator load prediction radial basis function neural network model, which specifically comprises the steps of establishing a radial basis function neural network model and training the model;
creating a radial basis function neural network model:
the radial basis function neural network adopts a three-layer network structure, and the input layer vector comprises the flow of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator at the last two moments, and also comprises the digging track of the tip of the bucket, the driving force of the three groups of hydraulic actuators under the action of structural load and the driving force of the three groups of hydraulic actuators which are actually measured; the hidden layer adopts 22 neurons, and a Gaussian function is used as an activation function; since the excavator load is difficult to directly measure and characterize, but is reflected in the driving forces of the three sets of hydraulic actuators, the driving forces of the boom hydraulic actuator, the arm hydraulic actuator and the bucket hydraulic actuator at the current moment are selected as network outputs;
model training:
acquiring pressure and flow signals of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator in the actual excavating process of the excavator, and constructing a training set; the training of the radial basis function neural network model consists of a self-organizing learning phase and a supervised learning phase: self-organizing learning stage using k-means clusteringMethod and p nearest neighbor algorithm to determine center c of hidden layer neuron j And width sigma j The method comprises the steps of carrying out a first treatment on the surface of the In the supervised learning stage, the connection weight w between the hidden layer and the output layer is adaptively updated on line by using a recursive least square algorithm with forgetting factors, and the calculation formula is as follows:
wherein P (k) and L (k) are intermediate variables of an algorithm, phi (k) is an output vector of a hidden layer at the moment k, lambda is a forgetting factor, and y (k) is a system output value at the moment k;
step three, collecting pressure and flow signals of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator of the excavator, and inputting the collected signals into a control unit;
step four, preprocessing the signals acquired in the step three as original signals, and specifically comprising the following steps:
step A, denoising an original acquisition signal by utilizing wavelet filtering in a control unit, so as to reduce background noise and interference of a vibration high-frequency signal; selecting Daubechies 5 wavelet as wavelet basis function, adopting five-level wavelet decomposition and soft threshold function;
step B, analyzing the structural load dynamics characteristics of the excavator by using the filtered data; converting translational motion of a hydraulic actuator into relative rotational motion of a working device by utilizing an excavator kinematics model, and obtaining an excavating track of the bucket tip; the excavator dynamic model is utilized to calculate the load of the excavator structure, the influence of load uncertainty is reduced, the model prediction precision is improved, and the method is as follows:
wherein, theta,The rotation angle, the angular speed and the angular acceleration of the three working devices; m (θ) i ) Is an inertial matrix;characterizing the effects of centrifugal force and coriolis force as square terms of velocity; g (θ) i ) Is a gravity influence item; τ i Driving torque for each working device; f (F) sboom 、F sarm 、F sbucket Thrust forces of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator are respectively; e (E) 1 、E 2 、E 3 The hydraulic arm is an acting arm of force of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator to the joint;
step C, normalization: the signal is normalized as follows:
wherein x is the original acquired data, x min 、x max For the minimum value and the maximum value of a certain characteristic signal in the original acquired data, x scale The signal value after normalization processing;
step five, constructing a feature vector according to the preprocessed signals, wherein the specific steps are as follows:
constructing eigenvectors of the preprocessed signals according to the sequence of the flow of the hydraulic actuator, the excavating track of the bucket tip, the driving force of the hydraulic actuator of the working device under the action of structural load and the driving force of the actually measured hydraulic actuator, wherein the eigenvectors are as follows:
X(t)＝[q i (t-1),q i (t-2),J y (t-1),J y (t-2),F si (t-1),F si (t-1),F i (t-1),F i (t-2)]
wherein i=1, 2,3 represent the hydraulic actuators of the boom, arm, bucket, respectively; q is the normalized discrete data of the flow signal of the hydraulic actuator; j (J) y Digging track discrete data for the normalized bucket tip; f (F) si Discrete driving force data of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator under the action of the normalized structural load; f (F) i Discrete data of driving force signals of the normalized actual-measurement movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator;
inputting the constructed feature vector into an excavator load prediction radial basis function neural network model, outputting the excavator load at the current moment, and the calculation formula is as follows:
wherein phi is j (t) is the output of the jth hidden layer neuron at time t;the predicted thrust of the hydraulic actuator of the movable arm, the bucket rod and the bucket at the t moment.
The embodiment of the invention provides the excavator load prediction method based on the radial basis function neural network, which does not depend on a precise physical model and priori knowledge of material property parameters, is more in line with the complex and changeable actual working conditions of the excavator working environment, and meets the requirement of the real-time performance of the excavator load prediction; the load dynamic model of the excavator structure is fused with the radial basis function neural network model, so that the influence of load uncertainty can be reduced, the load prediction of the excavator can be realized more accurately, and the prediction result can be applied to the development of an intelligent excavating system.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.
Claims (1)
1. The excavator load prediction method based on the radial basis function neural network is characterized by comprising the following steps of:
step one, combining structural load, passive load and interaction load in the working process of the excavator into the overall uncertainty of the load of the excavator;
step two, establishing an excavator load prediction radial basis function neural network model;
step three, collecting pressure and flow signals of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator of the excavator, and inputting the collected signals into a control unit;
step four, preprocessing the signals acquired in the step three as original signals;
fifthly, constructing a feature vector according to the preprocessed signals;
step six, inputting the constructed feature vector into an excavator load prediction radial basis function neural network model, and outputting the excavator load at the current moment;
in the first step, structural load is generated by gravity and inertia of a working device, passive load is generated by joint friction and viscous damping of hydraulic oil, and interaction load is generated by interaction of a bucket and materials;
the second step specifically comprises the steps of creating a radial basis function neural network model and training the model;
creating a radial basis function neural network model:
the radial basis function neural network adopts a three-layer network structure, and the input layer vector comprises the flow of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator at the last two moments, and also comprises the digging track of the tip of the bucket, the driving force of the three groups of hydraulic actuators under the action of structural load and the driving force of the three groups of hydraulic actuators which are actually measured; the hidden layer adopts 22 neurons, and a Gaussian function is used as an activation function; since the excavator load is difficult to directly measure and characterize, but is reflected in the driving forces of the three sets of hydraulic actuators, the driving forces of the boom hydraulic actuator, the arm hydraulic actuator and the bucket hydraulic actuator at the current moment are selected as network outputs;
model training:
acquiring pressure and flow signals of a movable arm hydraulic actuator, a bucket rod hydraulic actuator and a bucket hydraulic actuator in the actual excavating process of the excavator, and constructing a training set; the training of the radial basis function neural network model consists of a self-organizing learning phase and a supervised learning phase: the self-organizing learning stage utilizes a k-means clustering algorithm and a p nearest neighbor algorithm to determine the center c of hidden layer neurons j And width sigma j The method comprises the steps of carrying out a first treatment on the surface of the In the supervised learning stage, the connection weight w between the hidden layer and the output layer is adaptively updated on line by using a recursive least square algorithm with forgetting factors, and the calculation formula is as follows:
wherein P (k) and L (k) are intermediate variables of an algorithm, phi (k) is an output vector of a hidden layer at the moment k, lambda is a forgetting factor, and y (k) is a system output value at the moment k;
the specific steps of preprocessing the original acquisition signals in the fourth step are as follows:
step A, denoising an original acquisition signal by utilizing wavelet filtering in a control unit, so as to reduce background noise and interference of a vibration high-frequency signal; selecting Daubechies 5 wavelet as wavelet basis function, adopting five-level wavelet decomposition and soft threshold function;
step B, analyzing the structural load dynamics characteristics of the excavator by using the filtered data; converting translational motion of a hydraulic actuator into relative rotational motion of a working device by utilizing an excavator kinematics model, and obtaining an excavating track of the bucket tip; the excavator dynamic model is utilized to calculate the load of the excavator structure, the influence of load uncertainty is reduced, the model prediction precision is improved, and the method is as follows:
wherein, theta,The rotation angle, the angular speed and the angular acceleration of the three working devices; m (θ) i ) Is an inertial matrix; />Characterizing the effects of centrifugal force and coriolis force as square terms of velocity; g (θ) i ) Is a gravity influence item; τ i Driving torque for each working device; f (F) sboom 、F sarm 、F sbucket Thrust forces of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator are respectively; e (E) 1 、E 2 、E 3 The hydraulic arm is an acting arm of force of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator to the joint;
step C, normalization; the signal is normalized as follows
Wherein x is the original acquired data, x min 、x max For the minimum value and the maximum value of a certain characteristic signal in the original acquired data, x scale The signal value after normalization processing;
in the fifth step, the specific steps of constructing the feature vector according to the preprocessed signal are as follows:
constructing eigenvectors of the preprocessed signals according to the sequence of the flow of the hydraulic actuator, the excavating track of the bucket tip, the driving force of the hydraulic actuator of the working device under the action of structural load and the driving force of the actually measured hydraulic actuator, wherein the eigenvectors are as follows:
X(t)＝[q i (t-1),q i (t-2),J y (t-1),J y (t-2),F si (t-1),F si (t-1),F i (t-1),F i (t-2)]
wherein i=1, 2,3 represent the hydraulic actuators of the boom, arm, bucket, respectively; q is the normalized discrete data of the flow signal of the hydraulic actuator; j (J) y Digging track discrete data for the normalized bucket tip; f (F) si Discrete driving force data of the movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator under the action of the normalized structural load; f (F) i Discrete data of driving force signals of the normalized actual-measurement movable arm hydraulic actuator, the bucket rod hydraulic actuator and the bucket hydraulic actuator;
in the sixth step, a calculation formula of the excavator load at the current moment is output as follows:
wherein phi is j (t) is the output of the jth hidden layer neuron at time t;the predicted thrust of the hydraulic actuator of the movable arm, the bucket rod and the bucket at the t moment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211347698.5A CN115478574B (en) | 2022-10-31 | 2022-10-31 | Excavator load prediction method based on radial basis function neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211347698.5A CN115478574B (en) | 2022-10-31 | 2022-10-31 | Excavator load prediction method based on radial basis function neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115478574A CN115478574A (en) | 2022-12-16 |
| CN115478574B true CN115478574B (en) | 2024-03-19 |
Family
ID=84396779
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211347698.5A Active CN115478574B (en) | 2022-10-31 | 2022-10-31 | Excavator load prediction method based on radial basis function neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115478574B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110565711A (en) * | 2019-09-20 | 2019-12-13 | 太原科技大学 | Track control system and track planning method for backhoe hydraulic excavator |
| KR20200012584A (en) * | 2018-07-27 | 2020-02-05 | 건설기계부품연구원 | Method of Measuring Work Load for Excavators and system thereof |
| CN112681443A (en) * | 2021-01-19 | 2021-04-20 | 山西创智卓越科技有限公司 | Excavating robot joint track control method and control system |
| CN113657687A (en) * | 2021-08-30 | 2021-11-16 | 国家电网有限公司 | Power load prediction method based on feature engineering and multi-path deep learning |
| CN113988369A (en) * | 2021-09-23 | 2022-01-28 | 上海三一重机股份有限公司 | Prediction model training method and prediction method based on prediction model |
| CN114792067A (en) * | 2022-04-08 | 2022-07-26 | 大连理工大学 | Excavator working phase classification and identification method based on ResNet neural network |
-
2022
- 2022-10-31 CN CN202211347698.5A patent/CN115478574B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20200012584A (en) * | 2018-07-27 | 2020-02-05 | 건설기계부품연구원 | Method of Measuring Work Load for Excavators and system thereof |
| CN110565711A (en) * | 2019-09-20 | 2019-12-13 | 太原科技大学 | Track control system and track planning method for backhoe hydraulic excavator |
| CN112681443A (en) * | 2021-01-19 | 2021-04-20 | 山西创智卓越科技有限公司 | Excavating robot joint track control method and control system |
| CN113657687A (en) * | 2021-08-30 | 2021-11-16 | 国家电网有限公司 | Power load prediction method based on feature engineering and multi-path deep learning |
| CN113988369A (en) * | 2021-09-23 | 2022-01-28 | 上海三一重机股份有限公司 | Prediction model training method and prediction method based on prediction model |
| CN114792067A (en) * | 2022-04-08 | 2022-07-26 | 大连理工大学 | Excavator working phase classification and identification method based on ResNet neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115478574A (en) | 2022-12-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111931411B (en) | Duhem dynamic hysteresis modeling method for piezoelectric driving micro-positioning platform | |
| CN106125574B (en) | Piezoelectric ceramics mini positioning platform modeling method based on DPI model | |
| CN101105126B (en) | Logging-while-drilling orientation measurement error compensation method based on micro-quartz angular rate sensor | |
| CN112276944A (en) | Man-machine cooperation system control method based on intention recognition | |
| Huo et al. | Intelligent prediction for digging load of hydraulic excavators based on RBF neural network | |
| CN114169230A (en) | Robot dynamics parameter identification method | |
| Liu et al. | Modeling of hysteresis in piezoelectric actuator based on adaptive filter | |
| İlman et al. | Generalized input preshaping vibration control approach for multi-link flexible manipulators using machine intelligence | |
| CN114310911B (en) | Driving joint dynamic error prediction and compensation system and method based on neural network | |
| CN110504878B (en) | Soft measurement method for rotor speed and displacement of bearingless permanent magnet synchronous motor | |
| CN109815525B (en) | Multi-objective optimization design method for dynamic characteristics of feeding system | |
| CN115478574B (en) | Excavator load prediction method based on radial basis function neural network | |
| Ding et al. | Terminal force soft sensing of hydraulic manipulator based on the parameter identification | |
| CN111554355B (en) | Molecular dynamics calculation method based on non-von neumann architecture | |
| Lin et al. | DSP-based self-organising fuzzy controller for active suspension systems | |
| CN111539157A (en) | Dynamic load time domain identification method based on multilayer perceptron | |
| Chen | Determination of flutter derivatives via a neural network approach | |
| Borges et al. | A neural network-based inversion method of a feedback linearization controller applied to a hydraulic actuator | |
| Le | Design of intelligent controller using type-2 fuzzy cerebellar model articulation and 3d membership functions | |
| Miettinen et al. | Bidirectional lstm-based soft sensor for rotor displacement trajectory estimation | |
| Park et al. | Data driven analytics (machine learning) for system characterization, diagnostics and control optimization | |
| Liu et al. | [Retracted] The Use of Deep Learning Model for Effect Analysis of Conventional Friction Power Confinement | |
| Wu et al. | Simulation Analysis of Dynamic Characteristics of AC Motor Based on BP Neural Network Algorithm | |
| Zhang et al. | A study on trajectory control system of hydraulic excavators based on multi-domain physical model | |
| Liu et al. | Application of LabVIEW and MATLAB hybrid programming in unmanned helicopter system simulation |
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 |