CN108363836B - Multi-working-condition self-adaptive industrial robot health degree assessment method and system - Google Patents

Multi-working-condition self-adaptive industrial robot health degree assessment method and system Download PDF

Info

Publication number
CN108363836B
CN108363836B CN201810045777.8A CN201810045777A CN108363836B CN 108363836 B CN108363836 B CN 108363836B CN 201810045777 A CN201810045777 A CN 201810045777A CN 108363836 B CN108363836 B CN 108363836B
Authority
CN
China
Prior art keywords
working condition
data
health degree
robot
health
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
Application number
CN201810045777.8A
Other languages
Chinese (zh)
Other versions
CN108363836A (en
Inventor
张开桓
蔡一彪
吴芳基
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou AIMS Intelligent Technology Co Ltd
Original Assignee
Hangzhou AIMS Intelligent Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hangzhou AIMS Intelligent Technology Co Ltd filed Critical Hangzhou AIMS Intelligent Technology Co Ltd
Priority to CN201810045777.8A priority Critical patent/CN108363836B/en
Publication of CN108363836A publication Critical patent/CN108363836A/en
Application granted granted Critical
Publication of CN108363836B publication Critical patent/CN108363836B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06395—Quality analysis or management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/04—Ageing analysis or optimisation against ageing

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Evolutionary Computation (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Educational Administration (AREA)
  • Development Economics (AREA)
  • General Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • Biophysics (AREA)
  • Computer Hardware Design (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Game Theory and Decision Science (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Manipulator (AREA)

Abstract

The invention relates to a multi-working-condition self-adaptive industrial robot health degree assessment method and system, which comprises the following steps: acquiring running state data, working condition information data and design parameter data of each joint of the industrial robot; processing the running state data to obtain a state characteristic database, and matching the working condition information data to obtain a working condition information database; establishing a correlation model of the state characteristics and the working condition parameters, and carrying out multi-dimensional adaptive processing on the state characteristic data through the correlation model to obtain characteristic data with desensitized working conditions; calculating health degree indexes of all joints based on the working condition desensitization characteristic data; calculating and evaluating the overall health degree of the robot based on the health degree indexes of all joints; and predicting the remaining service life of the robot based on the health degree of the whole robot, the working condition information data and the design parameter data. By the method and the system, the accuracy of health degree evaluation and residual service life prediction of the industrial robot can be effectively improved, and the accuracy and the adaptability of the model and the system can be continuously improved through online self-learning.

Description

Multi-working-condition self-adaptive industrial robot health degree assessment method and system
Technical Field
The invention relates to the field of robot health management and service life prediction, in particular to a multi-working-condition self-adaptive industrial robot health degree assessment method and system.
Background
In recent years, the mode of 'robot changing' is rapidly rising in the process of transformation and upgrading from the domestic manufacturing industry to modernization, automation and intelligence. Along with the rapid increase of the number of robots used in China, the problems of operation, management, safety, maintenance, repair and the like are becoming more and more serious. On an automatic production line taking a robot as a main part, the problems of unexpected shutdown caused by various faults of the robot or product quality fluctuation caused by performance degradation and the like can cause huge economic loss. Therefore, how to effectively evaluate the health state of the robot and timely make effective preventive maintenance measures is very important.
The predictive maintenance and health management (PHM) technology can accurately sense the performance state of the robot on one hand, provide reliable index basis for 'maintenance according to the situation' and effectively solve the problems of insufficient maintenance and excessive maintenance in the traditional mode by monitoring the running states of the robot and parts thereof on a production line and combining an effective health degree evaluation method; on the other hand, the state can be predicted and managed in advance based on indexes such as health degree, the fault shutdown is avoided, the stable operation of a production system is guaranteed, and the value creation targets of cost reduction, efficiency improvement and quality stabilization on an industrial production line are further achieved. In the technology, the evaluation of the health degree plays a key role in the starting and the ending, so the effectiveness, the accuracy and the reliability of the evaluation method are important.
However, the application field of the robot is wide, and the related operation tasks are complex and various, so that the robot often works under a complex and changeable working condition environment. The information of the changed working conditions such as the operation type, the operation speed, the load and the like can be coupled in the operation state parameters of the robot, so that the real performance state and the change rule of the robot are covered. The traditional robot health degree evaluation method does not consider the operation condition data of the robot, and cannot accurately reflect the real health state of the robot, so that the obtained evaluation result is greatly different from the actual situation.
Disclosure of Invention
The invention provides a multi-working-condition self-adaptive industrial robot health degree assessment method and system, aiming at the problem of accurate assessment of health degree of an industrial robot under the multi-working condition in the prior art.
A multi-working-condition self-adaptive industrial robot health degree assessment method comprises the following steps:
acquiring operating state data, working condition information data and robot-related design parameter data of each joint of an industrial robot, processing the operating state data to obtain a state characteristic database, and matching the state characteristic database with corresponding working condition information data to obtain a working condition information database;
establishing a characteristic and working condition parameter association model through a working condition information database and a state characteristic database, and carrying out multi-dimensional adaptive processing on characteristic data in the state characteristic database through the association model to obtain characteristic data with desensitized working conditions;
standardizing the characteristic data desensitized to the working condition, and combining a distance method to obtain health degree indexes of all joints;
calculating the whole health degree of the robot according to the health degree index of each joint and a preset health degree threshold value of each joint, and performing overall evaluation on the health state of the robot according to the whole health degree of the robot;
and predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data and the design parameter data.
As an implementation manner, the processing of the operation state data refers to performing feature extraction and feature selection on the operation state data;
the feature extraction comprises the basic feature extraction of a signal time domain and a signal frequency domain and the secondary feature extraction after the signal is transformed and decomposed;
the characteristic selection means that the characteristics with definite representation meaning or tendency are selected by combining the physical meaning or the requirement of the characteristics to obtain required state characteristic data; and if the characteristics without clear representation meaning or tendency need to be selected, screening by combining the dimensionality reduction of a principal component analysis method to obtain required state characteristic data.
As an implementation manner, the specific steps of establishing a characteristic and working condition parameter association model through a working condition information database and a state characteristic database, and performing multidimensional adaptive processing on the characteristic data in the state characteristic database through the association model to obtain the characteristic data with desensitized working conditions include:
decomposing the characteristic data in the working condition information database into a plurality of mutually independent working condition parameter dimensions, establishing association models of each characteristic in the characteristic database and the working condition parameters under different dimensions, and recording the association models as first association models;
and matching the selected characteristic data to the corresponding working condition parameter dimension, and compensating and eliminating the influence on the working condition according to the established first correlation model under the working condition parameter dimension to obtain the characteristic data with desensitized working condition.
As an implementation mode, the specific steps of performing normalization processing on the characteristic data of desensitization of the working condition and obtaining the health index of each joint by combining a distance method include:
calculating Euclidean distance D from the selected characteristic vector to a healthy reference vector through standardized characteristic data, wherein the healthy reference vector is obtained by averaging standardized characteristic value data corresponding to the characteristic vector selected when the robot leaves a factory and is in an initial stable state when the robot leaves the factory;
and obtaining a normalization function through Sigmoid function transformation, and carrying out normalization processing on the Euclidean distance D through the normalization function to obtain the health degree index of each joint.
As an implementation manner, the transformation by Sigmoid function results in a normalization function, which is expressed as:
Figure BDA0001550859880000031
wherein w is a scale parameter, t is a smoothing parameter, and HI represents a health index of each joint.
As an implementation manner, the specific operation of calculating the overall health degree of the robot according to the index of the health degree of each joint and a preset threshold of the health degree of each joint, and performing overall evaluation on the health state of the robot through the overall health degree of the robot is as follows:
constructing a health radar map through the health degree indexes of all joints, and calculating the health degree of the whole machine according to the area of the health radar map and a preset health degree threshold, wherein the health degree of the whole machine is expressed as follows:
Figure BDA0001550859880000032
wherein HT is a preset health threshold, A is the current state health radar map area, A1J is 1,2, which is an area of a health degree radar map when all the joint health degrees are 1.
As an implementation manner, the specific steps of predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data and the design parameter data include:
establishing a correlation model between the remaining service life and the health degree of the whole machine according to the health degree of the whole machine and the design parameter data, and recording the correlation model as a second correlation model;
establishing a correlation model between the influence degree of the service life and the working condition parameters by combining the influence degrees of various working conditions on the service life and the design parameter data, and recording the correlation model as a third correlation model;
and establishing a residual service life prediction model after the working condition parameters are corrected by combining the two correlation models according to the health degree data of the whole machine, and predicting the residual service life under the actual working condition.
As an implementation, the method further comprises the following steps:
an updating step: the state characteristic database and the working condition information database matched with the state characteristic database can be continuously updated according to the increase of the running state data and the working condition information data of the robot;
self-learning step: the first correlation model, the second correlation model and the third correlation model can be self-learned and optimized.
A multi-working condition self-adaptive industrial robot health degree evaluation system comprises the following modules:
a data acquisition module: acquiring running state data, working condition information data and robot related design parameter data of each joint of the industrial robot;
a data processing module: processing the running state data to obtain a state characteristic database, and matching the state characteristic database with corresponding working condition information data to obtain a working condition information database;
the working condition self-adaptive processing module comprises: establishing a characteristic and working condition parameter association model through a working condition information database and a state characteristic database, and carrying out multi-dimensional adaptive processing on characteristic data in the state characteristic database through the association model to obtain characteristic data with desensitized working conditions;
each joint health degree calculation module: standardizing the characteristic data desensitized to the working condition, and combining a distance method to obtain health degree indexes of all joints;
the whole machine health degree evaluation module: calculating the whole health degree of the robot according to the health degree index of each joint and a preset health degree threshold value of each joint, and performing overall evaluation on the health state of the robot according to the whole health degree of the robot;
a remaining service life prediction module: predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data and the design parameter data;
updating the self-learning module: the state characteristic database and the working condition information database matched with the state characteristic database can be continuously updated according to the increase of the running state data and the working condition information data of the robot; the first correlation model, the second correlation model and the third correlation model can be self-learned and optimized.
Due to the adoption of the technical scheme, the invention has the remarkable technical effects that:
the invention provides a multi-working-condition self-adaptive industrial robot health degree evaluation method and system, aiming at the situation that the working conditions of an industrial robot are complex and changeable under different application scenes, the correlation model self-adaptive processing is carried out to realize working condition desensitization, and health degree calculation and evaluation based on an Euclidean distance method and residual service life prediction combined with the working conditions are completed, so that the accuracy of the industrial robot health degree evaluation and the residual service life prediction is effectively improved, and the accuracy and the adaptability of the model and the system can be continuously improved through on-line self-learning.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic overall flow diagram of the present invention;
FIG. 2 is a schematic flow diagram of FIG. 1;
FIG. 3 is a schematic diagram of the overall system architecture of the present invention;
FIG. 4 is a diagram illustrating the behavior parameter variation corresponding to the time sequence of the multi-behavior sample according to an embodiment of the present invention;
FIG. 5 is a comparison of RMS value conditions of a third joint of an industrial robot before and after adaptive processing according to an embodiment of the invention;
6 a-6 f are comparison results of health evaluation of two industrial robots with different use time lengths under multiple working conditions in the specific embodiment;
FIG. 7 is a comparison of health radar maps of two industrial robots with different use durations under a certain working condition in an embodiment;
FIG. 8 is a comparison of radar maps of health of an industrial robot before and after a simulated fault under certain operating conditions in an embodiment of the present invention;
fig. 9 shows the predicted remaining service life of an industrial robot under multiple operating conditions in an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples, which are illustrative of the present invention and are not to be construed as being limited thereto.
Example 1:
a health degree evaluation method of a multi-working-condition self-adaptive industrial robot is shown in figure 1 and comprises the following steps:
s1, acquiring running state data, working condition information data and robot related design parameter data of each joint of the industrial robot, processing the running state data to obtain a state characteristic database, and matching the corresponding working condition information data with the state characteristic database to obtain a working condition information database;
s2, establishing a characteristic and working condition parameter association model through a working condition information database and a state characteristic database, and carrying out multi-dimensional adaptive processing on the characteristic data in the state characteristic database through the association model to obtain characteristic data with desensitized working conditions;
s3, carrying out standardization processing on the characteristic data of the working condition desensitization, and combining a distance method to obtain health degree indexes of each joint;
s4, calculating the overall health degree of the robot according to the health degree indexes of the joints and preset health degree threshold values of the joints, and performing overall evaluation on the health state of the robot according to the overall health degree of the robot;
and S5, predicting the remaining service life of the robot according to the whole machine health degree, the working condition information data and the design parameter data.
Further, in step S1, the processing the operation state data refers to performing feature extraction and feature selection on the operation state data;
the feature extraction comprises the basic feature extraction of a signal time domain and a signal frequency domain and the secondary feature extraction after the signal is transformed and decomposed;
the characteristic selection means that the characteristics with definite representation meaning or tendency are selected by combining the physical meaning or the requirement of the characteristics to obtain required state characteristic data; and if the characteristics without clear representation meaning or tendency need to be selected, screening by combining the dimensionality reduction of a principal component analysis method to obtain required state characteristic data.
In step S2, the step of establishing a characteristic and condition parameter association model by using the condition information database and the state characteristic database, and performing multidimensional adaptive processing on the characteristic data in the state characteristic database by using the association model to obtain characteristic data with desensitized conditions includes:
s21, decomposing the characteristic data in the working condition information database into a plurality of mutually independent working condition parameter dimensions, establishing a correlation model of each characteristic and working condition parameter in the characteristic database under different dimensions, and recording the correlation model as a first correlation modelijA first correlation model, where i 1,2,.. and P are P selected features, and j1, 2.. and K are K operating condition parameter dimensions;
s22, matching the selected characteristic data to corresponding working condition parameter dimensions, wherein the number of the working condition parameter dimensions is K, and compensating and eliminating the influence on the working condition under the K working condition parameter dimensions according to the establishment of the first correlation model to obtain the characteristic data with desensitized working conditions, which is specifically expressed as:
Figure BDA0001550859880000071
wherein i 1,2iFor the ith selected feature value, VjIs the working condition parameter corresponding to the j dimension,
Figure BDA0001550859880000074
the value of the ith selected feature after all operating condition effects have been removed, here,
Figure BDA0001550859880000075
the characteristic data of the desensitization of the obtained working condition is obtained.
In step S3, the step of normalizing the characteristic data for desensitization of the operating condition and obtaining the health index of each joint by combining the distance method includes:
s31, carrying out standardization processing on the characteristic data with the working condition desensitization to obtain standardized characteristic data, wherein the standardization processing is carried out to eliminate dimension and magnitude influences to obtain a standardized characteristic value S without the working condition influencesiWherein, (i ═ 1, 2.... P), further, the normalization processing method selects dispersion normalization so that the normalized feature value falls within [0,1 ], n]Interval:
Figure BDA0001550859880000072
wherein max and min are respectively the maximum value and the minimum value of the number of samples; siThe value of the normalized characteristic is represented,
Figure BDA0001550859880000073
the value of the ith selected feature after all operating conditions have been removed.
S32, feature by normalizationData calculate the Euclidean distance D of the sample vector to the healthy reference vector, here the healthy reference vector yBThe health reference vector y is obtained by averaging the normalized characteristic value data corresponding to the sample set in the initial stable state when the robot leaves the factoryBExpressed as:
Figure BDA0001550859880000081
wherein, the sample set of the initial steady state is n multiplied by P dimension, n represents the number of samples, and P is the characteristic dimension;
the euclidean distance D is expressed as:
Figure BDA0001550859880000082
where y denotes a sample vector, yBRepresenting a health reference vector; the sample vectors are selected eigenvectors, and the samples are combined into a group of eigenvectors selected when leaving a factory;
and S33, obtaining a normalization function through Sigmoid function transformation, and carrying out normalization processing on the Euclidean distance D through the normalization function to obtain the health degree index of each joint.
In this embodiment, the normalization function is obtained through Sigmoid function transformation, and the normalization function is expressed as:
Figure BDA0001550859880000083
wherein w is a scale parameter, t is a smoothing parameter, and HI represents a health index of each joint.
In step S4, the specific operation of calculating the overall health degree of the robot according to the index of the health degree of each joint and the preset threshold of the health degree of each joint, and performing overall evaluation on the health state of the robot according to the overall health degree of the robot is as follows:
constructing a health radar map through the health degree indexes of all joints, and calculating the health degree of the whole machine according to the area of the health radar map and a preset health degree threshold, wherein the health degree of the whole machine is expressed as follows:
Figure BDA0001550859880000084
wherein HT is a preset health threshold, A is the current state health radar map area, A1J is 1,2, which is an area of a health degree radar map when all the joint health degrees are 1.
More specifically, in step S5, the specific step of predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data, and the design parameter data includes:
s51, establishing a correlation model between the remaining service life and the whole machine health degree according to the whole machine health degree and the design parameter data, recording the correlation model as a second correlation model, and using MRULRepresents;
s52, establishing a correlation model between the life influence degree and the working condition parameters by combining the influence degree of various working conditions on the service life and the design parameter data, recording the correlation model as a third correlation model, and using IMjRepresenting j as 1, 2.. and K as K operating condition parameter dimensions;
and S53, establishing a residual service life prediction model after working condition parameter correction according to the health degree data of the whole machine and combining the two correlation models, predicting the residual service life under the actual working condition, and more specifically, correcting according to the working condition when predicting the residual service life so that the correlation prediction value is more accurate. The method for predicting the residual service life in association and correcting the residual service life according to the working condition comprises the following steps:
Figure BDA0001550859880000091
wherein RUL is the predicted residual service life after the working condition parameters are corrected, RUL0To essentially predict remaining service life without regard to the effects of operating conditions, IFjThe effect of j-th working condition parameter dimension on the service lifeLoudness, H is the health, Parm, obtained in step S4iFor the relevant design parameter of the robot, VjThe parameters are the operating parameters corresponding to the j-th dimension, wherein j is 1, 2.
In this embodiment, the method further comprises the steps of:
an updating step: the state characteristic database and the working condition information database matched with the state characteristic database can be continuously updated according to the increase of the running state data and the working condition information data of the robot;
self-learning step: self-learning and optimization of the first correlation model MijSecond correlation model MRULAnd a third correlation model IMj。
A health degree evaluation system of an industrial robot with adaptive multi-working condition is shown in figure 3 and comprises the following modules:
the data acquisition module 1: acquiring running state data, working condition information data and robot related design parameter data of each joint of the industrial robot;
the data processing module 2: processing the running state data to obtain a state characteristic database, and matching the state characteristic database with corresponding working condition information data to obtain a working condition information database;
and the working condition self-adaptive processing module 3: establishing a characteristic and working condition parameter association model through a working condition information database and a state characteristic database, and carrying out multi-dimensional adaptive processing on characteristic data in the state characteristic database through the association model to obtain characteristic data with desensitized working conditions;
each joint health degree calculation module 4: standardizing the characteristic data desensitized to the working condition, and combining a distance method to obtain health degree indexes of all joints;
the whole health degree evaluation module 5: calculating the whole health degree of the robot according to the health degree index of each joint and a preset health degree threshold value of each joint, and performing overall evaluation on the health state of the robot according to the whole health degree of the robot;
remaining service life prediction module 6: predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data and the design parameter data;
updating the self-learning module 7: the state characteristic database and the working condition information database matched with the state characteristic database can be continuously updated according to the increase of the running state data and the working condition information data of the robot; the first correlation model, the second correlation model and the third correlation model can be self-learned and optimized.
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
The technical effects of the invention are described by combining the practical cases as follows:
the industrial robot described in this embodiment includes two identical six-axis articulated industrial robots of the same model, which have different service lives and other design parameters.
Referring to fig. 1 and 2, collected relevant design parameter data of the industrial robot comprises design service life, rated speed, rated load, service life, theoretical decline rule parameters and the like, the working condition information data of the robot mainly comprises real-time integral operation speed and load of the robot, and the operation state data of each joint of the robot mainly comprises torque and current data of each joint; the state feature database is obtained by extracting the time domain features of signals and the frequency domain features and selecting the features of the torque or current data of each joint. The feature selection is mainly selected according to the signal variation characteristics, and the mainly selected features are root mean square value (RMS), rectified mean value (ARV), peak-to-peak value (P-P) and standard deviation (STD), which are all data processing processes in the early stage and prepare for subsequent modeling and prediction.
In order to verify the health assessment method for the multi-working-condition adaptive industrial robot provided by the invention, in this embodiment, a plurality of working conditions caused by speed and load changes are designed for the six-axis joint industrial robot, and the working condition parameter changes corresponding to the time sequence of the sample are shown in fig. 4.
Fig. 5 shows a comparison between the RMS of the third joint in the time series of the sample before and after the multi-condition adaptive processing, and it can be seen that, before the condition adaptive processing, the time-domain characteristic RMS value is greatly affected by the operating speed and the applied load, and as a whole, the time-domain characteristic RMS value shows an increasing trend along with the speed and the load, and the fluctuation increases at a higher speed. After the adaptive working condition processing is carried out by adopting the method provided by the invention, the RMS value basically realizes the desensitization of the working condition, but simultaneously retains the fluctuation information at high speed, in FIG. 5, the upper diagram shows the diagram when the adaptive working condition processing is not carried out, the diagram shows that the time domain characteristic RMS value is influenced by the running speed and the applied load, the lower diagram shows the diagram when the adaptive working condition processing is carried out, and the RMS value basically realizes the desensitization of the working condition, but simultaneously retains the fluctuation information at high speed.
In this embodiment, two six-axis joint industrial robots of the same model, the same working condition, and different service lives are further evaluated and compared for health degree, the changes of working condition parameters corresponding to two sample time sequences are shown in fig. 4, and fig. 6a to 6e are joint health degree index comparisons labeled as J1 to J6 obtained after steps S1 to S3, wherein each joint health degree index of a newly purchased industrial robot using only 6 months is basically close to an initial health reference value, and the health degree value is basically stable and the health condition is good when the working condition changes; the other is that the health degree of other joints except the joint marked as J4 of the robot used for 30 months is reduced to different degrees, and the health degree loss can be visually reflected when the health degree is increased along with the service life through the plurality of figures.
Fig. 7 is a health degree radar map corresponding to each state, which is obtained after health degree indexes of joints labeled J1-J6 in two samples are calculated, and the health degree radar map can represent that the method provided by the invention can effectively realize multi-working-condition self-adaptive health degree evaluation of the industrial robot.
Further, in order to evaluate the effect of the method for evaluating the health degree of the multi-working-condition adaptive industrial robot in the potential fault state, in the embodiment, the simple fault state is simulated by strongly binding the joints, which are labeled as J3, of the industrial robot with the service life of 6 months. Fig. 8 shows the comparison effect of health before and after the simulated fault after steps S1-S4, and it can be seen that, compared with the normal state, the health degree of the joints marked as J1, J4, J5 and J6 has no obvious change, the health degree value of the joint marked as J3 directly affected by the simulated fault is obviously reduced, and the health degree value of the joint marked as J2 indirectly affected by the simulated fault is reduced in a small degree.
In order to verify the effect of the method for estimating the health of the industrial robot under the adaptive multi-working conditions provided by the invention on predicting the remaining service life of the whole industrial robot, in the embodiment, the six-axis multi-joint industrial robot is operated under various working condition parameters shown in fig. 4, and the remaining service life under different working conditions is predicted after the steps S1-S5, and fig. 9 is a corresponding prediction result, so that for the six-axis multi-joint industrial robot with the design service life of about 15 years, the predicted remaining service life under the low-speed and low-load operation state is basically consistent with the difference between the design service life and the actual service life, which indicates the effectiveness of the service life prediction, and meanwhile, when the working conditions are changed according to the sequence shown in fig. 4, the remaining service life under the corresponding working conditions is attenuated to different degrees according to the working conditions, the influence of parameters such as working conditions on the residual service life of the industrial robot in the actual application environment is reflected, so that the residual service life of the industrial robot can be more accurately predicted according to the actual application working conditions, the environment and the like, and more effective maintenance scheduling and predictive maintenance are performed.
In addition, it should be noted that the specific embodiments described in the present specification may differ in the shape of the components, the names of the components, and the like. All equivalent or simple changes of the structure, the characteristics and the principle of the invention which are described in the patent conception of the invention are included in the protection scope of the patent of the invention. Various modifications, additions and substitutions for the specific embodiments described may be made by those skilled in the art without departing from the scope of the invention as defined in the accompanying claims.

Claims (8)

1. A health degree assessment method for a multi-working-condition self-adaptive industrial robot is characterized by comprising the following steps:
acquiring operating state data, working condition information data and robot-related design parameter data of each joint of an industrial robot, processing the operating state data to obtain a state characteristic database, and matching the state characteristic database with corresponding working condition information data to obtain a working condition information database;
decomposing the characteristic data in the working condition information database into a plurality of mutually independent working condition parameter dimensions, establishing association models of each characteristic in the characteristic database and the working condition parameters under different dimensions, and recording the association models as first association models;
matching the selected characteristic data to the corresponding working condition parameter dimension, and compensating and eliminating the influence on the working condition according to the established first correlation model under the working condition parameter dimension to obtain characteristic data with desensitized working condition;
standardizing the characteristic data desensitized to the working condition, and combining a distance method to obtain health degree indexes of all joints;
calculating the whole health degree of the robot according to the health degree index of each joint and a preset health degree threshold value of each joint, and performing overall evaluation on the health state of the robot according to the whole health degree of the robot;
and predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data and the design parameter data.
2. The method for evaluating the health degree of the multi-working-condition adaptive industrial robot according to claim 1, wherein the processing of the operation state data refers to feature extraction and feature selection of the operation state data;
the feature extraction comprises the basic feature extraction of a signal time domain and a signal frequency domain and the secondary feature extraction after the signal is transformed and decomposed;
the characteristic selection means that the characteristics with definite representation meaning or tendency are selected by combining the physical meaning or the requirement of the characteristics to obtain required state characteristic data; and if the characteristics without clear representation meaning or tendency need to be selected, screening by combining the dimensionality reduction of a principal component analysis method to obtain required state characteristic data.
3. The method for evaluating the health degree of the multi-working-condition self-adaptive industrial robot according to claim 1, wherein the specific steps of carrying out standardization processing on the characteristic data desensitized to the working conditions and combining a distance method to obtain the health degree index of each joint comprise:
standardizing the characteristic data with desensitized working conditions to obtain standardized characteristic data;
calculating Euclidean distance D from the selected feature vector to a healthy reference vector through standardized feature data, wherein the healthy reference vector is obtained by averaging the standardized feature data corresponding to the selected feature vector when the robot leaves a factory and is in an initial stable state;
and obtaining a normalization function through Sigmoid function transformation, and carrying out normalization processing on the Euclidean distance D through the normalization function to obtain the health degree index of each joint.
4. The method for assessing the health of a multi-condition adaptive industrial robot according to claim 3, wherein the normalization function is obtained by Sigmoid function transformation, and is expressed as:
Figure DEST_PATH_IMAGE001
wherein w is a scale parameter, t is a smoothing parameter, HI represents a health index of each joint, and D represents an Euclidean distance.
5. The method for evaluating the health degree of the multi-working-condition self-adaptive industrial robot according to claim 4, wherein the specific operation of calculating the overall health degree of the robot according to the index of the health degree of each joint and a preset threshold value of the health degree of each joint and integrally evaluating the health state of the robot through the overall health degree of the robot is as follows:
constructing a health radar map through the health degree indexes of all joints, and calculating the health degree of the whole machine according to the area of the health radar map and a preset health degree threshold, wherein the health degree of the whole machine is expressed as follows:
Figure 41363DEST_PATH_IMAGE002
HT is a preset health threshold, a is an area of a current state health radar map, a1 is an area of a health radar map when all joint health degrees are 1, and j is 1, 2.
6. The method for assessing the health degree of the multi-working-condition adaptive industrial robot according to claim 5, wherein the specific steps of predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data and the design parameter data comprise:
establishing a correlation model between the remaining service life and the health degree of the whole machine according to the health degree of the whole machine and the design parameter data, and recording the correlation model as a second correlation model;
establishing a correlation model between the influence degree of the service life and the working condition parameters by combining the influence degrees of various working conditions on the service life and the design parameter data, and recording the correlation model as a third correlation model;
and establishing a residual service life prediction model after the working condition parameters are corrected by combining the two correlation models according to the health degree data of the whole machine, and predicting the residual service life under the actual working condition.
7. The multi-condition adaptive industrial robot health assessment method according to claim 1 or 6, further comprising the steps of:
an updating step: the state characteristic database and the working condition information database matched with the state characteristic database can be continuously updated according to the increase of the running state data and the working condition information data of the robot;
self-learning step: the first correlation model, the second correlation model and the third correlation model can be self-learned and optimized.
8. A multi-working condition self-adaptive industrial robot health degree evaluation system is characterized by comprising the following modules:
a data acquisition module: acquiring running state data, working condition information data and robot related design parameter data of each joint of the industrial robot;
a data processing module: processing the running state data to obtain a state characteristic database, and matching the state characteristic database with corresponding working condition information data to obtain a working condition information database;
the working condition self-adaptive processing module comprises: decomposing the characteristic data in the working condition information database into a plurality of mutually independent working condition parameter dimensions, establishing a correlation model of each characteristic in the characteristic database and a working condition parameter under different dimensions, recording the correlation model as a first correlation model, matching the selected characteristic data to the corresponding working condition parameter dimensions, compensating and eliminating the influence on the working condition under the working condition parameter dimensions according to the established first correlation model, and obtaining the characteristic data with desensitized working condition;
each joint health degree calculation module: standardizing the characteristic data desensitized to the working condition, and combining a distance method to obtain health degree indexes of all joints;
the whole machine health degree evaluation module: calculating the whole health degree of the robot according to the health degree index of each joint and a preset health degree threshold value of each joint, and performing overall evaluation on the health state of the robot according to the whole health degree of the robot;
a remaining service life prediction module: predicting the remaining service life of the robot according to the health degree of the whole robot, the working condition information data and the design parameter data;
updating the self-learning module: the state characteristic database and the working condition information database matched with the state characteristic database can be continuously updated according to the increase of the running state data and the working condition information data of the robot; the first correlation model, the second correlation model and the third correlation model can be self-learned and optimized.
CN201810045777.8A 2018-01-17 2018-01-17 Multi-working-condition self-adaptive industrial robot health degree assessment method and system Active CN108363836B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810045777.8A CN108363836B (en) 2018-01-17 2018-01-17 Multi-working-condition self-adaptive industrial robot health degree assessment method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810045777.8A CN108363836B (en) 2018-01-17 2018-01-17 Multi-working-condition self-adaptive industrial robot health degree assessment method and system

Publications (2)

Publication Number Publication Date
CN108363836A CN108363836A (en) 2018-08-03
CN108363836B true CN108363836B (en) 2021-07-20

Family

ID=63006363

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810045777.8A Active CN108363836B (en) 2018-01-17 2018-01-17 Multi-working-condition self-adaptive industrial robot health degree assessment method and system

Country Status (1)

Country Link
CN (1) CN108363836B (en)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109239082B (en) * 2018-09-21 2021-01-26 杭州安脉盛智能技术有限公司 Tobacco shred structure quality online detection method and system based on machine vision technology
CN109711453B (en) * 2018-12-21 2022-05-13 广东工业大学 Multivariable-based equipment dynamic health state evaluation method
CN111126822B (en) * 2019-12-19 2023-04-28 佛山科学技术学院 Industrial robot health assessment method, device and storage medium
CN112115575A (en) * 2020-07-29 2020-12-22 北京奔驰汽车有限公司 Equipment lubricating oil state evaluation system and method
WO2022041064A1 (en) * 2020-08-27 2022-03-03 Rethink Robotics Gmbh Method and apparatus for robot joint status monitoring
CN113032985B (en) * 2021-03-11 2024-04-26 北京必创科技股份有限公司 Intelligent evaluation method and device for service life of wireless sensing equipment
CN115139336B (en) * 2021-03-31 2024-07-16 中国科学院沈阳自动化研究所 Data acquisition and screening system and method for industrial robot health monitoring
CN114418042B (en) * 2021-12-30 2022-07-22 智昌科技集团股份有限公司 Industrial robot operation trend diagnosis method based on cluster analysis
CN116061233B (en) * 2023-02-06 2024-07-23 杭州亿恒科技有限公司 Health monitoring method of in-service industrial joint robot
CN117428760B (en) * 2023-10-10 2024-03-29 无锡蔚动智能科技有限公司 Joint module control system and method based on artificial intelligence
CN118365316B (en) * 2024-06-20 2024-09-03 深圳墨影科技有限公司 Predictive maintenance system and method for robot

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006021268A (en) * 2004-07-07 2006-01-26 Nissan Motor Co Ltd Service life time predicting method of robot, operation pattern making method of robot and their program
CN102313577A (en) * 2011-06-24 2012-01-11 上海大学 Equipment health state evaluation and recession prediction method based on multi-channel sensing signals
CN103558565B (en) * 2013-10-14 2016-06-01 中国科学院电工研究所 A kind of Three-leg type magnetic field detection robot
CN107016235B (en) * 2017-03-21 2020-06-19 西安交通大学 Equipment running state health degree evaluation method based on multi-feature adaptive fusion

Also Published As

Publication number Publication date
CN108363836A (en) 2018-08-03

Similar Documents

Publication Publication Date Title
CN108363836B (en) Multi-working-condition self-adaptive industrial robot health degree assessment method and system
CN108280543B (en) Working condition self-adaptive equipment health degree evaluation method based on classification regression mode
CN109459993B (en) Online adaptive fault monitoring and diagnosing method for process industrial process
EP1982301B1 (en) Method of condition monitoring
CN110530650B (en) Method for monitoring performance state of heavy-duty gas turbine based on generalized regression neural network and box diagram analysis
CN116383636A (en) Coal mill fault early warning method based on PCA and LSTM fusion algorithm
CN116028887B (en) Analysis method of continuous industrial production data
CN115270491A (en) Offshore wind power operation and maintenance platform design method based on multivariate information fusion
CN111638707A (en) Intermittent process fault monitoring method based on SOM clustering and MPCA
CN106682159A (en) Threshold configuration method
CN117498432A (en) New energy power system electric energy quality assessment method
CN111445105A (en) Power plant online performance diagnosis method and system based on target value analysis
CN114962390A (en) Hydraulic system fault diagnosis method and system and working machine
CN117422447A (en) Transformer maintenance strategy generation method, system, electronic equipment and storage medium
CN116258467B (en) Electric power construction management and control system
CN115456222A (en) Remote intelligent predictive maintenance operation and maintenance service method
CN108844662A (en) A kind of numerically-controlled machine tool electrical cabinet state evaluating method
CN113869502A (en) Deep neural network-based bolt tightening failure reason analysis method
CN113406537A (en) Quantitative evaluation method for fault degree of power equipment
Starr et al. Data fusion applications in intelligent condition monitoring
CN118430092B (en) Data general acquisition method based on MCC system
CN118037280B (en) Corrugated paper production line maintenance and fault diagnosis system
CN118313519B (en) Electromechanical full life cycle prediction modeling method and system
CN118396252B (en) MES client data analysis optimization method based on cloud computing
CN118171897B (en) Process editing method based on MES system

Legal Events

Date Code Title Description
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