A predictive monitoring and maintenance method for heavy-duty industrial robot joints
By constructing degradation expressions and optimizing maintenance cost prediction models, the maintenance standards for heavy-loaded industrial robot joints are determined, preventive maintenance is realized before failure, maintenance costs and failure efficiency are reduced, and maintenance standards are solved in the lack of maintenance standards in the existing technology. The problem.
Patent Information
- Application Number
- CN202411110794.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-14
AI Technical Summary
The lack of maintenance standards before joint failure of heavy-load industrial robots in the prior art leads to high maintenance costs when the system fails and the inability to effectively prevent system failure.
By constructing a degradation expression, the maximum degradation of the robot joint is obtained, and the maintenance degradation is set based on this amount. The maintenance cost prediction model is optimized using genetic algorithms to determine the setting time and degradation of the minimum maintenance cost, so as to achieve predictive monitoring and maintain.
Maintenance before the robot joint failure reduces failure efficiency and maintenance costs, avoids damage to other parts due to joint bearing failure, and only new bearings need to be replaced, reducing maintenance complexity.
Smart Images

Figure CN118656981B_ABST
Abstract
Description
Technical Field
[0001] The present invention generally relates to the technical field of industrial robot maintenance, and in particular to a predictive monitoring and maintenance method for heavy-load industrial robot joints. Background Art
[0002] In modern industrial automation, heavy-duty industrial robot joint servo motors are key joints that ensure efficient and precise operation of the robot. Their performance directly affects the stability and reliability of industrial robots. For mechanical systems with heavy production tasks such as industrial robots, sudden or unreasonable maintenance plans will disrupt the normal production rhythm.
[0003] For industrial robot servo motor bearings, when they fail, the motor function can be restored by replacing new bearings. For complex mechanical systems, system downtime will result in expensive downtime penalty costs.
[0004] In order to reduce the losses caused by system failure, preventive maintenance can be performed before the system joints fail, thereby reducing the maintenance cost of system failure. However, in the prior art, there is no specific standard or supporting solution for the degree to which the system needs maintenance before failure. Summary of the invention
[0005] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide a predictive monitoring and maintenance method for heavy-duty industrial robot joints.
[0006] The present invention provides a predictive monitoring and maintenance method for a heavy-duty industrial robot joint, comprising:
[0007] Obtain the maximum degradation amount of the robot joint under the condition of maintaining normal working;
[0008] The maintenance degradation amount and the first setting time for detecting the degradation amount of the robot joint are set according to the maximum degradation amount; the maintenance degradation amount indicates the minimum degradation amount required for maintenance of the robot joint; the maintenance degradation amount is less than the maximum degradation amount;
[0009] At the first set time, detecting the degradation amount of the robot joint to obtain a first actual degradation amount;
[0010] When the first actual degradation amount is greater than or equal to the maintenance degradation amount, the robot joint is maintained.
[0011] According to the technical solution provided by the present invention, the step of obtaining the maximum degradation amount of the robot joint under the condition of maintaining normal working state includes:
[0012] Constructing the functional relationship between the impact damage and the impact amplitude under the influence of temperature increase, and obtaining the first linear function;
[0013] The functional relationship between the impact damage and the impact amplitude under the influence of the current impact is constructed to obtain the second linear function;
[0014] Obtain the function of the natural degradation rate as the total degradation amount grows, and obtain the third linear function;
[0015] Constructing a total degradation expression of the robot joint according to the first linear function, the second linear function and the third linear function; the total degradation expression is used to represent the trend of the total degradation amount changing over time;
[0016] Obtain the total service time of the robot joints under normal working conditions;
[0017] The total service time is substituted into the total degradation amount expression to calculate the maximum degradation amount.
[0018] According to the technical solution provided by the present invention, the steps of setting the first setting time of the maintenance degradation amount and detecting the robot joint degradation amount according to the maximum degradation amount include:
[0019] Obtain a maintenance cost prediction model; the maintenance cost prediction model is used to express the relationship between the maintenance cost, the time of detecting the degradation amount of the robot joint and the degradation amount that needs to be maintained for the robot joint; wherein the time of detecting the degradation amount of the robot joint and the degradation amount that needs to be maintained for the robot joint are inputs, and the maintenance cost is output;
[0020] According to the maintenance cost prediction model, an optimization algorithm is used to optimize the time for detecting the degradation amount of the robot joints and the degradation amount that requires maintenance of the robot joints, so as to obtain the first set time and the maintenance degradation amount when the maintenance cost is at the minimum value.
[0021] According to the technical solution provided by the present invention, the optimization algorithm is a genetic algorithm;
[0022] According to the maintenance cost prediction model, the step of optimizing the time for detecting the degradation amount of the robot joint and the degradation amount for which the robot joint needs to be maintained by using a genetic algorithm to obtain the first set time and the maintenance degradation amount when the maintenance cost is at a minimum includes:
[0023] Set the search space, algorithm parameters, and termination conditions;
[0024] Setting a fitness function according to the maintenance cost prediction model;
[0025] In the iterative process, a set of iteration moments and iteration maintenance degradation amounts are calculated each time;
[0026] Substituting the iteration time and the iteration maintenance degradation amount into the fitness function to calculate the iteration fitness;
[0027] After the termination condition is reached, the iteration time and the iteration maintenance degradation amount corresponding to the minimum value of all the iteration fitnesses in the iteration process are used as the first set time and the maintenance degradation amount.
[0028] According to the technical solution provided by the present invention, the step of obtaining the maintenance cost prediction model includes:
[0029] Obtain the cost of detecting the degeneration amount of the robot joints to obtain the detection cost;
[0030] Obtain the cost of maintenance when the robot joints are not failed, and obtain the non-failure maintenance cost;
[0031] Obtain the cost of maintenance when the robot joint fails, and obtain the failure maintenance cost;
[0032] Obtain the loss caused by downtime per unit time and get the downtime cost;
[0033] Obtain the probability that the robot joint needs maintenance before failure, and obtain the maintenance probability before failure;
[0034] Obtain the probability that the robot joint needs maintenance when it fails, and obtain the failure maintenance probability;
[0035] Get the expected downtime required for robot maintenance;
[0036] The relationship between the detection cost, non-failure maintenance cost, failure maintenance cost, downtime cost, non-failure maintenance probability, failure maintenance probability, expected downtime time and the time for detecting the amount of robot joint degradation and the total maintenance cost is constructed to obtain the maintenance cost prediction model; the maintenance cost prediction model includes the unknown time for detecting the amount of robot joint degradation and the amount of degradation that requires maintenance of the robot joints.
[0037] According to the technical solution provided by the present invention, the step of obtaining the expected value of the downtime required for robot maintenance includes:
[0038] Obtain the reliability prediction model of robot joints;
[0039] Constructing a probability density function and a distribution function of the robot downtime according to the reliability prediction model;
[0040] The expected value of the downtime duration is calculated based on the probability density function and the distribution function.
[0041] According to the technical solution provided by the present invention, the step of obtaining the reliability prediction model of the robot joint includes:
[0042] Constructing a function of the probability that the total degradation amount of the robot joint is less than the maximum degradation amount over time, and obtaining a non-failure probability function;
[0043] The total degradation expression is substituted into the non-failure probability function to calculate and obtain the reliability prediction model.
[0044] According to the technical solution provided by the present invention, at the first set moment, the degradation amount of the robot joint is detected to obtain the first actual degradation amount, comprising:
[0045] At the first set time, detecting a vibration signal of a joint of the robot;
[0046] Obtaining a vibration amplitude according to the vibration signal;
[0047] A first actual degradation amount is calculated according to the vibration amplitude.
[0048] According to the technical solution provided by the present invention, when the first actual degradation amount is less than the maintenance degradation amount, the following steps are performed:
[0049] The failure time of the robot joint is calculated according to the initial time when the robot joint starts to serve and the total service time;
[0050] Selecting a second set time between the expiration time and the first set time;
[0051] At the second set time, the degradation amount of the robot joint is detected again to obtain a second actual degradation amount;
[0052] When the second actual degradation amount is greater than or equal to the maintenance degradation amount, maintenance is performed on the robot joint.
[0053] According to the technical solution provided by the present invention, when the second actual degradation amount is less than the maintenance degradation amount, the following steps are performed:
[0054] When the robot joint has been in service until the failure moment, the robot joint is maintained.
[0055] The beneficial effects of the present invention are:
[0056] The maintenance degradation amount is set according to the maximum degradation amount when the robot joint maintains normal operation, so that the maintenance degradation amount is less than the maximum degradation amount; and the degradation amount of the robot joint is detected at the first set time before the robot joint fails to obtain the first actual degradation amount. When the first actual degradation amount is greater than or equal to the maintenance degradation amount, the robot joint is maintained. In most cases, this method will maintain the robot joint before it fails, which significantly reduces the failure rate. At the same time, the maintenance degradation amount is set according to the maximum degradation amount, and the judgment standard for maintaining the robot joint is given. That is, after the degradation amount reaches the maintenance degradation amount, the robot joint is maintained before failure. Because no actual failure occurs, the other parts will not be additionally damaged due to the failure of the joint bearing; when performing maintenance, only new joint bearings need to be replaced, and other parts do not need to be replaced, which reduces maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:
[0058] Figure 1 A flowchart of a predictive monitoring and maintenance method for a heavy-load industrial robot joint is provided;
[0059] Figure 2 Schematic diagram of the degradation amount changing with time. DETAILED DESCRIPTION
[0060] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.
[0061] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0062] Please refer to Figure 1 The present invention provides a predictive monitoring and maintenance method for a heavy-duty industrial robot joint, comprising:
[0063] Obtain the maximum degradation amount of the robot joint under the condition of maintaining normal working;
[0064] The maintenance degradation amount and the first setting time for detecting the degradation amount of the robot joint are set according to the maximum degradation amount; the maintenance degradation amount indicates the minimum degradation amount required for maintenance of the robot joint; the maintenance degradation amount is less than the maximum degradation amount;
[0065] At the first set time, detecting the degradation amount of the robot joint to obtain a first actual degradation amount;
[0066] When the first actual degradation amount is greater than or equal to the maintenance degradation amount, the robot joint is maintained.
[0067] Specifically, the time when the robot joint starts to operate normally is taken as the initial time of service, and the time when the mechanical structure fails due to wear, impact and other factors is taken as the failure time. Since there are differences between the design life and actual service life of robot joints of the same model but different models, the failure time in this embodiment means the time when failure occurs when the design life is reached.
[0068] This means that if the robot joints are kept in service, working and operating continuously, the actual failure time may be earlier or later than the failure time corresponding to the design life.
[0069] Specifically, the maximum degradation amount indicates the total degradation amount of the robot joint just before failure.
[0070] The maintenance degradation amount is set to a value smaller than the maximum degradation amount.
[0071] The first set time is set between the initial time and the expiration time.
[0072] Further, when the first actual degradation amount is less than the maintenance degradation amount, the following steps are performed:
[0073] The failure time of the robot joint is calculated according to the initial time when the robot joint starts to serve and the total service time;
[0074] Selecting a second set time between the expiration time and the first set time;
[0075] At the second set time, the degradation amount of the robot joint is detected again to obtain a second actual degradation amount;
[0076] When the second actual degradation amount is greater than or equal to the maintenance degradation amount, maintenance is performed on the robot joint.
[0077] Among them, the total service time is theoretically equal to the design life; if the design life of the robot joints of this model is unknown, it can be obtained through experiments.
[0078] The second set time is set to an intermediate value between the first set time and the failure time.
[0079] Further, when the second actual degradation amount is less than the maintenance degradation amount, the following steps are performed:
[0080] When the robot joint has been in service until the failure moment, the robot joint is maintained.
[0081] In summary, the first detection of robot joints in this embodiment includes the following three situations:
[0082] Case 1: The actual degradation amount at the first set moment is less than the maintenance degradation amount. This means that the degradation degree of the robot joint is lower than expected, and the actual service life may be longer than the design life. It is necessary to further observe its degradation and then perform preventive maintenance.
[0083] Case 2: At the first set moment, the actual degradation amount is greater than or equal to the maintenance degradation amount, and less than the maximum degradation amount. This means that the actual degradation degree of the robot joint is basically in line with expectations, and preventive maintenance can be performed at this time to avoid failure.
[0084] Case 3: The actual degradation amount at the first set moment is greater than or equal to the maximum degradation amount. This indicates that the degradation degree of the robot joint is higher than expected. The actual service life must be less than the design life, and preventive maintenance is very necessary.
[0085] For situations 2 and 3, perform maintenance normally.
[0086] For case 1, refer to Figure 2 , which is a schematic diagram showing the change of degradation amount over time; the horizontal axis is time, and the vertical axis is degradation amount; the origin of the coordinates represents the initial moment when the robot joint begins to serve; the interval between the failure moment and the moment when the robot joint degradation amount is detected is It can be seen that the first set time is earlier than the failure time corresponding to the design life, and the first actual degradation amount detected is less than the maintenance degradation amount. It is necessary to further observe the degradation situation: after a period of time, the degradation amount of the robot joint is detected again, and the size of the maintenance degradation amount is judged again, including 3 situations:
[0087] Case 4: The actual degradation amount at the second set time is less than the maintenance degradation amount. This indicates that the degradation degree of the robot joint is still lower than expected, and the actual service life may be significantly longer than the design life. At this time, it is only necessary to perform preventive maintenance on it at the failure time corresponding to the design life. When the joint bearing is in service until the failure time, it is very likely that it will not actually fail. After maintenance, the reason why the actual service life is significantly longer than the design life can be analyzed.
[0088] Case 5: The actual degradation amount at the second set time is greater than or equal to the maintenance degradation amount and less than the maximum degradation amount. At this time, preventive maintenance is performed to avoid failure.
[0089] Case 6: The actual degradation amount at the second set time is greater than or equal to the maximum degradation amount. This means that between the first set time and the second set time, the robot joint was affected by additional factors and the degradation degree was higher than expected. This situation requires immediate preventive maintenance and further analysis of the reasons for rapid degradation in the subsequent process.
[0090] This method can maintain the robot joints before they fail in most cases, significantly reducing the failure rate.
[0091] At the same time, the maintenance degradation amount is set according to the maximum degradation amount, and the judgment standard for maintenance of robot joints is given.
[0092] If the robot joint is maintained before failure after the degradation reaches the maintenance degradation, the remaining parts will not be damaged due to the failure of the joint bearing because the failure has not actually occurred. When performing maintenance, only the joint bearing needs to be replaced without replacing other parts, which reduces maintenance costs.
[0093] Furthermore, the step of obtaining the maximum degradation amount of the robot joint under the condition of maintaining normal working state includes:
[0094] Constructing the functional relationship between the impact damage and the impact amplitude under the influence of temperature increase, and obtaining the first linear function;
[0095] The functional relationship between the impact damage and the impact amplitude under the influence of the current impact is constructed to obtain the second linear function;
[0096] Obtain the function of the natural degradation rate as the total degradation amount grows, and obtain the third linear function;
[0097] Constructing a total degradation expression of the robot joint according to the first linear function, the second linear function and the third linear function; the total degradation expression is used to represent the trend of the total degradation amount changing over time;
[0098] Obtain the total service time of the robot joints under normal working conditions;
[0099] The total service time is substituted into the total degradation amount expression to calculate the maximum degradation amount.
[0100] Specifically, since the degradation of spherical bearings is mainly caused by wear, the degradation amount caused by wear is X ( t ) can be considered to follow the general path degradation:
[0101] Formula 1;
[0102] in, represents the initial value of the degradation amount, βrepresents the degradation rate, both of which can be constants or random variables, t Indicates time, i represents the degradation parameter.
[0103] During the motor startup process, the current in the motor stator winding is several times that of normal operation, and a lot of heat will be generated in the winding. For industrial robot joint servo motors, the motor starts and stops frequently and periodically. In addition, the compact structure of the joint servo motor makes it difficult to dissipate heat, and the temperature accumulation leads to accelerated degradation of the bearings.
[0104] Since the arrival of the shock is periodic and not random; however, the amplitude of the shock is related to the heat and is random, and the amplitude of each shock should be similar, so the amplitude of this shock can be described using a normal distribution.
[0105] Also, because the units of measurement for the impact amplitude and the degradation are different, the first conversion factor is quoted a To establish the linear functional relationship between impact damage and impact amplitude, the impact damage distribution is: . Assuming that you have experienced n The sum of the impact degradation caused by the motor start-stop cycle is S 1( t ) is expressed as formula 2:
[0106] Formula 2:
[0107] in, a represents the first conversion coefficient, Indicates impact damage under the influence of temperature rise, T i Indicates i The moment of impact, It means that the mean is m W , the standard deviation is d W Normal distribution N 01 . Formula 2 is the first linear function.
[0108] Specifically, bearing current is one of the main causes of bearing failure, so in the study of the reliability or life of robot joint bearings, bearing current should be taken into consideration.
[0109] This embodiment designs relevant experiments and arranges corresponding experimental platforms. The principle is to connect a resistor in series with the positive electrode of the power supply and then connect it to the point where the electric cylinder on the aluminum foil applies force.
[0110] The bearing is insulated from the electric cylinder by the insulating layer. When the electric cylinder exerts force, the outer ring of the bearing is connected to the positive electrode of the power supply. The negative electrode of the power supply is connected to the shaft through a carbon brush, which is equivalent to the inner ring of the bearing being connected to the negative electrode of the power supply. At this time, a path is formed to flow through the bearing, and the voltage on the bearing is collected through an oscilloscope.
[0111] In addition, this test bench can use electric cylinders to load mechanical loads on the spherical bearings while applying electrical loads to them, so that the bearings can restore their working conditions in real operation to the greatest extent possible. The vibration signals of the spherical bearings are then collected through vibration sensors and transmitted to an oscilloscope.
[0112] This method only needs to use an oscilloscope to monitor the relationship between the vibration amplitude of the spherical bearing and the maintenance degradation amount, that is, to compare the root mean square of the vibration amplitude and the maintenance degradation amount.
[0113] Since the discharge events of spherical plain bearings are random, the number of bearing discharges should be measured within a sufficiently long observation period.
[0114] In this embodiment, 20 seconds is selected to estimate the number of discharges. The oscilloscope saves each image for 2ms, and a image is saved every 4 seconds. Therefore, the average number of discharges can be written as Yes m for:
[0115] Formula 3;
[0116] in, For the i The number of discharges counted in 0 images.
[0117] In this embodiment, the average number of discharges at ten moments was calculated, and it was found that the discharge events were more frequent at the initial start-up, then gradually decreased and gradually stabilized. Therefore, the exponential model can be used to approximate the discharge events. The exponential model has unknown parameters, which are obtained by fitting with the least squares method combined with the data collected in the previous text.
[0118] The voltage values recorded when the spherical plain bearings are broken down are distributed and statistically analyzed. At this time, only the magnitude of the bearing voltage is considered, and all spherical plain bearing voltages are regularized.
[0119] Specifically, the modeling of the spherical bearing current takes into account its damage characteristics. The high temperature each time the current breaks down will melt the bearing raceway and form micron-sized pits. When the number of impacts accumulates, special patterns will be formed, which will aggravate the wear of the bearing and accelerate degradation.
[0120] The current shock of the spherical bearing occurs when the motor starts and ends when the motor stops. The number of bearing current shocks in each cycle is completely random. Therefore, a non-homogeneous Poisson process can be used to deal with the bearing current shock, refer to formula 4:
[0121] Formula 4;
[0122] in, Indicates i cycle u The impact frequency of the moment, c 1 and c 2 represents the parameter of the current impulse frequency, which is obtained by fitting the experimental data; Indicates i Period ends at time t Number of impacts N i ( t )equal m i The probability of the next i With i The moment of impact T i In i consistent), u represents the moment, ! represents the factorial, e represents the natural logarithm, t i Indicates the start and stop of the motor i The starting time of a cycle, t i+1 Indicates the motor start and stop i +1 cycle start time, i Is a positive integer.
[0123] Regarding the amplitude of the current impact, we can refer to the bearing voltage every time the oil film is broken down. The impact amplitude is linearly related to the bearing voltage, that is, the greater the voltage, the greater the amplitude. The experimental results show that the bearing breakdown voltage follows a normal distribution, which is the same as the start-stop impact.
[0124] Introducing the second conversion factor b To establish the relationship between current impact damage and impact amplitude, the distribution is , indicating that the mean is , the standard deviation is Normal distribution N 02 . It is expressed by Formula 5:
[0125] Formula 5;
[0126] in, represents the sum of the current impact degradation, It represents the impact damage under the influence of current impact. Formula 5 is the second linear function; T j Indicatesj The moment of impact, j Is a positive integer.
[0127] As the degradation increases, the same impact will cause greater damage, so the natural degradation rate of the spherical bearing will accelerate after the impact. As shown in Formula 6:
[0128] Formula 6;
[0129] in, Y ( t ) is the total degradation amount, X ( t ) is the amount of degradation due to wear, f ( X ( t )) is the first state function, which represents the influence of natural degradation process on impact damage.
[0130] S ( t ) is the second state function, which represents the impact of the impact process on the natural degradation process; in this embodiment, it is equal to S 1( t )and S 2( t Formula 6 is the third linear function.
[0131] The first state function can be a linear increasing function, expressed as f ( X ( t ))= cX ( t )+ d .
[0132] in, c and d Both represent state function parameters, which are given based on experience.
[0133] According to the previous analysis, under the influence of temperature increase, the first linear functional relationship between impact damage and impact amplitude is as shown in Formula 2;
[0134] Under the influence of current shock, the second linear function relationship between shock damage and shock amplitude is as shown in Formula 5;
[0135] The third linear function of the natural degradation rate and the total degradation amount is as shown in Formula 6;
[0136] Combining the above three factors, the total degradation is Y ( t ) can be expressed as:
[0137] Formula seven.
[0138] in, Indicates time t The motor start and stop i The number of spherical plain bearing current impacts in a cycle; a represents the first conversion coefficient, b represents the second conversion coefficient;
[0139] T i , T j Respectively represent i The moment of the first impact and the j The moment of impact.
[0140] Further, the steps of setting the first setting time for maintaining the degradation amount and detecting the degradation amount of the robot joint according to the maximum degradation amount include:
[0141] Obtain a maintenance cost prediction model; the maintenance cost prediction model is used to express the relationship between the maintenance cost, the time of detecting the degradation amount of the robot joint and the degradation amount that needs to be maintained for the robot joint; wherein the time of detecting the degradation amount of the robot joint and the degradation amount that needs to be maintained for the robot joint are inputs, and the maintenance cost is output;
[0142] According to the maintenance cost prediction model, an optimization algorithm is used to optimize the time for detecting the degradation amount of the robot joints and the degradation amount that requires maintenance of the robot joints, so as to obtain the first set time and the maintenance degradation amount when the maintenance cost is at the minimum value.
[0143] Specifically, the steps of obtaining the maintenance cost prediction model include:
[0144] Obtain the cost of detecting the degeneration amount of the robot joints to obtain the detection cost;
[0145] Obtain the cost of maintenance when the robot joints are not failed, and obtain the non-failure maintenance cost;
[0146] Obtain the cost of maintenance when the robot joint fails, and obtain the failure maintenance cost;
[0147] Obtain the loss caused by downtime per unit time and get the downtime cost;
[0148] Obtain the probability that the robot joint needs maintenance before failure, and obtain the maintenance probability before failure;
[0149] Obtain the probability that the robot joint needs maintenance when it fails, and obtain the failure maintenance probability;
[0150] Obtaining an expected downtime duration required for robot maintenance; the expected downtime duration includes an unknown degradation amount that requires maintenance of the robot joints;
[0151] The relationship between the detection cost, non-failure maintenance cost, failure maintenance cost, downtime cost, non-failure maintenance probability, failure maintenance probability, expected downtime time and the time for detecting the amount of robot joint degradation and the total maintenance cost is constructed to obtain the maintenance cost prediction model; the maintenance cost prediction model includes the unknown time for detecting the amount of robot joint degradation and the amount of degradation that requires maintenance of the robot joints.
[0152] The detection cost is M d , the non-failure maintenance cost is M pk The failure maintenance cost is M ck The downtime cost is M s , the interval between the failure moment and the moment of detecting the robot joint degradation is , then the total maintenance cost M It can be expressed as:
[0153] Formula XIII;
[0154] in, l is the number of spherical plain bearings that require preventive maintenance, p is the number of spherical plain bearings that have failed during the test, P 01 ( B ) is l The probability function of the number of joints that need preventive maintenance, that is, the probability of no failure maintenance; P 02 ( C )for p The probability function of the number of joints that have failed and need replacement maintenance is the failure maintenance probability. E ( s ) is the expected downtime duration, k and q All are positive integers.
[0155] Furthermore, the step of obtaining the expected value of the downtime required for robot maintenance includes:
[0156] Obtain the reliability prediction model of robot joints;
[0157] Constructing a probability density function and a distribution function of the robot downtime according to the reliability prediction model;
[0158] The expected value of the downtime duration is calculated based on the probability density function and the distribution function.
[0159] Furthermore, the steps of obtaining the reliability prediction model of the robot joint include:
[0160] Construct a function of the probability that the total degradation amount of the robot joint is less than the maximum degradation amount over time, and obtain a non-failure probability function; that is, Formula 8;
[0161] The total degradation expression is substituted into the non-failure probability function to calculate and obtain the reliability prediction model.
[0162] Specifically, in order to calculate h The probability that a joint needs preventive maintenance is defined as a set {Sp}, where the elements in the set represent the failure states of multiple spherical bearings.
[0163] When the degradation amount of the spherical bearing is greater than or equal to the maintenance degradation amount, the corresponding element in the set is 1, otherwise it is 0. h joints, whose {Sp} has 2 h Case 2: {Sp}=[0,0,…,0],[1,0,…,0],[0,1,…,0],…,[1,1,…,1]. Then P 1( B ) in a certain situation in {Sp} can be expressed as:
[0164] Formula 14;
[0165] in, represents the probability of the number of joints requiring preventive maintenance in a situation {Sp}, S pk Indicates the first failure state of the kth spherical plain bearing, 0 indicates failure, and 1 indicates no failure. is the probability that the first actual degradation amount of the kth spherical bearing is greater than or equal to the maintenance degradation amount and less than the maximum degradation amount, h represents the number of joints, k is a positive integer; It means summing all the cases in the set {Sp}.
[0166] Similarly, a set {Sc} is defined to indicate that the joint is in a state of failure and requires replacement maintenance. P 02 ( C ) in a certain situation in {Sc} can be expressed as:
[0167] Formula 15;
[0168] in, represents the probability of the number of failed joints in a situation of {Sc}, is the probability that the first actual degradation amount of the kth spherical bearing is greater than or equal to the maximum degradation amount, S ck represents the second failure state of the kth spherical bearing, 0 represents failure, and 1 represents non-failure; It represents the sum of all cases in the set {Sc}.
[0169] Specifically, the expected downtime E ( s )for:
[0170] Formula 16;
[0171] in, f g ( t ) is the spherical plain bearing t The probability density function of the downtime at the time of failure is, F ( t ) is the spherical plain bearing t The distribution function of the downtime at the failure time.
[0172] Formula XVII;
[0173] Formula XVIII;
[0174] in, The specific value representing the reliability of all joint bearings of the robot, represents the reliability prediction model, k and h Compared with Formula 14 and Formula 15 k and h Consistent.
[0175] Specifically, the process of constructing the reliability prediction model is as follows: first, an initial expression as shown in Formula 8 is established to express t The degradation at the moment Y ( t ) is less than the maximum degradation D probability.
[0176]
[0177] Formula 8;
[0178] in, D represents the maximum degradation amount, P ( N i( t )= m i ) indicates time t The motor start and stop i The number of spherical bearing current impacts in a cycle is equal to m i Substituting formula 7 into formula 8, we get formula 9, which represents the reliability prediction model:
[0179]
[0180] Formula 9;
[0181] in, G i express , subject to the mean , the standard deviation is Normal distribution N 03 ; t i Indicates the motor start and stop i The starting time of a cycle, Indicates infinity.
[0182] is the distribution function of the degradation rate β, for The distribution function of for G i The distribution function of express of n Heavy Stieltjes convolution, express of n Heavy Stieltjes Convolution, here n The number of motor start-stop cycles n Consistent. z 1 means The independent variable, z 2 means The independent variable of .
[0183] Specifically, the first actual degradation amount of the spherical plain bearing is The probability of being less than the maintenance degradation amount is :
[0184]
[0185] Formula 10;
[0186] in, represents the first actual degradation amount, Indicates the actual amount of natural degradation.
[0187] The probability that the first actual degradation amount of the spherical plain bearing is greater than or equal to the maintenance degradation amount and less than the maximum degradation amount for:
[0188]
[0189] Formula XI;
[0190] in, t i-1 Indicates the motor start and stop i -The start time of a cycle, a 1 represents the ratio of the maintenance degradation amount to the maximum degradation amount.
[0191] The probability that the first actual degradation amount of the spherical plain bearing is greater than or equal to the maximum degradation amount for:
[0192]
[0193] Formula twelve.
[0194] Specifically, substitute Formula 10 to Formula 12 into Formula 9, and substitute Formula 9, Formula 14 to Formula 18 into Formula 13 to obtain the maintenance cost prediction model: , where the input is t , maintenance degradation a 1 D , the output is the maintenance cost:
[0195] Formula 19;
[0196] in, h Indicates the number of joints, which is set to 3 in this implementation.
[0197] Specifically, the optimization algorithm is a genetic algorithm;
[0198] According to the maintenance cost prediction model, the step of optimizing the time for detecting the degradation amount of the robot joint and the degradation amount for which the robot joint needs to be maintained by using a genetic algorithm to obtain the first set time and the maintenance degradation amount when the maintenance cost is at a minimum includes:
[0199] Set the search space, algorithm parameters, and termination conditions;
[0200] Setting a fitness function according to the maintenance cost prediction model;
[0201] In the iterative process, a set of iteration moments and iteration maintenance degradation amounts are calculated each time;
[0202] Substituting the iteration time and the iteration maintenance degradation amount into the fitness function to calculate the iteration fitness;
[0203] After the termination condition is reached, the iteration time and the iteration maintenance degradation amount corresponding to the minimum value of all the iteration fitnesses in the iteration process are used as the first set time and the maintenance degradation amount.
[0204] Specifically, Formula 19 is set as the fitness function,
[0205] Since the failure time is known, and the maintenance degradation can be expressed as a multiple of the maximum degradation, that is, a 1 times the maximum degradation D Therefore, in the actual optimization process, the parameters that need to be optimized are t and a 1.
[0206] The search space is set to t In the interval (0, 600), unit: minute, a 1 is in the interval (0, 1); the initial population size is set to 30, and each population contains t and a 1 Two chromosomes. Crossover and mutation operations are performed during the optimization process, with a crossover probability of 0.6 and a mutation probability of 0.2; the maximum number of iterations is 100, which serves as the termination condition.
[0207] By using the globally optimal iteration time and the iteration maintenance degradation amount as the first set time and the maintenance degradation amount, the optimization result of the algorithm can be closer to the real optimal solution.
[0208] Furthermore, at the first set moment, the degradation amount of the robot joint is detected to obtain a first actual degradation amount, comprising:
[0209] At the first set time, detecting a vibration signal of a joint of the robot;
[0210] Obtaining the vibration amplitude according to the vibration signal; this belongs to the prior art and will not be described in detail here;
[0211] A first actual degradation amount is calculated according to the vibration amplitude.
[0212] Specifically, the root mean square of the vibration amplitude is taken as the first actual degradation amount.
[0213] The above description is only a preferred embodiment of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present invention (but not limited to) to form a technical solution.
Claims
1. A predictive monitoring and maintenance method for heavy-duty industrial robot joints, characterized in that: include: Obtain the maximum degradation amount of the robot joint under the condition of maintaining normal working; The maintenance degradation amount and the first setting time for detecting the degradation amount of the robot joint are set according to the maximum degradation amount; the maintenance degradation amount indicates the minimum degradation amount required for maintenance of the robot joint; the maintenance degradation amount is less than the maximum degradation amount; At the first set time, detecting the degradation amount of the robot joint to obtain a first actual degradation amount; When the first actual degradation amount is greater than or equal to the maintenance degradation amount, performing maintenance on the robot joint; The steps of obtaining the maximum degradation amount of the robot joint under the condition of maintaining normal working state include: Constructing the functional relationship between the impact damage and the impact amplitude under the influence of temperature increase, and obtaining the first linear function; The functional relationship between the impact damage and the impact amplitude under the influence of the current impact is constructed to obtain the second linear function; Obtain the function of the natural degradation rate as the total degradation amount grows, and obtain the third linear function; Constructing a total degradation expression of the robot joint according to the first linear function, the second linear function and the third linear function; the total degradation expression is used to represent the trend of the total degradation amount changing over time; Obtain the total service time of the robot joints under normal working conditions; The total service time is substituted into the total degradation amount expression to calculate the maximum degradation amount.
2. The predictive monitoring and maintenance method for a heavy-duty industrial robot joint according to claim 1 is characterized in that: The steps of setting the maintenance degradation amount and detecting the first setting time of the robot joint degradation amount according to the maximum degradation amount include: Obtain a maintenance cost prediction model; the maintenance cost prediction model is used to express the relationship between the maintenance cost, the time of detecting the degradation amount of the robot joint and the degradation amount that needs to be maintained for the robot joint; wherein the time of detecting the degradation amount of the robot joint and the degradation amount that needs to be maintained for the robot joint are inputs, and the maintenance cost is output; According to the maintenance cost prediction model, an optimization algorithm is used to optimize the time for detecting the degradation amount of the robot joints and the degradation amount that requires maintenance of the robot joints, so as to obtain the first set time and the maintenance degradation amount when the maintenance cost is at the minimum value.
3. The predictive monitoring and maintenance method for a heavy-duty industrial robot joint according to claim 2 is characterized in that: The optimization algorithm is a genetic algorithm; According to the maintenance cost prediction model, the step of optimizing the time for detecting the degradation amount of the robot joint and the degradation amount for which the robot joint needs to be maintained by using a genetic algorithm to obtain the first set time and the maintenance degradation amount when the maintenance cost is at a minimum includes: Set the search space, algorithm parameters, and termination conditions; Setting a fitness function according to the maintenance cost prediction model; In the iterative process, a set of iteration moments and iteration maintenance degradation amounts are calculated each time; Substituting the iteration time and the iteration maintenance degradation amount into the fitness function to calculate the iteration fitness; After the termination condition is reached, the iteration time and the iteration maintenance degradation amount corresponding to the minimum value of all the iteration fitnesses in the iteration process are used as the first set time and the maintenance degradation amount.
4. The predictive monitoring and maintenance method for a heavy-duty industrial robot joint according to claim 2, characterized in that: The steps to obtain the maintenance cost prediction model include: Obtain the cost of detecting the degeneration amount of the robot joints to obtain the detection cost; Obtain the cost of maintenance when the robot joints are not failed, and obtain the non-failure maintenance cost; Obtain the cost of maintenance when the robot joint fails, and obtain the failure maintenance cost; Obtain the loss caused by downtime per unit time and get the downtime cost; Obtain the probability that the robot joint needs maintenance before failure, and obtain the maintenance probability before failure; Obtain the probability that the robot joint needs maintenance when it fails, and obtain the failure maintenance probability; Get the expected downtime required for robot maintenance; The relationship between the detection cost, non-failure maintenance cost, failure maintenance cost, downtime cost, non-failure maintenance probability, failure maintenance probability, expected downtime time and the time for detecting the amount of robot joint degradation and the total maintenance cost is constructed to obtain the maintenance cost prediction model; the maintenance cost prediction model includes the unknown time for detecting the amount of robot joint degradation and the amount of degradation that requires maintenance of the robot joints.
5. The predictive monitoring and maintenance method for a heavy-duty industrial robot joint according to claim 4, characterized in that: The steps to derive the expected downtime required for robot maintenance include: Obtain the reliability prediction model of robot joints; Constructing a probability density function and a distribution function of the robot downtime according to the reliability prediction model; The expected value of the downtime duration is calculated based on the probability density function and the distribution function.
6. A predictive monitoring and maintenance method for a heavy-duty industrial robot joint according to claim 5, characterized in that: The steps to obtain the reliability prediction model of the robot joints include: Constructing a function of the probability that the total degradation amount of the robot joint is less than the maximum degradation amount over time, and obtaining a non-failure probability function; The total degradation expression is substituted into the non-failure probability function to calculate and obtain the reliability prediction model.
7. The predictive monitoring and maintenance method for heavy-duty industrial robot joints according to claim 1 is characterized in that: At the first set time, the step of detecting the degradation amount of the robot joint to obtain the first actual degradation amount includes: At the first set time, detecting a vibration signal of a joint of the robot; Obtaining a vibration amplitude according to the vibration signal; A first actual degradation amount is calculated according to the vibration amplitude.
8. The predictive monitoring and maintenance method for a heavy-duty industrial robot joint according to claim 1, characterized in that: When the first actual degradation amount is less than the maintenance degradation amount, the following steps are performed: The failure time of the robot joint is calculated according to the initial time when the robot joint starts to serve and the total service time; Selecting a second set time between the expiration time and the first set time; At the second set time, the degradation amount of the robot joint is detected again to obtain a second actual degradation amount; When the second actual degradation amount is greater than or equal to the maintenance degradation amount, maintenance is performed on the robot joint.
9. The predictive monitoring and maintenance method for heavy-duty industrial robot joints according to claim 8, characterized in that: When the second actual degradation amount is less than the maintenance degradation amount, the following steps are performed: When the robot joint has been in service until the failure moment, the robot joint is maintained.
