Maritime work equipment fatigue life prediction method based on operation and maintenance big data

Abstract

The invention discloses a maritime work equipment fatigue life prediction method based on operation and maintenance big data, and the method comprises the steps: selecting a fault phenomenon as a key feature value through the historical operation and maintenance big data of maritime work equipment, calculating the fault rate of the fault phenomenon, determining a key node through the fault rate, and carrying out the prediction of the fatigue life of the maritime work equipment. Then calculating an overall dynamic adjustment factor of the maritime work equipment through the key node, establishing a three-dimensional simulation model of the maritime work equipment for fatigue life analysis, and then establishing a maritime work equipment fatigue life prediction model based on a BP neural network for fatigue life prediction, so as to obtain a damage value and a scrapping evaluation coefficient of the key node; and performing state evaluation on the maritime work equipment. By processing the operation and maintenance data and determining the reasonable operation and maintenance period of the key node by using the neural network algorithm, the normal operation of the maritime work equipment is ensured, the maintenance frequency of the equipment is reduced, and the normal use time of the equipment is prolonged.

Description

Technical Field

[0001] The present invention relates to the field of marine engineering equipment, and in particular to a method for predicting fatigue life of marine engineering equipment based on operation and maintenance big data. Background Art

[0002] During the construction of marine engineering projects, some marine equipment is often used, such as cutting equipment, welding equipment, transportation equipment, etc. Marine equipment will inevitably be damaged during long-term use, and regular maintenance is required to ensure its normal operation. However, after long-term use, some marine equipment often fails during use. For example, transportation equipment often fails during transportation, causing related construction materials to fail to arrive at the site on time, resulting in a waste of construction time and increased labor time.

[0003] At the same time, during long-term operation, offshore equipment will produce fatigue damage, which can easily cause damage to key structures. In order to improve the reliability and safety of engineering equipment, it is necessary to determine the fatigue stress of the equipment and avoid the problem of damage to the mechanical structure. This requires the prediction of the fatigue life of offshore equipment. At present, most of the fatigue life prediction methods use the finite element simulation method. This method needs to set certain assumptions during the prediction, resulting in inaccurate fatigue life prediction and affecting the accuracy of fatigue life prediction. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a method for predicting the fatigue life of offshore engineering equipment based on operation and maintenance big data. This method reduces the maintenance frequency of the equipment and increases the normal use time of the equipment by predicting the fatigue life of each key node in the operation of the offshore engineering equipment.

[0005] The main steps of the fatigue life prediction method of offshore equipment based on operation and maintenance big data of the present invention are as follows:

[0006] Step 1: Collect historical operation and maintenance big data of offshore equipment, classify and statistically process the historical operation and maintenance data, and obtain the historical fault information set of offshore equipment, including historical data of the equipment, fault phenomenon, operation and maintenance plan, operation time and other information. Select the fault phenomenon as the key feature value and calculate the fault rate s of the fault phenomenon. i ;

[0007] Step 2: Failure rate s i The value range is divided into intervals, and the causes of the fault phenomena in the operation and maintenance solutions corresponding to the fault phenomena with a failure rate greater than or equal to 0.1 are determined as key nodes. i <0.1, indicating that the probability of the fault phenomenon is low and can be ignored, 0.1≤s i ​<1, indicating that the probability of the fault phenomenon occurring is high and it is determined to be a critical node;

[0008] Step 3: Based on the key nodes determined in step 2, determine the maximum value of the normal operation time T of each key node based on historical operation and maintenance data. max and minimum value T min , calculate the dynamic adjustment factor DA of each key node i ;

[0009] Step 4: Dynamic adjustment factor DA of each key node determined in step 3 i , according to the failure rate s i Set the failure rate weight coefficient α of each key node i , calculate the overall dynamic adjustment factor DA of offshore equipment 总：

[0010] Step 5: Establish a 3D model of the offshore equipment in the finite element analysis software, and determine the overall dynamic adjustment factor DA of the offshore equipment according to step 4. 总 Conduct static analysis to obtain the stress distribution of the equipment during operation;

[0011] Step 6: Import the stress distribution state obtained by static analysis into the finite element analysis software to perform modal analysis on the offshore equipment. The first two natural frequencies ω1 and ω2 of the offshore equipment are obtained through modal analysis. The mass damping coefficient α and the stiffness damping coefficient β are obtained according to the first two natural frequencies ω1 and ω2.

[0012] Step 7: Using the mass damping coefficient α and stiffness damping coefficient β obtained in step 6 as constraint conditions, perform transient dynamic analysis on the offshore equipment, set the time to 100s, and the time step to 0.5s, solve the stress-time curve of the key nodes of the offshore equipment with potential safety hazards, and extract the stress values ​​{σ1, σ2…σ j}, according to the stress values ​​of each key node {σ1, σ2…σ j}Carry out fatigue life simulation analysis of offshore equipment and obtain fatigue life curves of each key node σ i -N i , i=1,2,3…j;

[0013] Step 8: Set a vibration signal collector at each key node during the operation of the offshore equipment to obtain the vibration signals {x1, x2…x i}, normalize the collected vibration signal to remove the influence of noise;

[0014] Step 9: Construct a fatigue life prediction model for offshore equipment based on BP neural network, input the normalized vibration signal into the BP neural network fatigue life prediction model, use the BP neural network algorithm to predict fatigue life, and obtain the predicted stress cycle number n of the key nodes of offshore equipment. i ;

[0015] Step 10: According to the mineral fatigue damage theory, the predicted stress cycle number n i The number of stress cycles N of the simulated fatigue life i , and obtain the damage value D of each key node i If the damage value D i If the damage value D is greater than or equal to 0.8, it indicates that the key node needs maintenance. i If it is less than 0.8, it means that the node is in normal use and can be used continuously;

[0016] Step 11: Calculate the normal operation time T of replacing key node components of marine engineering equipment through operation and maintenance data t , and the total operating time T of the marine engineering equipment is obtained 总 , and obtain the scrap assessment coefficient θ of the marine engineering equipment;

[0017] Step 12: Determine the status of the marine engineering equipment according to its scrapping assessment coefficient θ. If the scrapping assessment coefficient θ is greater than or equal to 0.9, it indicates that the equipment is in an aging state and it is recommended to scrap it. If the scrapping assessment coefficient θ is less than 0.9, it indicates that the equipment is in normal use and can be used continuously.

[0018] The beneficial effects of the present invention are as follows: This method determines the key nodes of operation and maintenance of offshore engineering equipment by utilizing a large amount of data in the operation and maintenance process of offshore engineering equipment, analyzes the structural stress and fatigue life of the key nodes of offshore engineering equipment, and uses a big data processing method to predict the fatigue life of offshore engineering equipment using a neural network algorithm at the key nodes, thereby determining a reasonable operation and maintenance cycle of the key nodes, ensuring the normal operation of offshore engineering equipment, reducing the maintenance frequency of equipment, and increasing the normal use time of equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a method for predicting fatigue life of offshore equipment based on operation and maintenance big data according to the present invention;

[0020] Figure 2 It is a schematic diagram of the structure of the fatigue life prediction model of marine equipment based on BP neural network of the present invention;

[0021] Figure 3 This is a flow chart of fatigue life prediction of marine engineering equipment based on BP neural network of the present invention. DETAILED DESCRIPTION

[0022] The method of the present invention will be further described below in conjunction with the accompanying drawings.

[0023] As shown in the accompanying drawings, a method for predicting fatigue life of offshore equipment based on operation and maintenance big data of the present invention comprises the following steps:

[0024] Step 1: Collect historical operation and maintenance big data of offshore equipment, classify and statistically process the historical operation and maintenance data, and obtain the historical fault information set of offshore equipment, including historical data of the equipment, fault phenomenon, operation and maintenance plan, operation time and other information. Select the fault phenomenon as the key feature value and calculate the fault rate s of the fault phenomenon. i ,

[0025]

[0026] In the formula, s i is the failure rate of a certain fault phenomenon, N is the statistical number of historical operation and maintenance big data, n i The number of statistical operations and maintenance for a certain fault phenomenon.

[0027] Step 2: Failure rate s i Divide the value range into intervals, and identify the causes of the fault phenomena in the operation and maintenance solutions corresponding to the fault phenomena with a failure rate greater than or equal to 0.1 as key nodes;

[0028] 0 i <0.1, indicating that the probability of the fault phenomenon is low and can be ignored;

[0029] 0.1≤s i <1, indicating that the probability of the fault phenomenon occurring is high and it is determined to be a critical node;

[0030] Step 3: Based on the key nodes determined in step 2, determine the maximum value of the normal operation time T of each key node based on historical operation and maintenance data. max and minimum value T min , calculate the dynamic adjustment factor DA of each key node i ,

[0031]

[0032] In the formula, DA i is the dynamic adjustment factor of each key node, T imax is the maximum value of normal operation time in historical operation and maintenance data, T imin The maximum value of normal operation time in historical operation and maintenance data, T i T is the effective time that a key node has been operating normally. iref is the reference running time. ​

[0033] Step 4: Dynamic adjustment factor DA of each key node determined in step 3 i , according to the failure rate s i Set the failure rate weight coefficient α of each key node i , calculate the overall dynamic adjustment factor DA of offshore equipment 总：

[0034]

[0035] In the formula, DA 总 is the overall dynamic adjustment factor of offshore equipment, DA i is the dynamic adjustment factor of each key node, α i is the failure rate weight coefficient, i is the i-th key node, and j is the number of key nodes.

[0036] Step 5: Establish a 3D model of the offshore equipment in the finite element analysis software, and determine the overall dynamic adjustment factor DA of the offshore equipment according to step 4. 总 Conduct static analysis to obtain the stress distribution of the equipment during operation;

[0037] Step 6: Import the stress distribution state obtained by static analysis into the finite element analysis software to perform modal analysis on the offshore equipment. The first two natural frequencies ω1 and ω2 of the offshore equipment are obtained through modal analysis. The mass damping coefficient α and the stiffness damping coefficient β are obtained by solving the first two natural frequencies ω1 and ω2. The calculation formula is as follows:

[0038]

[0039] Where: ε1, ε2 are damping ratios, which are set to 0.8 according to the structural parameters of offshore equipment.

[0040] Step 7: Using the mass damping coefficient α and stiffness damping coefficient β obtained in step 6 as constraint conditions, perform transient dynamic analysis on the offshore equipment, set the time to 100s, and the time step to 0.5s, solve the stress-time curve of the key nodes of the offshore equipment with potential safety hazards, and extract the stress values ​​{σ1, σ2…σ j}, according to the stress values ​​of each key node {σ1, σ2…σ j}Carry out fatigue life simulation analysis of offshore equipment and obtain fatigue life curves of each key node σ i -N i , i=1,2,3…j;

[0041] Step 8: Set a vibration signal collector at each key node during the operation of the offshore equipment to obtain the vibration signals {x1, x2…x i}, the collected vibration signal is normalized to remove the influence of noise. The function used for normalization is:`

[0042]

[0043] In the formula, x i is the collected value of the vibration signal of the i-th key node, x imax is the maximum value of the collected value of the vibration signal of the i-th key node, x imin is the minimum value of the collected values ​​of the vibration signal of the i-th key node.

[0044] Step 9: Construct a fatigue life prediction model for offshore equipment based on BP neural network, input the normalized vibration signal into the BP neural network fatigue life prediction model, use the BP neural network algorithm to predict fatigue life, and obtain the predicted stress cycle number n of the key nodes of offshore equipment. i， The specific steps are as follows:

[0045] The first step is to build a fatigue life prediction model for offshore equipment based on BP neural network and establish the corresponding relationship between the stress value of each key node and the fatigue life;

[0046]

[0047] Where N′ i is the predicted fatigue life, σ i is the stress value of the key node, p i is the stress value weight coefficient of each key node, b is the bias value, i is the i-th key node, and j is the number of key nodes.

[0048] The second step is to convert the stress values ​​of each key node {σ1, σ2…σ j} is input into the fatigue life prediction model of offshore equipment based on BP neural network, the fatigue life prediction model of offshore equipment based on BP neural network is trained, and the predicted fatigue life curve σ of each key node obtained by prediction output is obtained i -N′ i ;

[0049] The third step is to calculate the predicted fatigue life N′ i The fatigue life N obtained by analysis i For comparison, If the error Δ>10%, return to the first step to adjust the stress value weight coefficient p of each key node. iand bias value b, retrain the fatigue life prediction model of offshore equipment using the BP neural network, and retrain the predicted fatigue life N′ i The fatigue life N obtained by analysis i The comparison is carried out until Δ≤10%, indicating that the fatigue life prediction model of offshore equipment based on BP neural network meets the prediction requirements;

[0050] The fourth step is to convert the normalized vibration signals of each key node into the stress / strain response of each key node through the frequency domain method, and obtain the stress power spectrum density G (f i );

[0051] Step 5: Substitute the stress power spectrum density G(f i ) is input into the BP neural network fatigue life prediction model, and the BP neural network algorithm is used to predict the fatigue life to obtain the predicted stress cycle number n at each key node of the offshore equipment. i .

[0052] Step 10: According to the mineral fatigue damage theory, the predicted stress cycle number n i The number of stress cycles N of the simulated fatigue life i , and obtain the damage value D of each key node i If the damage value D i If the damage value D is greater than or equal to 0.8, it indicates that the key node needs maintenance. i If it is less than 0.8, it indicates that the node is in normal use and can be used continuously.

[0053]

[0054] Where D i is the fatigue damage value, n i is the predicted stress cycle number of the i-th key node, N i is the number of simulated fatigue life stress cycles of the i-th key node, i is the i-th key node, and j is the number of key nodes.

[0055] Step 11: Calculate the normal operation time T of replacing key node components of marine engineering equipment through operation and maintenance data t , and the total operating time T of the marine engineering equipment is obtained 总 , we get the scrapping assessment coefficient θ of marine engineering equipment,

[0056]

[0057]

[0058] Where θ is the scrap assessment coefficient of marine engineering equipment, T 总 is the total operating time of the marine engineering equipment, Y m is the normal operation time of the key node components when they are replaced, T0 is the total operation time specified by the equipment before leaving the factory, x is the tth replacement value, and t is the total number of replacements.

[0059] Step 12: Determine the status of the marine engineering equipment according to its scrapping assessment coefficient θ. If the scrapping assessment coefficient θ is greater than or equal to 0.8, it indicates that the equipment is in an aging state and it is recommended to scrap it. If the scrapping assessment coefficient θ is less than 0.8, it indicates that the equipment is in normal use and can be used continuously.

Claims

1. A method for predicting fatigue life of offshore equipment based on operation and maintenance big data, characterized in that The following steps are involved: Step 1: Collect historical operation and maintenance big data of offshore equipment, classify and statistically process the historical operation and maintenance data, and obtain the historical fault information set of offshore equipment, including historical data of the equipment, fault phenomenon, operation and maintenance plan, operation time and other information. Select the fault phenomenon as the key feature value and calculate the fault rate s of the fault phenomenon. i ; Step 2: Failure rate s i The value range is divided into intervals, and the causes of the fault phenomena in the operation and maintenance solutions corresponding to the fault phenomena with a failure rate greater than or equal to 0.1 are determined as key nodes. i <0.1, indicating that the probability of the fault phenomenon is low and can be ignored, 0.1≤s i <1, indicating that the probability of the fault phenomenon occurring is high and it is determined to be a critical node;​ Step 3: Based on the key nodes determined in step 2, determine the maximum value of the normal operation time T of each key node based on historical operation and maintenance data. max and minimum value T min , calculate the dynamic adjustment factor DA of each key node i ; Step 4: Dynamic adjustment factor DA of each key node determined in step 3 i , according to the failure rate s i Set the failure rate weight coefficient α of each key node i , calculate the overall dynamic adjustment factor DA of offshore equipment 总 : Step 5: Establish a 3D model of the offshore equipment in the finite element analysis software, and determine the overall dynamic adjustment factor DA of the offshore equipment according to step 4. 总 Conduct static analysis to obtain the stress distribution of the equipment during operation; Step 6: Import the stress distribution state obtained by static analysis into the finite element analysis software to perform modal analysis on the offshore equipment. The first two natural frequencies ω1 and ω2 of the offshore equipment are obtained through modal analysis. The mass damping coefficient α and the stiffness damping coefficient β are obtained according to the first two natural frequencies ω1 and ω2. Step 7: Using the mass damping coefficient α and stiffness damping coefficient β obtained in step 6 as constraint conditions, perform transient dynamic analysis on the offshore equipment, set the time to 100s, and the time step to 0.5s, solve the stress-time curve of the key nodes of the offshore equipment with potential safety hazards, and extract the stress values ​​{σ1, σ2…σ j }, according to the stress values ​​of each key node {σ1, σ2…σ j } Conduct fatigue life simulation analysis of offshore equipment and obtain fatigue life curves of key nodes σ i -N i , i=1,2,3…j; Step 8: Set a vibration signal collector at each key node during the operation of the offshore equipment to obtain the vibration signals {x1, x2…x i }, normalize the collected vibration signal to remove the influence of noise; Step 9: Construct a fatigue life prediction model for offshore equipment based on BP neural network, input the normalized vibration signal into the BP neural network fatigue life prediction model, use the BP neural network algorithm to predict fatigue life, and obtain the predicted stress cycle number n of the key nodes of offshore equipment. i ; Step 10: According to the mineral fatigue damage theory, the predicted stress cycle number n i The number of stress cycles N of the simulated fatigue life i , and obtain the damage value D of each key node i If the damage value D i If it is greater than or equal to 0.8, it indicates that the key node needs maintenance; If the damage value D i If it is less than 0.8, it means that the node is in normal use and can be used continuously; Step 11: Calculate the normal operation time T of replacing key node components of marine engineering equipment through operation and maintenance data t , and the total operating time T of the marine engineering equipment is obtained 总 , and obtain the scrap assessment coefficient θ of the marine engineering equipment; Step 12: Determine the status of the marine engineering equipment according to its scrapping assessment coefficient θ. If the scrapping assessment coefficient θ is greater than or equal to 0.9, it indicates that the equipment is in an aging state and it is recommended to scrap it. If the scrapping assessment coefficient θ is less than 0.9, it indicates that the equipment is in normal use and can be used continuously.

2. The method for predicting fatigue life of offshore equipment based on operation and maintenance big data according to claim 1 is characterized in that: Step 9: The specific steps of using BP neural network algorithm to predict fatigue life are as follows: The first step is to build a fatigue life prediction model for offshore equipment based on BP neural network and establish the corresponding relationship between the stress value of each key node and the fatigue life; Where N' i is the predicted fatigue life, σ i is the stress value of the key node, p i is the stress value weight coefficient of each key node, b is the bias value, i is the i-th key node, and j is the number of key nodes. The second step is to convert the stress values ​​of each key node {σ1, σ2…σ j } is input into the fatigue life prediction model of offshore equipment based on BP neural network, the fatigue life prediction model of offshore equipment based on BP neural network is trained, and the predicted fatigue life curve σ of each key node obtained by prediction output is obtained i -N' i ; The third step is to calculate the predicted fatigue life N' i The fatigue life N obtained by analysis i For comparison, If the error Δ>10%, return to the first step to adjust the stress value weight coefficient p of each key node. i and bias value b, retrain the fatigue life prediction model of offshore equipment using the BP neural network, and retrain the predicted fatigue life N' i The fatigue life N obtained by analysis i The comparison is carried out until Δ≤10%, indicating that the fatigue life prediction model of offshore equipment based on BP neural network meets the prediction requirements; The fourth step is to convert the normalized vibration signals of each key node into the stress / strain response of each key node through the frequency domain method, and obtain the stress power spectrum density G (f i ); Step 5: Substitute the stress power spectrum density G(f i ) is input into the BP neural network fatigue life prediction model, and the BP neural network algorithm is used to predict the fatigue life to obtain the predicted stress cycle number n at each key node of the offshore equipment. i .

Images