Online monitoring of wind turbine blade root bolt looseness and fatigue life prediction method

Abstract

The present invention discloses a method for online monitoring of the looseness of a wind turbine blade root bolt group and predicting fatigue life, which belongs to the field of wind power and includes the following steps: determining the swing direction of the blade of the monitoring unit; arranging strain gauges; collecting strain data; setting early warning and alarm rules; establishing a K value curve for monitoring the bolt group; detecting the most unfavorable bolt axial force; calculating the average axial force of the most unfavorable bolt; calculating the tension increment in the screw under the fatigue load of the most unfavorable bolt; calculating the equivalent stress amplitude and average stress of the most unfavorable bolt; and determining the remaining fatigue life of the most unfavorable bolt. The present invention combines the bolt loosening area with the fatigue life prediction of the most unfavorable bolt in the area, which not only improves the accuracy of the blade root bolt life prediction, but also can realize the early detection and early treatment of the blade root bolt loosening problem through the accurate identification of the looseness of the local bolt group, and eliminate the possibility of fatigue fracture.

Description

Technical Field

[0001] The invention relates to the field of wind power, and in particular to a method for online monitoring of the looseness of a wind turbine blade root bolt group and predicting its fatigue life. Background Art

[0002] The blade root bolts of wind turbines are subjected to repeated loads for a long time, which can easily cause significant loss of bolt pre-tension, leading to loosening of the blade root bolts (i.e., the actual bolt pre-tension P0 is less than the designed pre-tension P 0S ) and fatigue fracture occurs, and the blade bolt fracture mostly occurs in the blade swing direction (0° and 180° directions).

[0003] The bolt connection of wind turbine blade root is different from that of traditional steel structure. The bolts of blade root and wind turbine head hub flange connection (hereinafter referred to as flange bolts) are under repeated load for a long time, which can easily cause the actual pre-tension P0 loss of the bolts, resulting in fatigue stress amplitude Δσ of the connection bolts under external force. s Significant increase, such as Figure 1 As shown, if not solved in time, it will lead to premature fatigue fracture of the bolt; Figure 1 It shows the relationship between the bolt and the stress amplitude under different pre-tension states. The three curves represent the actual axial forces of the three bolts (these three axial forces are our own assumptions). P0 means that the bolt has no prestress loss (that is, the bolt axial force when the wind turbine is just built, and the bolt pre-tightening force is applied in place), 0.8 P0 means that the bolt has less prestress loss (that is, the bolt axial force after the wind turbine has been running for a period of time or the bolt itself is not tightened, only 0.8 P0), and 0.5 P0 means that the bolt has a large prestress loss (that is, the wind turbine has been running for a long time, and the bolt prestress loss is large under fatigue load or the bolt itself is not tightened, only 0.5 P0). When there is no external load (such as no wind load or the wind turbine is in a shutdown state), under the same bending moment amplitude ΔM, the fatigue stress amplitude Δσ of the three bolts s The increase is very small, the fatigue stress amplitude Δσ s Small, long fatigue life (the fatigue life of the three bolts is almost the same at this time), such as Figure 1 As shown in the left half. With the increase of external load, under the same bending moment amplitude ΔM, the fatigue stress amplitude Δσ of the three bolts s There will be significant differences, such as Figure 1 Shown in the right half.

[0004] At present, bolt loosening monitoring is mainly aimed at detecting the looseness of a single bolt. The method can only perform quantitative detection of bolt loosening, and cannot evaluate the looseness degree and predict fatigue life. Bolt and bolt loosening angle real-time monitoring system and method, wind turbine connection bolt loosening monitoring system are all aimed at a single bolt. When monitoring the looseness of the bolt, it is necessary to arrange the sensor on each bolt, which will inevitably cause the problem of installing a large number of monitoring sensors and high costs.

[0005] The finite element method is currently mainly used to predict the life of bolts. The finite element model not only needs to consider factors such as the structural nonlinearity, material nonlinearity and contact nonlinearity of the bolts, but also needs to consider the problem of uneven stiffness around the bolts. The actual operation is relatively cumbersome; the patent document with publication number CN109000873B discloses a method for predicting the loosening life of bolts, and the patent document with publication number CN117113710A discloses a method for predicting the fatigue life of wind turbine bolts based on intelligent monitoring. These two patent documents do not combine the bolt pre-tension loss and bolt life prediction in actual projects, and have little significance in actual projects; the patent document with publication number CN114491854A discloses a method for rapid prediction of bolt life based on local stress-strain theory and finite element simulation. It performs finite element analysis and simulation on each bolt, which is not only time-consuming and labor-intensive, but also cannot meet the needs of real-time monitoring in actual projects. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a method for online monitoring of the looseness of a wind turbine blade root bolt group and predicting fatigue life with a simple algorithm and high prediction accuracy.

[0007] The technical solution of the present invention to solve the above technical problem is: a method for online monitoring of the looseness of a wind turbine blade root bolt group and fatigue life prediction, comprising the following steps:

[0008] S1: Determine the blade swing direction of the monitored blade root bolt;

[0009] S2: strain gauges are arranged in the blade swing direction and away from the blade root end, with the end away from the blade root end as the upper monitoring point and the blade swing direction as the lower monitoring point;

[0010] S3: When the wind turbine starts to run, the strain data of the upper monitoring point and the lower monitoring point are collected respectively;

[0011] S4: The strain ratio K value curves of the upper monitoring point and the lower monitoring point after the wind turbine has been in operation for one year are used as the standard curve, and the strain ratio K value curves of the upper monitoring point and the lower monitoring point after the wind turbine has been in operation for one year are used as the comparison curve;

[0012] S5: Compare the standard curve and the comparison curve, set the finite element calculated K value curves of equivalent loosening of 1 bolt and equivalent loosening of 3 bolts as the early warning line and the alarm line, respectively, and set the early warning and alarm rules;

[0013] S6: Establish a K value curve for monitoring bolt groups to evaluate the loosening and deterioration trend of the monitoring bolt groups;

[0014] S7: Determine the bolt loosening area based on the K value curve of the monitoring bolt group, and use ultrasonic measurement technology to detect the most unfavorable bolt axial force. The most unfavorable bolt is the bolt with the largest degree of loosening;

[0015] S8: The main engine manufacturer obtains the fatigue load of the most unfavorable bolt among the blade root bolts under monitoring and calculates the average axial force of the most unfavorable bolt;

[0016] S9: Calculate the tensile force increment in the screw under the most unfavorable fatigue load of the bolt;

[0017] S10: Calculate the equivalent stress amplitude and average stress of the most unfavorable bolt;

[0018] S11: Substitute the calculated average stress and equivalent stress amplitude into the bolt fatigue curve bundle for calculation to determine the remaining fatigue life of the most unfavorable bolt;

[0019] S12: Based on the looseness monitoring results of blade root bolts and the remaining fatigue life prediction results of the most unfavorable bolts, submit an online monitoring report on the looseness of wind turbine blade root bolts and fatigue life prediction to the owner, and put forward treatment suggestions for possible problems.

[0020] In the above-mentioned method for online monitoring of the looseness of the wind turbine blade root bolt group and predicting fatigue life, in step S1, in actual engineering, the blade bolt fracture occurs in the blade swing direction, that is, 0° and 180° directions, so the blade root bolt monitoring position is clear, and it is sufficient to monitor the blade swing direction.

[0021] In the above-mentioned method for online monitoring of the looseness of the root bolt group of a wind turbine and prediction of fatigue life, in step S2, a staggered arrangement scheme of upper and lower monitoring points is adopted, and the lower monitoring points are respectively set at the positions of 0° and 180° in the swing direction of the blade. and , the next monitoring point Set up monitoring points on both sides and , the next monitoring point Set the upper monitoring point above and away from the blade root end , the next monitoring point Set up monitoring points on both sides and , the next monitoring point Set the upper monitoring point above and away from the blade root end , where and are configured as the main measurement group, and and and and and and are configured as the slave measurement group. The lower monitoring points are arranged in a denser pattern according to needs, while only one upper monitoring point needs to be arranged.

[0022] In the above method for online monitoring of the loosening degree and fatigue life prediction of the wind turbine blade root bolt group, in step S2, according to the Saint-Venant boundary conditions, there are distance requirements for the positions of the upper and lower monitoring points. In actual engineering, the distance from the upper monitoring point to the flange connection is taken to be ≥ 0.7d, where d is the tower diameter, to ensure that the strain at the upper monitoring point only reflects the magnitude of the load and completely eliminates the influence of the flange boundary conditions.

[0023] In the above method for online monitoring of the loosening degree and fatigue life prediction of the wind turbine blade root bolt group, in step S5, the warning and alarm rules are as follows:

[0024] If K ≤ the warning line, it is judged to be safe, and the countermeasure is not to take any measures;

[0025] If the warning line < K ≤ the alarm line, it is judged as a warning, and the countermeasure is to check whether there is any single bolt loosening and continuously monitor;

[0026] If K > the alarm line, it is judged as an alarm, and the countermeasure is to immediately conduct bolt fatigue life prediction.

[0027] In the above method for online monitoring of the loosening degree and fatigue life prediction of the wind turbine blade root bolt group, the specific process of step S6 is as follows:

[0028] There are two states of the loosened bolt under the action of external force. Different from the normal bolt, there is a situation where the actual pre-tension P0 of the bolt is less than the designed pre-tension. As a result, under the action of external force, the pressure on the connection surface drops to 0, causing the connection surface to disengage, resulting in a change in the stress state and a decrease in the flange connection stiffness. Therefore, there are two stress states of the bolt under the action of external tensile force N: normal stress and loosened stress;

[0029] Normal stress state:

[0030] The normal stress state is that the external tensile force N is less than the pre-pressure C of the flange connection surface, and the stress of the bolt satisfies the following equilibrium equation:

[0031] (1)

[0032] Where: ΔP is the tension increment in the screw; ΔC is the reduction in flange contact pressure; P is the bolt axial force; at this time, most of the external force overcomes the preload, and only a small part of ΔP is applied to the bolt:

[0033] (2)

[0034] Where: k b and k c are the vertical stiffness of screw and flange respectively;

[0035] Loose stress state:

[0036] As the external tension N gradually increases, the preload C of the flange connection surface will be completely lost, and the connection surface will be disengaged. At this time, the external tension N is all borne by the bolts, and ΔP is:

[0037] (3)

[0038] Due to the change in the force transmission mode, the N and P curves will overlap in the loose state, and the flange connection stiffness k3 in the loose state will be reduced, which is smaller than the flange connection stiffness k1 in the normal stress state, but larger than the slope k2 of the P curve in the normal stress state. The relationship between the three is:

[0039] k2<k3<k1(4)

[0040] According to Saint-Venant's principle, compared with the stress in the flange connection area, the vertical stress in the blade area far away from the flange connection area will not be affected by the degree of connection looseness;

[0041] To this end, the loosening influence coefficient of the flange connection bolt group is introduced, that is, the strain ratio K is:

[0042] K = ε2 / ε1 (5)

[0043] Where ε1 is the strain at the upper monitoring point, and ε2 is the strain at the lower monitoring point;

[0044] The detachment point A and height h of the strain ratio K value curve of the upper and lower monitoring points are determined by the finite element calculation results of the connecting flange bolt group and the cylinder wall.

[0045] In the above-mentioned method for online monitoring of the looseness of the blade root bolt group of a wind turbine and predicting fatigue life, in step S8, the minimum value N of the fatigue load on the most unfavorable bolt among the blade root bolts being monitored is obtained from the host manufacturer. min , the maximum value of fatigue load on the most unfavorable bolt N max , and the average axial force P of the most unfavorable bolt is obtained mean :

[0046] (6).

[0047] In the above-mentioned method for online monitoring of the looseness of the root bolt group of wind turbines and prediction of fatigue life, in step S9, for the running wind turbine blades, the actual pre-tension force P0 of the bolts, the minimum value N of the fatigue load on the most unfavorable bolts min and the maximum value of fatigue load on the most unfavorable bolt N max , due to the change in the stress state of the loose bolt, the actual fatigue stress amplitude changes, which can be divided into the following three situations:

[0048] Case 1: N max ≤P0;

[0049] When P0 has no loss, the pre-stress on the flange connection interface exists, the interface is not detached, the bolt is in normal working condition, ΔP meets the fatigue strength requirements, and the remaining fatigue life meets the use requirements;

[0050] ΔP=P2-P1(7)

[0051] ΔP≈0 (8)

[0052] Where P2 is N max ≤P0, at N max The actual axial force of the most unfavorable bolt under action; P1 is N max ≤P0, at N min The actual axial force of the most unfavorable bolt under action;

[0053] The second case: N min ≤P0<N max ;

[0054] Assume that the critical bolt axial force for interface disengagement is Ps. If N max >P0≥Ps, the interface is disengaged and the bolt axial force changes completely with the external load; when N min When <P0<Ps, the pre-pressure of the flange connection interface exists, the interface is not loose, and the bolt is in normal working condition; so when N min ≤P0<N max When ΔP is calculated as follows:

[0055] ΔP=N max -P3 (9)

[0056] Where P3 is N min ≤P0<N max When N min The actual axial force of the most unfavorable bolt under action;

[0057] The third case: N min >P0;

[0058] If N appears min>P0, the bolt pretension is completely lost, the interface is completely detached, and the bolt axial force is completely controlled by the external load, then:

[0059] ΔP=N max -N min (10)

[0060] It can be seen that the fatigue life prediction of the bolt is related to P0 and N max 、N min When P0 has no loss, N max ≤P0, ΔP≈0, the pre-pressure on the flange connection surface always exists, and the remaining life meets the use requirements; when N min ≤P0<N max , ΔP=N max -P3, perform remaining life prediction to determine whether it meets the use requirements; when N min >P0,ΔP=N max -N min , the pre-stress on the flange connection surface is completely lost, ΔP As the external force changes in amplitude, the remaining life is predicted to determine whether it meets the use requirements.

[0061] In the above-mentioned method for online monitoring of the looseness of the wind turbine blade root bolt group and fatigue life prediction, in step S10, the equivalent stress amplitude of the most unfavorable bolt is calculated according to the following formula: and mean stress ;

[0062] (11)

[0063] (12)

[0064] In the formula, A e is the effective area of ​​the bolt.

[0065] The beneficial effects of the present invention are:

[0066] 1. Compared with most loosening monitoring technologies that require sensors to be installed on individual bolts, the innovation of the present invention is to achieve online monitoring of the looseness of a local bolt group, that is, based on the Saint-Venant principle, by attaching strain gauges to the bolts and the ends away from the bolts, collecting strain data at different measuring points to draw a K curve of the loosening influence coefficient of the connecting bolt group, and by determining the breakaway point A and height h of the K value curve, the looseness of the local bolt group can be accurately identified. The present invention has the characteristics of low cost, simple operation and high identification accuracy.

[0067] 2. The present invention combines the bolt loosening area with the fatigue life prediction of the most unfavorable bolt in the area, that is, the ultrasonic bolt detection technology is used to determine the residual pretension of the most unfavorable bolt (with the smallest pretension) in the loose bolt group, and the stress amplitude of the most unfavorable bolt under the action of external force considering the loss of pretension is calculated. Combined with the S-N curve of the bolt material, the remaining life prediction of the most unfavorable bolt can be achieved. Different from the current bolt life prediction method, the present invention combines the bolt loosening area with the fatigue life prediction of the most unfavorable bolt in the area, which not only improves the accuracy of the blade root bolt life prediction, but also can achieve early detection and early treatment of the blade root bolt loosening problem through accurate identification of the degree of looseness of the local bolt group, eliminate the possibility of fatigue fracture, and can maximize the recovery of economic losses for enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 This is a diagram showing the relationship between the bolt and the stress amplitude under different pre-tension states.

[0069] Figure 2 It is a flow chart of the present invention.

[0070] Figure 3 This is the plan view of the strain arrangement for monitoring the looseness of flange bolt groups.

[0071] Figure 4 This is an elevation view of the strain arrangement scheme for monitoring the loosening of flange bolt groups.

[0072] Figure 5 This is a schematic diagram of flange bolt group loosening monitoring and early warning.

[0073] Figure 6 Schematic diagram of two stress states of the bolt, where (a) represents the normal stress state and (b) represents the loose state.

[0074] Figure 7 Schematic diagram of the relationship between N, P and C under different stress states.

[0075] Figure 8 Schematic diagram of vertical strain distribution of tower wall at different heights of flange connection, where (c) indicates that the bolts are not loose, and (d) indicates that the three bolts are completely loose.

[0076] Fig. 9 It is the K value curve.

[0077] Fig.10 N max When ≤P0, the relationship curve between bolt pretension and external load.

[0078] Fig.11 N min <P0<N max The relationship curve between bolt pretension and external load is shown in Figure 2.

[0079] Fig.12 N min >P0, relationship curve between bolt pretension and external load.

[0080] Fig.13 Schematic diagram of fatigue curve bundle of bolts considering different degrees of looseness. DETAILED DESCRIPTION

[0081] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0082] like Figure 2 As shown, a method for online monitoring of the looseness of a wind turbine blade root bolt group and predicting fatigue life includes the following steps:

[0083] S1: Determine the blade swing direction of the monitored blade root bolt.

[0084] In the step S1, in actual engineering, the blade bolt fracture occurs in the blade swing direction, that is, 0° and 180° directions, so the blade root bolt monitoring position is clear, and only the blade swing direction needs to be monitored.

[0085] S2: strain gauges are arranged in the blade swing direction and away from the blade root end, with the end away from the blade root end as the upper monitoring point and the blade swing direction as the lower monitoring point.

[0086] In step S2, the problem of reduced monitoring and evaluation accuracy due to the large location of the monitoring points and the loose area is solved by encrypting the monitoring points. At the same time, in order to reduce the number of monitoring points as much as possible, the staggered arrangement of the upper and lower monitoring points is adopted. Figure 3 and Figure 4 In the figure, 1 is the flange, 2 is the bolt, 3 is the bolt loosening area, and 4 is the tower wall. The lower monitoring points are set at 0° and 180° in the blade swing direction. and , the next monitoring point Set up monitoring points on both sides and , the next monitoring point Set the upper monitoring point above and away from the blade root end , the next monitoring point Set up monitoring points on both sides and , the next monitoring point Set the upper monitoring point above and away from the blade root end ,in , The main test group, , and , and , and are configured as the lower monitoring group. The lower monitoring points are arranged in a denser pattern according to needs, while only one upper monitoring point is arranged. In this way, while minimizing the number of monitoring points as much as possible, the monitoring cost is reduced, and the requirements for monitoring accuracy can be ensured.

[0087] According to the Saint-Venant boundary conditions, there are distance requirements for the positions of the upper and lower monitoring points. In actual projects, the distance from the upper monitoring point to the flange connection is taken to be ≥0.7d, where d is the tower diameter, to ensure that the strain at the upper monitoring point only reflects the magnitude of the load and is completely free from the influence of the flange boundary conditions, thus ensuring the accuracy of the monitoring method.

[0088] S3: When the wind turbine starts operating, collect the strain data of the upper and lower monitoring points respectively.

[0089] S4: Take the strain ratio K value curves of the upper and lower monitoring points during one year of the wind turbine operation as the standard curve, and take the strain ratio K value curves of the upper and lower monitoring points during the subsequent one-year operation of the wind turbine as the comparison curve.

[0090] S5: As shown in Figure 5 , compare the standard curve and the comparison curve, and set the finite element calculation K value curves of equivalent loosening of 1 bolt and equivalent loosening of 3 bolts as the warning line and the alarm line respectively. In actual projects, bolt group loosening occurs, that is: bolts around the most unfavorable bolt are loosened to varying degrees, and the loosening degree is smaller the farther away from the most unfavorable bolt. Therefore, the equivalent hypothesis of bolt loosening effect is proposed. The effect of complete loosening of 1 bolt is equivalent to the different loss degrees of pre-tension of 3 bolts, that is, equivalent loosening of 1 bolt, and the effect of complete loosening of 3 bolts is equivalent to the different loss degrees of pre-tension of 5 bolts, that is, equivalent loosening of 3 bolts; set the warning and alarm rules. Figure 5 In , M1 is the safe area, M2 is the warning area, and M3 is the alarm area.

[0091] Follow the warning and alarm rules listed in Table 1 during work.

[0092]

[0093] The warning and alarm rules are as follows:

[0094] If K ≤ the warning line, it is judged as safe, and the corresponding measure is not to take any measures;

[0095] If K > the alarm line, it is judged as an alarm, and the corresponding measure is to immediately conduct bolt fatigue life prediction;

[0096] If the warning line < K ≤ the alarm line, it is judged as a warning, and the corresponding measure is to check whether there is serious loosening of a single bolt and continue to pay attention.

[0097] S6: Establish a K value curve for monitoring bolt groups and evaluate the loosening and deterioration trend of the monitoring bolt groups.

[0098] The specific process of step S6 is as follows:

[0099] There are two states of loose bolts under external force. Different from normal bolts, the actual pre-tension P0 of the bolt is less than the designed pre-tension. As a result, under the action of external force, the pressure on the connection surface drops to 0, causing the connection surface to be disengaged, resulting in a change in the stress state and a decrease in the stiffness of the flange connection. Therefore, the bolts are divided into two stress states: normal stress and loose stress under the action of external tension N. Figure 6 shown.

[0100] Normal stress state:

[0101] The normal stress state is that the external tension N is less than the preload C of the flange connection surface. Figure 6 As shown in (a), the bolt force satisfies the following equilibrium equation:

[0102] (1)

[0103] Where: ΔP is the tension increment in the screw; ΔC is the reduction in flange contact pressure; P is the bolt axial force; at this time, most of the external force overcomes the preload, and only a small part of ΔP is applied to the bolt:

[0104] (2)

[0105] Where: k b and k c are the vertical stiffness of screw and flange respectively.

[0106] Loose stress state:

[0107] like Figure 6 As shown in (b), as the external tension N gradually increases, the preload C of the flange connection surface will be completely lost, and the connection surface will be disengaged. At this time, the external tension N is all borne by the bolts, and ΔP is:

[0108] (3)

[0109] Due to the change in the force transmission mode, the N and P curves will overlap in the loose state, and the flange connection stiffness k3 in the loose state will be reduced, which is smaller than the flange connection stiffness k1 in the normal stress state, but larger than the slope k2 of the P curve in the normal stress state. The relationship between the three is:

[0110] k2<k3<k1(4)

[0111] According to Saint-Venant's principle, compared with the stress in the flange connection area, the vertical stress in the blade area far away from the flange connection area will not be affected by the degree of connection looseness, such as Figure 8As shown in (c); if three bolts of the wind turbine tower flange connection bolts are partially loose, Figure 8 As shown in (d), the cylinder wall strain around the flange will be affected, and the value and distribution will be different from when it is not loose.

[0112] To this end, the loosening influence coefficient of the flange connection bolt group is introduced, that is, the strain ratio K is:

[0113] K = ε2 / ε1 (5)

[0114] Where ε1 is the strain at the upper monitoring point, and ε2 is the strain at the lower monitoring point.

[0115] like Fig. 9 As shown in the figure, the finite element calculation results of the connecting flange bolt group and the cylinder wall can determine the detachment point A and height h of the strain ratio K value curve of the upper and lower monitoring points. The earlier point A appears, the greater the looseness of the bolt group. The size of h can not only reflect the looseness of the bolt group, but also reflect the severity of the detachment of the lower flange connection surface. Therefore, in the actual process, the detachment point A and height h of the K value curve are determined to achieve an accurate assessment of the looseness of the monitoring bolt group.

[0116] S7: Determine the bolt loosening area based on the K value curve of the monitoring bolt group, and use ultrasonic measurement technology to detect the most unfavorable bolt axial force. The most unfavorable bolt is the bolt with the greatest degree of loosening.

[0117] S8: The main engine manufacturer obtains the fatigue load of the most unfavorable bolt in the blade root bolts of the monitored blades and calculates the average axial force of the most unfavorable bolt.

[0118] Obtain the minimum value N of fatigue load on the most unfavorable bolt among the blade root bolts under monitoring from the host manufacturer min , the maximum value of fatigue load on the most unfavorable bolt N max , and the average axial force P of the most unfavorable bolt is obtained mean :

[0119] (6).

[0120] S9: Calculate the increment of tension in the screw under the most unfavorable fatigue load on the bolt.

[0121] For the operating wind turbine blades, the actual bolt pre-tension P0, the minimum value of the fatigue load on the most unfavorable bolt N min and the maximum value of fatigue load on the most unfavorable bolt N max , due to the change in the stress state of the loose bolt, the actual fatigue stress amplitude changes, which can be divided into the following three situations:

[0122] Case 1: N max ≤P0;

[0123] When P0 has no loss, the flange connection interface pre-pressure exists, the interface is not detached, and the bolts are in normal working condition, such as Fig.10 As shown, ΔP meets the fatigue strength requirements and the remaining fatigue life meets the service requirements;

[0124] ΔP=P2-P1(7)

[0125] ΔP≈0 (8)

[0126] Where P2 is N max ≤P0, at N max The actual axial force of the most unfavorable bolt under action; P1 is N max ≤P0, at N min The actual axial force of the most unfavorable bolt under action;

[0127] The second case: N min ≤P0<N max ;

[0128] like Fig.11 As shown in the figure, the critical bolt axial force for interface disengagement is assumed to be Ps. If N max >P0≥Ps, the interface is disengaged and the bolt axial force changes completely with the external load; when N min When <P0<Ps, the pre-pressure of the flange connection interface exists, the interface is not loose, and the bolt is in normal working condition; so when N min ≤P0<N max When ΔP is calculated as follows:

[0129] ΔP=N max -P3 (9)

[0130] Where P3 is N min ≤P0<N max When N min The actual axial force of the most unfavorable bolt under action;

[0131] The third case: N min >P0;

[0132] like Fig.12 As shown, if N min >P0, the bolt pretension is completely lost, the interface is completely detached, and the bolt axial force is completely controlled by the external load, then:

[0133] ΔP=N max -N min (10)

[0134] It can be seen that the fatigue life prediction of the bolt is related to P0 and N max 、N min When P0 has no loss, Nmax ≤P0, ΔP≈0, the pre-pressure on the flange connection surface always exists, and the remaining life meets the use requirements; when N min ≤P0<N max , ΔP=N max -P3, perform remaining life prediction to determine whether it meets the use requirements; when N min >P0,ΔP=N max -N min , the pre-stress on the flange connection surface is completely lost, ΔP As the external force changes in amplitude, the remaining life is predicted to determine whether it meets the use requirements.

[0135] S10: Calculate the equivalent stress amplitude and average stress of the most unfavorable bolt.

[0136] In step S10, the equivalent stress amplitude of the most unfavorable bolt is calculated as follows: and mean stress ;

[0137] (11)

[0138] (12)

[0139] Where: A e is the effective area of ​​the bolt.

[0140] S11: Substitute the calculated average stress and equivalent stress amplitude into the bolt fatigue curve bundle for calculation to determine the remaining fatigue life of the most unfavorable bolt.

[0141] like Fig.13 As shown in the figure, for fatigue problems of rolled or non-welded steel components such as bolts, their fatigue life is not only related to Δσ and fatigue times n, but also to σ mean related, Fig.13 Medium Δ σ 1 and n 1 are the equivalent stress amplitude and fatigue times in the first case, Δ σ 2 and n 2 are the equivalent stress amplitude and fatigue times in the second case, Δ σ 3 and n 3 are the equivalent stress amplitude and fatigue times in the third case. Under the same stress amplitude, σ mean The larger it is, the lower its fatigue curve is, the smaller the number of fatigue times is, and the shorter the fatigue life is.

[0142] However, for loose bolts, due to the decrease of P0, the bolts may partially or completely enter the second stage of stress under fatigue load (such as Figure 6As shown in Figure 2), Δσ increases significantly, although the σ of loosened bolts mean It is relatively small, but Δσ increases a lot, resulting in a significant decrease in fatigue life compared to that of non-loose bolts. This is also the main reason why broken bolts are often completely loosened bolts in actual engineering.

[0143] S12: Based on the looseness monitoring results of blade root bolts and the fatigue life prediction results of the most unfavorable bolts, submit an online monitoring report on the looseness of wind turbine blade root bolts and fatigue life prediction to the owner, and put forward treatment suggestions for possible problems.

Claims

1. A method for online monitoring of the looseness of a wind turbine blade root bolt group and predicting fatigue life, characterized in that: It includes the following steps: S1: Determine the blade flapping direction of the monitored blade root bolts; S2: Arrange strain gauges in the blade flapping direction and away from the blade root end. Take the position away from the blade root end as the upper monitoring point and the blade flapping direction as the lower monitoring point; S3: Collect the strain data of the upper monitoring point and the lower monitoring point respectively when the wind turbine starts to operate; S4: Take the strain ratio K value curve of the upper monitoring point and the lower monitoring point of the wind turbine operating for 1 year as the standard curve, and take the strain ratio K value curve of the upper monitoring point and the lower monitoring point of the wind turbine operating for one year later as the comparison curve; S5: Compare the standard curve and the comparison curve. Set the finite element calculation K value curves of equivalent loosening of 1 bolt and equivalent loosening of 3 bolts as the warning line and the alarm line respectively, and set the warning and alarm rules; S6: Establish the K value curve of the monitored bolt group and evaluate the loosening deterioration trend of the monitored bolt group; S7: According to the K value curve of the monitored bolt group, determine the bolt loosening area, and use ultrasonic measurement technology to detect the most unfavorable bolt axial force. The most unfavorable bolt is the bolt with the largest loosening degree; S8: The main engine manufacturer obtains the fatigue load borne by the most unfavorable bolt among the monitored blade root bolts of the blade and calculates the average axial force of the most unfavorable bolt; S9: Calculate the tensile force increment in the screw under the action of the fatigue load borne by the most unfavorable bolt; S10: Calculate the equivalent stress amplitude and average stress borne by the most unfavorable bolt; S11: Substitute the calculated average stress and equivalent stress amplitude into the bolt fatigue curve bundle for calculation to determine the remaining fatigue life of the most unfavorable bolt; S12: Based on the comprehensive monitoring of the loosening degree of the blade root bolt group and the prediction results of the remaining fatigue life of the most unfavorable bolt, submit an online monitoring and fatigue life prediction report on the loosening degree of the wind turbine blade root bolt group to the owner, and put forward treatment suggestions for possible problems.

2. The method for online monitoring of looseness of wind turbine blade root bolts and fatigue life prediction according to claim 1 is characterized in that: In step S1, in actual engineering, blade bolt fractures occur in the blade flapping direction, that is, the 0° and 180° directions. Therefore, the monitoring position of the blade root bolts is clear, and it is only necessary to monitor the blade flapping direction.

3. The method for online monitoring of looseness of wind turbine blade root bolts and fatigue life prediction according to claim 1 is characterized in that: In step S2, the staggered arrangement of the upper and lower monitoring points is adopted, and the lower monitoring points are respectively set at the positions of 0° and 180° in the blade swing direction. and , the next monitoring point Set up monitoring points on both sides and , the next monitoring point Set the upper monitoring point above and away from the blade root end , the next monitoring point Set up monitoring points on both sides and , the next monitoring point Set the upper monitoring point above and away from the blade root end ,in , The main test group, , and , and , and It is composed of a monitoring group, in which the lower monitoring points are arranged more densely as needed, while only one upper monitoring point is required.

4. The method for online monitoring of looseness of wind turbine blade root bolt groups and fatigue life prediction according to claim 3 is characterized in that: In step S2, according to the Saint-Venant boundary conditions, there are distance requirements for the positions of the upper monitoring point and the lower monitoring point. In actual engineering, the distance from the upper monitoring point to the flange connection is taken as ≥0.7d, where d is the tower diameter, to ensure that the strain at the upper monitoring point only reflects the load magnitude and completely eliminates the influence of the flange boundary conditions.

5. The method for online monitoring of looseness of wind turbine blade root bolts and fatigue life prediction according to claim 1 is characterized in that: In step S5, the warning and alarm rules are as follows: If K ≤ warning line, it is judged as safe, and the countermeasure is not to take measures; If warning line < K ≤ alarm line, it is judged as a warning, and the countermeasure is to check whether there is a single bolt loosening and continue to pay attention; If K > alarm line, it is judged as an alarm, and the countermeasure is to immediately conduct bolt fatigue life prediction.

6. The method for online monitoring of looseness of wind turbine blade root bolt groups and fatigue life prediction according to claim 1 is characterized in that: The specific process of step S6 is as follows: Under the action of external forces, there are two states of loosened bolts. Different from normal bolts, there is a situation where the actual pre-tension P0 of the bolt is less than the designed pre-tension. As a result, under the action of external forces, the connection surface pressure drops to 0, resulting in the disconnection of the connection surface, which causes the force state to change and the flange connection stiffness to decrease. Therefore, there are two force states of normal force and loosening for the bolt under the action of external tensile force N; Normal force state: The normal stress state is that the external tension N is less than the preload C of the flange connection surface, and the bolt force satisfies the following equilibrium equation: （1） Where: ΔP is the tension increment in the screw; ΔC is the reduction in flange contact pressure; P is the bolt axial force; at this time, most of the external force overcomes the preload, and only a small part of ΔP is applied to the bolt: （2） Where: k b and k c are the vertical stiffness of screw and flange respectively; Loose stress state: As the external tension N gradually increases, the preload C of the flange connection surface will be completely lost, and the connection surface will be disengaged. At this time, the external tension N is all borne by the bolts, and ΔP is: （3） Due to the change in the force transmission mode, the N and P curves will overlap in the loose state, and the flange connection stiffness k3 in the loose state will be reduced, which is smaller than the flange connection stiffness k1 in the normal stress state, but larger than the slope k2 of the P curve in the normal stress state. The relationship between the three is: k2<k3<k1(4) According to Saint-Venant's principle, compared with the stress in the flange connection area, the vertical stress in the blade area far away from the flange connection area will not be affected by the degree of connection looseness; To this end, the loosening influence coefficient of the flange connection bolt group is introduced, that is, the strain ratio K is: K = ε2 / ε1 (5) Where ε1 is the strain at the upper monitoring point, and ε2 is the strain at the lower monitoring point; The detachment point A and height h of the strain ratio K value curve of the upper and lower monitoring points are determined by the finite element calculation results of the connecting flange bolt group and the cylinder wall.

7. The method for online monitoring of looseness of wind turbine blade root bolts and predicting fatigue life according to claim 1, characterized in that: In step S8, the minimum value N of the fatigue load on the most unfavorable bolt among the blade root bolts to be monitored is obtained from the host manufacturer. min , the maximum value of fatigue load on the most unfavorable bolt N max , and the average axial force P of the most unfavorable bolt is obtained mean : （6）。 8. The method for online monitoring of looseness of wind turbine blade root bolts and fatigue life prediction according to claim 6 is characterized in that: In step S9, for the running wind turbine blade, the actual pre-tension force P0 of the bolt, the minimum value N of the fatigue load on the most unfavorable bolt min and the maximum value of fatigue load on the most unfavorable bolt N max , due to the change in the stress state of the loose bolt, the actual fatigue stress amplitude changes, which can be divided into the following three situations: Case 1: N max ≤P0; When P0 has no loss, the pre-stress on the flange connection interface exists, the interface is not detached, the bolt is in normal working condition, ΔP meets the fatigue strength requirements, and the remaining fatigue life meets the use requirements; ΔP=P2-P1(7) ΔP≈0 (8) Where P2 is N max ≤P0, at N max The actual axial force of the most unfavorable bolt under action; P1 is N max ≤P0, at N min The actual axial force of the most unfavorable bolt under action; The second case: N min ≤P0<N max ; set up The critical bolt axial force for interface disengagement is Ps. If N max >P0≥Ps, there is a disengagement at the interface, and the bolt axial force changes completely with the external load; when N min When <P0<Ps, the flange connection interface pre-pressure exists, the interface is not detached, and the bolt is in normal working condition; so when N min ≤P0<N max When ΔP is calculated as follows: ΔP=N max -P3(9) Where P3 is N min ≤P0<N max When N min The actual axial force of the most unfavorable bolt under action; The third case: N min >P0; If N appears min >P0, the bolt pretension is completely lost, the interface is completely detached, and the bolt axial force is completely controlled by the external load, then: ΔP=N max -N min (10) It can be seen that the fatigue life prediction of the bolt is related to P0 and N max 、N min When P0 has no loss, N max ≤P0, ΔP≈0, the pre-pressure on the flange connection surface always exists, and the remaining life meets the use requirements; when N min ≤P0<N max , ΔP=N max -P3, perform remaining life prediction to determine whether it meets the use requirements; when N min >P0,ΔP=N max -N min , the pre-stress on the flange connection surface is completely lost, ΔP As the external force changes in amplitude, the remaining life is predicted to determine whether it meets the use requirements.

9. The method for online monitoring of looseness of wind turbine blade root bolts and fatigue life prediction according to claim 1, characterized in that: In step S10, the equivalent stress amplitude of the most unfavorable bolt is calculated as follows: and mean stress ; （11） （12） In the formula, A e is the effective area of ​​the bolt.

Images