Life prediction method and system for key components of high-altitude rotary drilling rig based on digital twinning

Abstract

The application belongs to the technical field of data processing, and particularly relates to a high-altitude rotary drilling rig key component life prediction method and system based on digital twinning, which comprises the following steps: collecting environmental temperature and air pressure, using a kernel function to obtain a correction coefficient to scale a baseline matrix, combining working conditions to solve a dynamic equation to obtain a virtual reference response; then, total residual error is calculated, and an exponential weighted moving average filter is used to separate out a cumulative deviation trend and an environmental transient deviation; then, the cumulative trend is used to correct the damage degree and estimate the remaining life, the environmental transient deviation is used to obtain a correction amount, and the correction amount is combined with a material brittleness index and a lubricating oil film thermal attenuation coefficient to modulate a limit and then update a matrix through inertia smoothing. The application reduces the interference of environmental factors on the reference response and residual error analysis, and improves the accuracy and robustness of life prediction.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to a method and system for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins. Background Technology

[0002] In pile foundation construction, rotary drilling rigs in high-altitude areas are subjected to extreme environmental conditions for a long time. Low temperatures increase the brittleness of alloy materials, resulting in increased stiffness and decreased damping. Low air pressure weakens the heat capacity of the air, worsens heat dissipation conditions, and reduces the load-bearing capacity of the lubricating oil film.

[0003] In technologies for predicting the lifespan of critical components, a dynamic equation is typically constructed based on the baseline stiffness matrix and baseline damping matrix calibrated under normal temperature and pressure conditions, combined with operating parameters. The virtual reference response is then solved and compared with the state response acquired from the physical sample to assess component damage.

[0004] However, this type of fixed-parameter model fails to account for the effects of ambient temperature and air pressure on the mechanical properties of materials in real time, resulting in a systematic and time-varying benchmark deviation between the virtual response and the physical response. This makes it difficult to accurately reflect the dynamic behavior of components in real high-altitude environments, thus affecting the accuracy of life prediction.

[0005] The total virtual-real residual obtained by subtracting the physical state response from the virtual baseline response simultaneously couples the reversible fluctuations caused by transient changes in environmental temperature and pressure with the irreversible trend caused by component degradation, and the two differ significantly in time scale.

[0006] Existing residual processing methods lack intrinsic decoupling of components from different sources, making it easier to introduce errors from rapid environmental fluctuations into degradation assessments, or to pollute the environmental compensation pathway due to slow degradation trends, resulting in false alarms, missed reports, and directional biases in adaptive correction of model parameters in lifetime prediction.

[0007] Furthermore, when correcting model parameters based on residuals, if the adjustment is based solely on the error magnitude without fully considering constraints such as the brittle limit of materials at low temperatures and the thermal decay of lubricating oil films, it is easy to cause overshoot in the correction amount under extreme working conditions, leading to numerical oscillations in the matrix update process and affecting the convergence and robustness of the digital twin state estimation. Summary of the Invention

[0008] To address the technical problem of interference from high-altitude environmental factors on digital twin benchmark response and residual analysis, this invention provides solutions in the following aspects.

[0009] In a first aspect, the present invention provides a method for predicting the lifespan of key components of a high-altitude rotary drilling rig based on digital twins, employing the following technical solution: The method for predicting the lifespan of key components of a high-altitude rotary drilling rig based on digital twins includes: 1. A method for predicting the lifespan of key components of a high-altitude rotary drilling rig based on digital twins, characterized in that it includes:

[0010] Real-time ambient temperature and air pressure are obtained, and stiffness and damping environmental correction coefficients are obtained through preset kernel functions. The baseline stiffness matrix and baseline damping matrix are scaled to obtain the environmentally adaptive stiffness matrix and environmentally adaptive damping matrix. The dynamic equations are then solved in combination with the operating condition parameters to obtain the virtual reference response.

[0011] Calculate the total residual between the entity state response and the virtual baseline response, extract the cumulative deviation trend from the total residual using an exponentially weighted moving average filter, and obtain the environmental transient deviation based on the total residual and the cumulative deviation trend;

[0012] The cumulative damage level of the component is corrected by gradient descent based on the cumulative deviation trend, and the remaining life is calculated based on the corrected damage level.

[0013] The original correction values ​​for stiffness and damping are obtained from the environmental transient deviation. The material brittleness index and lubricating oil film thermal decay coefficient are calculated based on the current temperature. The original correction values ​​are then subjected to gain modulation and limiting. The environmental adaptability stiffness matrix and environmental adaptability damping matrix are updated by inertial smoothing recursion.

[0014] Preferably, obtaining the stiffness environment correction coefficient and the damping environment correction coefficient through a preset kernel function includes: when generating the stiffness environment correction coefficient, the parameters used include: the temperature value in the real-time ambient temperature and air pressure, a preset material embrittlement activation energy, a universal gas constant, a reference temperature, and the air pressure value in the real-time ambient temperature and air pressure; when generating the damping environment correction coefficient, the parameters used include: the temperature value in the real-time ambient temperature and air pressure, a preset viscosity-temperature attenuation empirical coefficient, a reference temperature, and the air pressure value in the real-time ambient temperature and air pressure.

[0015] Preferably, scaling the baseline stiffness matrix and baseline damping matrix to obtain the environmentally adaptable stiffness matrix and environmentally adaptable damping matrix includes: multiplying the stiffness environmental correction coefficient by the baseline stiffness matrix to obtain the environmentally adaptable stiffness matrix; and multiplying the damping environmental correction coefficient by the baseline damping matrix to obtain the environmentally adaptable damping matrix.

[0016] Preferably, the operating parameters include the power head speed and torque load.

[0017] Preferably, the step of extracting the cumulative deviation trend from the total residual using exponentially weighted moving average filtering, and obtaining the environmental transient deviation based on the total residual and the cumulative deviation trend, includes: using a preset trend smoothing coefficient to perform a weighted summation of the total residual of the current period and the cumulative deviation trend stored in the previous period to obtain the cumulative deviation trend of the current period; and subtracting the cumulative deviation trend of the current period from the total residual of the current period to obtain the environmental transient deviation.

[0018] Preferably, the gradient descent correction of the cumulative damage level of the component based on the cumulative deviation trend includes: applying a preset positive disturbance value to the cumulative damage level of the component to obtain the disturbance damage level; based on the disturbance damage level, resolving the dynamic equation using the environmental adaptability stiffness matrix, the environmental adaptability damping matrix, and the operating condition parameters to obtain the disturbance virtual reference response; determining the damage sensitivity based on the difference between the virtual reference response and the disturbance virtual reference response; and updating the cumulative damage level of the component based on the damage sensitivity, the cumulative deviation trend, and the preset gradient correction step size.

[0019] Preferably, the calculation of the material brittleness index and the lubricating oil film thermal decay coefficient includes: the material brittleness index is calculated using an S-shaped growth function based on the current temperature, a preset saturation embrittlement degree, a critical temperature for ductile-brittle transition, and a brittleness curve steepness coefficient; the lubricating oil film thermal decay coefficient is calculated using an exponential function based on the current temperature, the characteristic reference temperature of the lubricating oil, and the viscosity-temperature coefficient.

[0020] Preferably, the original correction amounts for stiffness and damping are obtained from the environmental transient deviation. The material brittleness index and lubricating oil film thermal decay coefficient are calculated based on the current temperature. Gain modulation and limiting are then applied to the original correction amounts, including: calculating the norm of the environmental transient deviation; multiplying the norm, a preset stiffness compensation gain coefficient, and the material brittleness index to obtain the original stiffness correction coefficient; multiplying the original stiffness correction coefficient by the baseline stiffness matrix to obtain the original stiffness correction amount matrix; and multiplying the norm, the preset damping compensation gain coefficient, and... The original damping correction coefficient is obtained by multiplying the lubricating oil film thermal decay coefficient, and then multiplied by the baseline damping matrix to obtain the original damping correction amount matrix. The original stiffness correction amount matrix is ​​limited so that its absolute value does not exceed the product of the preset maximum relative stiffness correction rate, the material brittleness index, and the baseline stiffness matrix. The original damping correction amount matrix is ​​also limited so that its absolute value does not exceed the product of the preset maximum relative damping correction rate, the lubricating oil film thermal decay coefficient, and the baseline damping matrix.

[0021] Preferably, the step of updating the environmental adaptability stiffness matrix and the environmental adaptability damping matrix recursively via inertial smoothing includes: multiplying the stiffness correction amount after limiting by a preset inertial smoothing factor to obtain a stiffness smoothing correction amount; multiplying the damping correction amount after limiting by the inertial smoothing factor to obtain a damping smoothing correction amount; adding the current environmental adaptability stiffness matrix to the stiffness smoothing correction amount to obtain an updated environmental adaptability stiffness matrix; and adding the current environmental adaptability damping matrix to the damping smoothing correction amount to obtain an updated environmental adaptability damping matrix.

[0022] Secondly, the present invention provides a life prediction system for key components of high-altitude rotary drilling rigs based on digital twins, which adopts the following technical solution: The life prediction system for key components of high-altitude rotary drilling rigs based on digital twins includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned life prediction method for key components of high-altitude rotary drilling rigs based on digital twins is implemented.

[0023] Beneficial effects:

[0024] 1. By establishing an environmental stress sensitivity kernel function with real-time temperature and air pressure as input, the stiffness and damping correction coefficients are dynamically obtained and the baseline matrix is ​​scaled to obtain a virtual reference response with environmental following capability, thereby reducing the system reference deviation caused by high-altitude low-temperature embrittlement and low-pressure heat dissipation degradation.

[0025] 2. In the environmental deviation compensation link, the low-temperature brittleness index of the material and the thermal decay coefficient of the lubricating oil film are calculated in real time. The original corrections for stiffness and damping are dynamically modulated and limited. An inertial smoothing factor is introduced to recursively update the model matrix, thereby limiting the corrections to a reasonable envelope defined by the material properties. This prevents overshoot under extreme conditions from causing parameter oscillations and model divergence, avoids inherent frequency errors and positive feedback instability, and ensures the convergence of the valuation and the stability of the life prediction of the digital twin during long-term operation. Attached Figure Description

[0026] Figure 1 The flowchart illustrates the steps of the method for predicting the lifespan of key components of a high-altitude rotary drilling rig based on digital twins in this invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] S1: Obtain real-time ambient temperature and air pressure.

[0030] In high-altitude areas, the daily temperature fluctuation range of the rotary drilling rig's operating environment is large. Low temperature increases the brittleness of the alloy materials in the power head transmission system, resulting in increased stiffness and decreased damping. Low air pressure reduces the air's heat capacity, worsens heat dissipation conditions, and affects the load-bearing capacity of the lubricating oil film.

[0031] If a preset baseline stiffness matrix is ​​used and baseline damping matrix All calibrations are based on ambient temperature and pressure conditions, which cannot reflect the material property changes induced by the aforementioned environment. and There must be a systematic and time-varying deviation between the virtual response obtained by solving the dynamic equation and the actual state response.

[0032] Therefore, when obtaining the virtual reference response, real-time environmental factors can be incorporated into the dynamic equations. Temperature and pressure sensors deployed on the drilling rig body can collect the current ambient temperature (T, in Kelvin) and pressure (P, in Pa) in real time. The stiffness environmental correction coefficient can then be obtained through a pre-set environmental stress sensitivity kernel function. and damping environment correction factor , used to scale the baseline matrix.

[0033] The kernel function obtains the stiffness environment correction coefficient. At that time, using real-time temperature T and air pressure P as inputs, the material embrittlement activation energy preset in the physical property parameter library is called. Universal gas constant R and modeling reference temperature , and the pressure correction polynomial f(P) solidified in the kernel function.

[0034] The specific calculation formula is as follows:

[0035]

[0036] In the formula, This is the stiffness scaling factor, with an empirical value of 0.15. It is dimensionless and used to control the overall magnitude of the correction. It can also be adjusted by the implementer according to the specific implementation scenario. The embrittlement activation energy is expressed in J / mol. It is determined by conducting Charpy impact tests on the same batch of materials at different temperatures to obtain brittle transition temperature profiles. The impact absorption energy at each temperature is then expressed as... After normalization, fit the Arrhenius relation. The least squares method is used to solve the problem. .

[0037] Get experience points It can be adjusted by the implementer according to the specific implementation scenario; the general gas constant R is taken as 8.314 J / (mol·K); The reference temperature for modeling is set to 298 Kelvin.

[0038] In order to use air pressure ratio The quadratic polynomial correction term for the independent variable: ,in Based on preliminary thermodynamic simulations and material testing, the coefficients were determined using empirical values. , , Similarly, the air pressure correction term for the damping channel... The coefficient is taken as an empirical value. , , These coefficients are fixed in the kernel function and will not be adjusted once fixed.

[0039] This exponential term makes it possible when the temperature is much lower hour, A value greater than 1 indicates that stiffness increases due to material embrittlement; the lower the air pressure, the higher the stiffness. Further modulate this correction magnitude.

[0040] Damping environmental correction factor The result also depends on the current temperature T and air pressure P, but it reflects the combined effect of temperature and air pressure on viscous damping characteristics.

[0041] The calculation formula is as follows:

[0042]

[0043] In the formula, β is the damping scaling factor, taken as an empirical value of 0.20, which is dimensionless; β is the empirical coefficient for viscosity-temperature attenuation, taken as an empirical value of 0.03. ; The value is 298K.

[0044] The pressure correction term for the damping channel has a structure similar to... Similar to, but with different coefficients in the corresponding polynomials, these coefficients are also determined by fitting previous experimental data and embedded in the kernel function. The system takes the experience value as the standard. , , .

[0045] This expression makes the temperature decrease. A value less than 1 reflects the law that the increased viscosity of lubricating oil at low temperatures leads to a decrease in damping, and that the decrease in air pressure is due to... The degree of damping reduction is further modulated.

[0046] At each simulation step, determine whether the changes in the current ambient temperature T and air pressure P compared to the previous cycle exceed the preset threshold, or whether it is the first run.

[0047] If the threshold is exceeded or this is the first run, the environment adaptability matrix is ​​obtained again: This retrieves the current environment's... and Then, the baseline stiffness matrix and baseline damping matrix Element-wise scaling is performed to obtain the updated environmental adaptability stiffness matrix. and environmentally adaptable damping matrix The scaling relationship is as follows:

[0048]

[0049]

[0050] In the formula, and These are the baseline stiffness and baseline damping matrices under normal temperature and pressure, respectively, which are pre-calculated and stored from the finite element model of the drilling rig drive system under conventional material mechanics parameters; otherwise, the environmental adaptability matrix updated recursively from the previous cycle is used. and .

[0051] The above and As the current environmental adaptability matrix, it is passed to the dynamic school kernel solver.

[0052] After obtaining the current environmental adaptability matrix, the current operating condition parameters are obtained from the rotary drilling rig control system, mainly including the power head speed n (in r / min) and torque load M (in N·m).

[0053] Construct the external load vector based on n and M Combined with the generalized quality matrix The dynamic equations of the transmission system are established as follows:

[0054]

[0055] In the formula, The generalized mass matrix is ​​automatically obtained from the finite element mesh and material density assignment. , and These are the nodal acceleration, velocity, and displacement response vectors, respectively.

[0056] Implicit time-domain solution is performed using the Newmark-β numerical integration method, with parameters selected as follows: , That is, the average acceleration method, which is unconditionally stable.

[0057] Integral step size Take the cycle corresponding to the highest engagement frequency of the power head. The specific value is determined by the sampling frequency. Decide: ,in The highest gear meshing frequency is calculated from the highest speed and the number of teeth, and the simulation state vector of each node in the power head transmission link is obtained.

[0058] Based on the corresponding locations of the installed sensors, the vibration response characteristics at the nodes are extracted to form a virtual reference response set. The specific extraction method is as follows:

[0059] Root mean square value of acceleration: time history of nodal acceleration exist Calculation within a second time window The unit is m / s².

[0060] Gear meshing frequency amplitude: Extract the meshing frequency by performing a fast Fourier transform on the nodal acceleration time history. ( The rotational speed of the power head. The amplitude at the gear tooth count (in m / s²) is given by 60, which is the conversion factor from minutes to seconds.

[0061] Axial trajectory characteristics: Take the displacement responses in two perpendicular directions (X, Y) on the same cross section to form a Lissajous figure, and calculate its major axis length. (Unit: m) is used as the characteristic value. Each feature in the model maintains the same definition and calculation method as the corresponding feature acquired by the physical sensor.

[0062] S2: Calculate the total residual between the physical state response and the virtual reference response.

[0063] An exponentially weighted moving average filter is used to extract the cumulative deviation trend from the total residual, and the environmental transient deviation is obtained based on the total residual and the cumulative deviation trend.

[0064] Entity state response set The data is acquired in real time through a data acquisition system connected to a network of vibration acceleration sensors installed on the power head drive link of the drilling rig. The signal is first filtered by a digital bandpass filter to remove high-frequency measurement noise.

[0065] Simultaneously, for each sensor channel, the absolute deviation between the current sampled value and the value of the previous cycle is calculated. If the deviation exceeds a preset threshold, it is determined to be a transient anomaly. The threshold for the acceleration channel is taken as an empirical value. The frequency amplitude channel threshold is taken as an empirical value. The threshold value for the axis trajectory displacement channel is taken as an empirical value. All of these can be adjusted by the implementer according to the specific implementation scenario.

[0066] For outliers, the valid value from the previous cycle is used to replace them, and the outlier count is recorded. If the same channel is outlier for three consecutive cycles, a sensor fault alarm is output, and lifespan prediction is paused. The number of consecutive outliers for the same channel (3 cycles) is an empirical value that can be adjusted by the implementer according to the specific implementation scenario.

[0067] In each virtual-to-real comparison cycle, the filtered data will be... The set of virtual reference responses obtained by solving in the same period The difference is performed on each feature component. The virtual reference response set used includes the root mean square value of acceleration, the amplitude of gear meshing frequency, and the shaft center trajectory features. The difference operation yields the total virtual and real residual vector. Its calculation formula is , Dimensions of each element and The corresponding features are the same.

[0068] Reversible fluctuations in lubricating oil film thickness and tooth clearance caused by diurnal temperature variations at high altitudes can occur within minutes to hours. This is reflected in the fact that irreversible degradation processes such as pitting on gear teeth and bearing wear cause vibration characteristics to slowly increase on a timescale of several weeks to several months. These two different timescales of change are reflected in the fact that... Superimposed in the middle.

[0069] If When used directly on a single compensation link, rapid temperature effects can interfere with the assessment of degradation trends, while slow degradation shifts can also cause misadjustments to environmental compensation.

[0070] Therefore, an exponentially weighted moving average filter is used to... Low-pass smoothing and exponentially weighted moving average filtering are well-known signal processing techniques and will not be elaborated further. The slowly changing cumulative deviation trend component is separated, and its iterative calculation formula is as follows.

[0071]

[0072] In the formula, This is a preset trend smoothing coefficient, taken as an empirical value of 0.02. It is dimensionless and can be adjusted by the implementer according to the specific implementation scenario. For the current period The total residual vector, This represents the cumulative deviation trend component stored from the previous period. For the updated current periodic trend component, this filter is applied to... Each feature component in the data is independently subjected to the above weighted moving average operation. A value of 0.02 ensures that the cumulative deviation trend component primarily tracks changes over time scales of several hours or more, while attenuating rapid environmental fluctuations on the order of minutes.

[0073] After obtaining the cumulative deviation trend component for the current period, calculate the environmental transient deviation component. The calculation formula is:

[0074]

[0075] In the formula, This represents the total residual vector for the current period. This represents the cumulative deviation trend component for the current cycle. It primarily reflects reversible disturbances caused by transient changes in ambient temperature and air pressure. This reflects the irreversible trend of permanent degradation of the component structure, thereby separating the total residual into the two independent components mentioned above, and completing the residual decoupling process of the current step.

[0076] S3: Apply gradient descent correction to the cumulative damage level of the component based on the cumulative deviation trend, and calculate the remaining life based on the corrected damage level.

[0077] In the power head transmission system of high-altitude rotary drilling rigs, the cumulative effect of degradation processes such as pitting on gear teeth and wear of bearing rolling elements gradually changes the load-bearing capacity and stiffness distribution of the material, resulting in a continuously expanding systematic offset between the virtual reference response calculated based on the current damage state and the physical state response.

[0078] The cumulative deviation trend component extracted by the aforementioned exponentially weighted moving average filter This reflects the slow growth trend of this irreversible degradation in the vibration characteristic space.

[0079] The cumulative damage level of the component is corrected by gradient descent based on the cumulative deviation trend, and the remaining life is calculated based on the corrected damage level.

[0080] Obtain the cumulative damage level of components , dimensionless, between 0 and 1, where 0 represents a brand new state and 1 represents a complete loss of bearing capacity.

[0081] When starting this component for the first time, it is not possible to know directly. The true value is therefore required based on the initial measured vibration response. With virtual benchmark response At this point, let's assume The calculated differences are used to reverse the calculation and determine the initial residuals. Then, by iterating several times using the same gradient descent approach as in S3, the initial value is obtained. If there is no prior information, conservatively initialize to .

[0082] Since the degree of cumulative damage to a component is an internal state variable of the nonlinear damage accumulation model of the material built into the digital twin, it cannot be directly measured from the outside. Its evolution can only be inferred by observing the changes in the virtual and real residuals.

[0083] Cumulative Deviation Trend Component The direction and magnitude of the current degradation bias are given in the multidimensional feature space, as well as the degree of damage at the end of the previous control cycle. .

[0084] like Non-zero means that the current The corresponding virtual response and the entity exhibit unexplained biases consistent with the degradation trend, therefore requiring numerical gradient descent to address this. Adjustments were made to make the virtual response more closely resemble the actual physical measurement.

[0085] To obtain the local sensitivity of the virtual response to the degree of damage, a numerical perturbation method is used. Make small perturbations.

[0086] Maintain the environmental adaptability stiffness matrix during the current cycle. Environmental Adaptive Damping Matrix and the speed of the power head and torque load When the operating parameters remain unchanged, for Apply a preset positive disturbance value , The empirical value is 0.01, which is dimensionless and can be adjusted by the implementer according to the specific implementation scenario. These are preset parameters.

[0087] Will As the degree of disturbance damage is determined, a complete dynamic re-solution process is triggered: using the generalized mass matrix. and the current environmental adaptability matrix , and based on and Constructed external load vector Construct the dynamic equations under the perturbation state. The simulation state vectors of each node after perturbation are obtained by solving the problem using the well-known Newmark-Beta numerical integration method, and then the virtual reference response is extracted. The corresponding characteristic components constitute the perturbation virtual reference response. .

[0088] The process of determining damage sensitivity is to Compared with the virtual baseline response before the disturbance Perform component-by-component subtraction and use the perturbation value Normalization.

[0089] The specific calculation formula is as follows:

[0090]

[0091] In the formula, Let be the damage sensitivity vector, whose dimension is... The same, each component is represented in The change in the corresponding vibration characteristic value when a unit change occurs is expressed as the ratio of the corresponding characteristic value to the degree of damage. For example, the sensitivity unit for the root mean square value component of acceleration is m / s², the sensitivity unit for the amplitude component of gear meshing frequency is m / s², and the sensitivity unit for the characteristic component of shaft trajectory is m.

[0092] The aforementioned preset parameter, with a value of 0.01, ensures the accuracy of the numerical derivative while avoiding excessive disturbances that could cause the dynamic response to exceed the near-linear range.

[0093] In obtaining damage sensitivity Then, the cumulative deviation trend component As a measure of current degradation error, along the sensitivity direction... Implement gradient descent updates.

[0094] The updated calculation formula is:

[0095]

[0096] In the formula, The corrected cumulative damage level of the component is dimensionless, and its value range is still limited to between 0 and 1; This is the preset gradient correction step size, taken as an empirical value of 0.1. It is dimensionless and used to control the single-step update magnitude. It is also a preset parameter.

[0097] The dot product operation between the cumulative bias trend component and the damage sensitivity vector maps the multidimensional residual features to a scalar, since... and The components have different units; for example, the root mean square sensitivity of acceleration has a unit of . The unit of frequency amplitude sensitivity is The dot product result numerically reflects the error contribution in each feature direction.

[0098] To ensure balanced weights for each feature, each component is dimensionless before the dot product: multiplied by the standard deviation of the corresponding feature in historical data. That is, in actual calculations The summation represents the magnitude of the projection degradation error in the current sensitivity direction.

[0099] like Components and Sensitivity If the corresponding components have the same sign and a large amplitude, then the dot product is positive and significant, driving... Comparison A significant increase indicates a worsening of the damage; conversely, if the residual and sensitivity are not aligned, the change is smaller.

[0100] Cumulative damage to components After being updated, it is written into the built-in material nonlinear damage accumulation model.

[0101] The damage update model is a well-known fatigue accumulator based on the modified Miner criterion, and it obtains the cumulative values ​​of each stress level and its corresponding number of cycles throughout the entire service life.

[0102] Current level of damage With the critical cumulative damage threshold To make a comparison, The empirical value of 0.95 was determined by statistical analysis of fatigue test data of the transmission component material of this model. It is dimensionless and is a preset parameter.

[0103] Therefore, the specific calculation method for the remaining lifespan is as follows: First, based on the power head rotation speed... and torque load The corresponding equivalent stress amplitude is obtained through the stored load spectrum mapping table. The units are MPa and the number of cycles per unit time. The unit is 1 / hour.

[0104] Then, based on the modified Miner linear accumulation criterion, the current damage level... The corresponding percentage of lifespan already consumed is .

[0105] Suppose that from the current moment until the damage reaches... The required number of remaining loops is ,in, The fatigue life cycle count under the standard load spectrum is obtained from the material's SN curve. The corresponding cycle number, To determine the unit cyclic damage contribution under the corresponding standard load, it is typically taken as... .

[0106] Final remaining service life The unit is hours. If the current operating conditions change, the calculation will be recalculated every cycle. and .

[0107] After this process, the damage level correction and lifetime prediction driven by degradation bias within the current cycle are completed, and the results are... This serves as a prediction of the remaining lifespan of key components provided by the current digital twin.

[0108] Meanwhile, in order to make the virtual baseline response of the next cycle more accurately reflect the current environmental conditions, it is necessary to update the model matrix using environmental transient biases to ensure that a continuous and stable lifetime prediction result can be obtained.

[0109] S4: Obtain the original corrections for stiffness and damping from the environmental transient deviations.

[0110] Within the current virtual-to-real comparison period, the environmental transient deviation vector It includes reversible disturbances caused by instantaneous fluctuations in ambient temperature and air pressure, and the current environmental adaptability stiffness matrix and environmentally adaptable damping matrix The virtual reference response has not yet fully followed the disturbance, resulting in a residual deviation between the virtual reference response and the physical state response. The drive obtains compensation corrections for stiffness and damping. Before obtaining these corrections, the current temperature is considered. The brittleness index of the material is calculated using the S-shaped growth function pre-stored in the physical property parameter library. The calculation formula is:

[0111]

[0112] In the formula, The degree of saturation embrittlement is represented by an empirical value of 2.5, which is dimensionless and reflects the maximum amplification factor of the stiffness improvement of the alloy material at extreme low temperatures. The critical temperature for the ductile-brittle transition is taken as an empirical value of 243 Kelvin. The steepness coefficient of the brittle curve is taken as an empirical value of 0.2K. , , and These are all preset parameters, but can also be adjusted by the implementer according to the specific implementation scenario. The value of is dimensionless; its larger value indicates greater material brittleness and a higher upper limit for allowable stiffness compensation. The empirical value can be adjusted by the implementer based on the specific implementation scenario. This represents a sigmoid function mapping over T.

[0113] Lubricating oil film thermal decay coefficient Based on the current temperature by the exponential function The calculation is as follows:

[0114]

[0115] In the formula, The reference temperature for the characteristics of the lubricating oil is taken as an empirical value of 313 Kelvin; The viscosity-temperature coefficient is taken as an empirical value of 0.04K. , and These are preset parameters, but can also be adjusted by the implementer based on the lubricant brand and the results of on-site oil sample analysis. Dimensionless, when Below hour A value greater than 1 indicates enhanced oil film load-bearing capacity, allowing for greater damping compensation. Higher than hour If the value is less than 1, the upper limit of damping compensation will decrease accordingly. This represents an exponential mapping to T.

[0116] Get and After that, firstly The components are dimensionless: Let ,in For the first The standard deviation of each characteristic component over a historical period without failures.

[0117] Then calculate the dimensionless norm. , Dimensionless, uniformly representing the comprehensive intensity of transient environmental deviations.

[0118] Then, the preset stiffness compensation gain coefficient is used. and damping compensation gain coefficient The original stiffness correction coefficients were obtained respectively. and original damping correction factor ,in Take an empirical value of 0.05 (m / s) ) , Take the empirical value of 0.08 (m / s) ) All of these are preset parameters.

[0119] Based on this correction factor, the original stiffness correction matrix is ​​defined as follows: The original damping correction matrix is , and These are the baseline stiffness matrix and baseline damping matrix at room temperature, respectively, which convert the overall intensity of environmental transient deviations into a matrix correction proportional to the baseline matrix.

[0120] Subsequently and Implement explicit bandwidth limiting, targeting Each element If its absolute value Maximum relative correction rate for stiffness exceeding the preset value and and The product of is then truncated to its absolute value. The symbols remain unchanged.

[0121] because This limit is equivalent to the coefficient. Truncation: If Then take ,otherwise ,in We take an empirical value of 0.1, which is dimensionless, as a preset parameter, and similarly adjust the original damping correction coefficient. Using the maximum relative correction rate of damping Limit the amplitude. Taking an empirical value of 0.15, which is dimensionless, as a preset parameter, we obtain... The original correction matrix after limiting is then scaled accordingly. and The preset parameters and empirical values ​​can be adjusted by the implementer according to the specific implementation scenario.

[0122] Applying the entire correction value after amplitude limiting at once may cause numerical oscillations in the simulation solver; therefore, an inertial smoothing factor is introduced. , Using an empirical value of 0.3, dimensionless, as a preset parameter, calculate the stiffness smoothing correction coefficient. and damping smoothing correction coefficient And construct the smoothing correction matrix. and .

[0123] Then the current environmental adaptability stiffness matrix and Add them together to obtain the initial update matrix. .

[0124] examine Whether it remains positive definite: Calculate its smallest eigenvalue. ,like ,but Otherwise, for Perform diagonal loading correction: ,in , It is an identity matrix.

[0125] Similarly, the environmental adaptability damping matrix is ​​handled: ,like but ,otherwise , The same rules are used for calculation.

[0126] This correction ensures the numerical stability of the dynamic solution through the aforementioned recursive update. and It is gradually approaching the true properties of the material under the current temperature and pressure.

[0127] It should be noted that the ambient temperature is measured at the beginning of each cycle. and air pressure Has there been a significant change compared to the previous cycle?

[0128] If temperature changes Kelvin or pressure change The kernel function will then be re-obtained. and and directly to the baseline matrix , Scaling yields new , This operation will overwrite the previously updated matrix, where, Kelvin and These are empirical values, which can be adjusted by the implementer based on the specific implementation scenario.

[0129] When the environment changes drastically, the original fine compensation based on the accumulation of small transient deviations can no longer reflect the new material properties, and a new benchmark needs to be established; conversely, if the environmental change does not exceed the threshold, the compensation in step S4 will continue to accumulate, enabling the model to adaptively track the slow environmental drift.

[0130] The two mechanisms work together: the former is responsible for rapid reset after significant environmental changes, while the latter is responsible for smooth follow-up under minor environmental fluctuations, thereby achieving accuracy in long-term life prediction in high-altitude environments.

[0131] This invention also discloses a life prediction system for key components of high-altitude rotary drilling rigs based on digital twins, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the life prediction method for key components of high-altitude rotary drilling rigs based on digital twins. The system also includes other components well-known to those skilled in the art, such as a communication bus and communication interface; their configurations and functions are known in the art and will not be described further here.

[0132] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make several improvements and substitutions without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A method for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins, characterized in that, include: Real-time ambient temperature and air pressure are obtained, and stiffness and damping environmental correction coefficients are obtained through preset kernel functions. The baseline stiffness matrix and baseline damping matrix are scaled to obtain the environmentally adaptive stiffness matrix and environmentally adaptive damping matrix. The dynamic equations are then solved in combination with the operating condition parameters to obtain the virtual reference response. Calculate the total residual between the entity state response and the virtual baseline response, extract the cumulative deviation trend from the total residual using an exponentially weighted moving average filter, and obtain the environmental transient deviation based on the total residual and the cumulative deviation trend; The cumulative damage level of the component is corrected by gradient descent based on the cumulative deviation trend, and the remaining life is calculated based on the corrected damage level. The original correction values ​​for stiffness and damping are obtained from the environmental transient deviations. The material brittleness index and lubricating oil film thermal decay coefficient are calculated based on the current temperature. Gain modulation and limiting are applied to the original correction values. The environmentally adaptive stiffness matrix and environmentally adaptive damping matrix are then updated via inertial smoothing recursion, including: Calculate the norm of the environmental transient deviation; Multiply the norm, the preset stiffness compensation gain coefficient, and the material brittleness index to obtain the original stiffness correction coefficient, and multiply the original stiffness correction coefficient with the baseline stiffness matrix to obtain the original stiffness correction amount matrix. Multiply the norm, the preset damping compensation gain coefficient, and the lubricating oil film thermal attenuation coefficient to obtain the original damping correction coefficient, and multiply the original damping correction coefficient with the baseline damping matrix to obtain the original damping correction amount matrix. The original stiffness correction matrix is ​​limited so that its absolute value does not exceed the product of the preset maximum relative stiffness correction rate and the material brittleness index and the baseline stiffness matrix; The original damping correction matrix is ​​limited so that its absolute value does not exceed the product of the preset maximum relative damping correction rate, the lubricating oil film thermal attenuation coefficient, and the baseline damping matrix. Multiply the stiffness correction amount after the amplitude limit by the preset inertia smoothing factor to obtain the stiffness smoothing correction amount; Multiply the damping correction amount after amplitude limiting by the inertial smoothing factor to obtain the damping smoothing correction amount; Add the current environmental adaptability stiffness matrix to the stiffness smoothing correction amount to obtain the updated environmental adaptability stiffness matrix; The current environmental adaptability damping matrix is ​​added to the damping smoothing correction amount to obtain the updated environmental adaptability damping matrix.

2. The method for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins according to claim 1, characterized in that, The process of obtaining the stiffness environment correction coefficient and damping environment correction coefficient through a preset kernel function includes: The parameters used in generating the stiffness environment correction coefficient include: the temperature value in the real-time ambient temperature and air pressure, the preset material embrittlement activation energy, the universal gas constant, the reference temperature, and the air pressure value in the real-time ambient temperature and air pressure. The parameters used in generating the damping environment correction coefficient include: the temperature value in the real-time ambient temperature and air pressure, the preset viscosity-temperature attenuation empirical coefficient, the reference temperature, and the air pressure value in the real-time ambient temperature and air pressure.

3. The method for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins according to claim 1, characterized in that, The scaling of the baseline stiffness matrix and baseline damping matrix to obtain the environmentally adaptive stiffness matrix and environmentally adaptive damping matrix includes: The environmental stiffness correction coefficient is multiplied by the baseline stiffness matrix to obtain the environmental adaptability stiffness matrix; The environmental adaptation damping matrix is ​​obtained by multiplying the damping environment correction coefficient by the baseline damping matrix.

4. The method for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins according to claim 1, characterized in that, The operating parameters include the power head speed and torque load.

5. The method for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins according to claim 1, characterized in that, The step of extracting the cumulative deviation trend from the total residual using exponentially weighted moving average filtering, and obtaining the environmental transient deviation based on the total residual and the cumulative deviation trend, includes: Using a preset trend smoothing coefficient, the total residual of the current period and the cumulative deviation trend stored in the previous period are weighted and summed to obtain the cumulative deviation trend of the current period. The environmental transient deviation is obtained by subtracting the cumulative deviation trend of the current period from the total residual of the current period.

6. The method for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins according to claim 1, characterized in that, Applying gradient descent correction to the cumulative damage level of the component based on the cumulative deviation trend includes: A preset positive disturbance value is applied to the cumulative damage level of the component to obtain the disturbance damage level; Based on the degree of disturbance damage, the dynamic equations are re-solved using the environmental adaptability stiffness matrix, the environmental adaptability damping matrix, and the operating condition parameters to obtain the virtual reference response to the disturbance. The damage sensitivity is determined based on the difference between the virtual reference response and the perturbation virtual reference response; The cumulative damage level of the component is updated based on the damage sensitivity, the cumulative deviation trend, and the preset gradient correction step size.

7. The method for predicting the lifespan of key components of high-altitude rotary drilling rigs based on digital twins according to claim 1, characterized in that, The calculation of the material brittleness index and the lubricating oil film thermal decay coefficient includes: The material brittleness index is calculated using an S-shaped growth function based on the current temperature, a preset degree of saturation embrittlement, a critical temperature for the ductile-brittle transition, and a brittleness curve steepness coefficient. The thermal decay coefficient of the lubricating oil film is calculated using an exponential function based on the current temperature, the characteristic reference temperature of the lubricating oil, and the viscosity-temperature coefficient.

8. A life prediction system for key components of high-altitude rotary drilling rigs based on digital twins, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the life prediction method for key components of high-altitude rotary drilling rigs based on digital twins according to any one of claims 1-7.

Images