Vertical mill gearbox gear life online intelligent regulation and control method and system

Abstract

This invention belongs to the field of vertical mill gearbox technology, specifically disclosing an online intelligent control method and system for the life of vertical mill gearbox gears. The method includes the following steps: acquiring the gear speed signal and gearbox vibration acceleration signal of the main test gearbox to determine whether the gear has entered the degradation stage; if the gear has entered the degradation stage, constructing a life prediction model and using the life prediction model to predict the remaining life of the vertical mill gearbox gears; combining the predicted remaining life of the vertical mill gearbox gears with the current expected remaining life, calculating the optimal control parameters based on the model-free adaptive control (MFAC) theory and introducing pseudo-partial derivatives (PPD) and a compact dynamic linearized (CFDL) model; and controlling the operating conditions of the vertical mill gearbox gears based on the optimal control parameters. Using this technical solution, the remaining life of the vertical mill gearbox gears is predicted, enabling real-time control of the gearbox gear operating state and extending the actual service life of the vertical mill gearbox gears to near the expected service life.

Description

Technical Field

[0001] This invention belongs to the field of vertical mill gearbox technology, and relates to a method and system for online intelligent control of gear life in vertical mill gearboxes. Background Technology

[0002] Vertical mill gearboxes are one of the core components of power transmission equipment in industries such as cement and mining, and their operational stability directly affects the efficiency and safety of the production line. However, due to their complex working environment, large load variations, and the fact that the gears and bearings inside the gearbox are subjected to harsh conditions such as high temperature, high humidity, and vibration for extended periods, they are prone to wear, failure, and even major accidents.

[0003] Therefore, conducting research on an online intelligent control system for the lifespan of vertical mill gearboxes is of great significance for extending the remaining service life of vertical mill gearboxes and improving the reliability and stability of the transmission system.

[0004] In practical applications of vertical mill equipment in cement, mining, and other industries, the actual service life of the equipment can often be effectively extended by reasonably adjusting the load on the equipment and optimizing the operating conditions of the gears without affecting the normal function of the equipment.

[0005] However, since most traditional gear life prediction systems are limited to the remaining life prediction stage, they cannot extend the actual service life to the expected life life and cannot achieve the goal of intelligent operation and maintenance of machinery. Summary of the Invention

[0006] The purpose of this invention is to provide an online intelligent control method and system for the life of gearbox gears in a vertical mill, so as to realize real-time control of the operating status of gearbox gears and extend the actual service life of gearbox gears in the vertical mill to close to the expected service life.

[0007] To achieve the above objectives, the basic solution of this invention is: an online intelligent control method for the gear life of a vertical mill gearbox, comprising the following steps:

[0008] Every 5 minutes, the gear speed signal and the vibration acceleration signal of the main test gearbox are collected for 1 second (intermittently).

[0009] Based on the 3σ criterion, it is determined whether the gear has entered the degradation stage. If the gear has entered the degradation stage, a life prediction model is constructed and used to predict the remaining life of the gearbox of the vertical mill.

[0010] Combining the predicted remaining life of the gearbox of the vertical mill with the current expected remaining life, the optimal control torque is calculated based on the model-free adaptive control MFAC theory and by introducing pseudo-partial derivative PPD and compact dynamic linearized CFDL model.

[0011] The operating conditions of the gearbox gears in the vertical mill are adjusted based on the optimal control parameters.

[0012] The working principle and beneficial effects of this basic scheme are as follows: This technical scheme collects the gear speed signal and housing vibration acceleration signal of the main test gearbox, and monitors in real time whether the gears have entered the degradation stage based on the 3σ criterion. When the gears enter the degradation stage, the remaining service life of the vertical mill gearbox is predicted, and then the optimal control parameters are output based on the predicted remaining service life and the current expected remaining service life. This achieves accurate prediction and online intelligent control of the remaining service life of the vertical mill gearbox, successfully extending the actual service life of the gearbox to close to the expected service life, ensuring the working efficiency and long-term stability of the gear transmission system, and reducing economic losses and safety risks caused by accidental gear damage.

[0013] Furthermore, the method for constructing the lifetime prediction model is as follows:

[0014] S1. Select the statistical features with the best trend and monotonicity of the gear speed signal and the vibration acceleration signal of the main test gearbox throughout the whole life test as the health indicators of the gear. Form a health indicator sequence throughout the whole life test. Construct life samples on the health indicator sequence with a window length of 512 and a step size of 1 using the sliding window method.

[0015] S2, construct a multi-dimensional fusion feature extractor, extract multi-dimensional feature information related to the degradation state and remaining lifespan in the lifetime samples in parallel, fuse the multi-dimensional feature information, and generate the final output of the multi-dimensional fusion feature extractor;

[0016] S3, Construct a prediction ensemble module. The prediction ensemble module includes multiple domain-specific predictors. Each domain-specific predictor integrates multiple prediction heads and predicts the remaining lifetime results of the multivariate fusion features of the corresponding specific source domain samples. The remaining lifetime results of multiple prediction heads are fused through an ensemble learning strategy to generate the remaining lifetime prediction output of the gearbox gears.

[0017] S4. Design a multi-level, multi-source domain adaptive strategy to perform local and global alignment of feature distributions in the source and target domains, as well as alignment of the output boundary of the predictor, thereby reducing the distribution differences between data features in different domains and improving the domain-invariant representation ability of features.

[0018] S5, construct the target network function using multi-level multi-source domain adaptive loss and prediction losses of source domain and target domain;

[0019] The target network function is optimized using backpropagation and gradient descent algorithms to obtain the target network.

[0020] A life prediction model is constructed to facilitate its use and enable accurate prediction of the remaining service life of gears in a vertical mill gearbox.

[0021] Furthermore, in step S1, the method used to select the statistical feature with the best trend and monotonicity between the gear speed signal and the vibration acceleration signal of the main test gearbox is linear regression analysis, specifically:

[0022] y i (t)=β i,0 +β i,1 t,

[0023]

[0024] Where, β i,0 ,β i,1 and y i (t) represents the intercept, slope, and corresponding statistical characteristic value of the i-th characteristic trend at time point t; It is the average time. K represents the average value of the i-th characteristic trend; K represents the length of the statistical characteristic data.

[0025] Obtain the statistical characteristics with the best trend and monotonicity from the gear speed signal and the vibration acceleration signal of the main test gearbox, which will be beneficial for subsequent use.

[0026] Furthermore, the health indicator with the best trend and monotonicity selected in S1 is the pulse factor in the speed signal, which is:

[0027]

[0028] Among them, s f x represents the impulse factor. i Let n represent the i-th sample point, and n represent the number of samples. It represents the absolute value of the mean of the sample data.

[0029] The calculation is simple and easy to use.

[0030] Furthermore, the final output of the multivariate fusion feature extractor in step S2 is:

[0031]

[0032] Among them, X n Let F represent the nth input sample. resnet (·)and These represent the ResNet branches of the residual network and their corresponding outputs, F and F respectively. bigru (·)and These represent the branches of the bidirectional gated loop unit (BIGRU) and their corresponding outputs; f nThis represents the final output of the multi-element fusion feature extractor.

[0033] Obtain fusion features for subsequent prediction.

[0034] Furthermore, the predictor in step S3 is:

[0035]

[0036] Where H represents the number of prediction heads integrated by each predictor. S represents the prediction result of the domain-specific predictor for the sample. n，j This indicates the predicted remaining service life of the gearbox gears.

[0037] The predictor has a simple structure and can generate the required prediction results.

[0038] Furthermore, the target network function is:

[0039] L = L y +k p ·L p +k w ·L w +k d ·L d ,

[0040] Where L represents the overall network optimization objective function, L y k represents the predicted total loss for the remaining useful life. p k w and k d L respectively P L W and L d The trade-off parameter; L P L represents the alignment loss at the predictor level. W L represents the Wasserstein distance loss. d This represents the loss of the domain discriminator during adversarial training.

[0041] Setting a target network function helps optimize prediction results.

[0042] Furthermore, based on the model-free adaptive control (MFAC) theory and by introducing the pseudo-partial derivative (PPD) and compact dynamic linearized CFDL model, the optimal control torque is calculated. The specific steps are as follows:

[0043] For a noiseless, ideal, single-input, single-output discrete-time nonlinear system, it can be expressed as:

[0044] h(k+1)=f(h(k),...,h(kn i ),l(k),...,l(kn j))

[0045] Where h(k) represents the system input at time k, i.e., the remaining service life prediction model based on monitoring data, and l(k) represents the system output at time k, i.e., the load on the gear during operation. i n j Let f(·) be two unknown positive numbers, and f(·) be an unknown nonlinear function.

[0046] When the function f(·) is continuously differentiable and the system satisfies the Lipschitz condition, i.e., for any k1≠k2, k1,k2≥0 and

[0047] If l(k1)≠l(k2), then:

[0048] |h(k1+1)-h(k2+1)|≤b||l(k1)-l(k2)||

[0049] Where b is a constant greater than 0;

[0050] For nonlinear systems that satisfy the above conditions, by introducing a time-varying parameter called PPD, they can be converted into CFDL models, i.e.:

[0051] h(k+1)=h(k)+φ(k)Δl(k)

[0052] Wherein, φ(k) is the time-varying parameter of PPD;

[0053] To avoid negative impacts or over-adjustment on system stability while predicting and adjusting the control input, the MFAC control input criterion function introduces a λ-weighting factor to limit the range of control input variation, thus maintaining the desired system output value. It can be expressed as:

[0054] J(l(k))=|h * (k+1)-h(k+1)| 2 +λ|l(k)-l(k-1)| 2

[0055] Where J(l(k)) represents the control input criterion function;

[0056] Substituting the linear model of the system into the criterion function and differentiating it with respect to the control input l(k), setting the derivative to zero and introducing a step size factor ρ, the final control algorithm can be obtained, namely:

[0057]

[0058] Where l(k) represents the output load at time k, ρ represents the step size factor, λ represents the weighting factor, and φ c(k) represents the time-varying parameter of PPD at time k, h * (k+1) represents the expected remaining lifetime of the system input at time k+1, and h(k) represents the predicted remaining lifetime output by the lifetime prediction model at time k.

[0059] The estimation algorithm and algorithm reset mechanism of PPD are expressed as follows:

[0060]

[0061] in, Indicates the relationship between PPD and φ at time k. c The estimated value of (k), yes The initial value of η represents the introduced step size factor, μ represents the penalty factor for the PPD estimate, Δl(k) represents the difference in output torque between time k and time k-1, Δh(k) represents the difference in remaining service life prediction between time k and time k-1, and ε is a sufficiently small difference.

[0062] Based on MFAC theory, pseudo-partial derivative (PPD) and compact dynamic linearization (CFDL) models are introduced to dynamically adjust control parameters. At the same time, weighted criterion functions, step size factors and algorithm reset mechanisms are added to the control algorithm to enhance its robustness and flexibility.

[0063] Furthermore, based on the optimal control parameters, the method for controlling the operating conditions of the gearbox gears in the vertical mill is as follows:

[0064] When the difference between the predicted remaining life of the vertical mill gearbox and the current expected remaining life exceeds 10 hours, count variables a and b are added according to the direction of the deviation:

[0065] If the predicted life of the gearbox gears of the vertical mill exceeds the current expected remaining life by more than 10 hours, the value of the count variable a increases by 1; conversely, if the predicted life of the gearbox gears of the vertical mill is less than the current expected remaining life by more than 10 hours, the value of the count variable b increases by 1.

[0066] If the cumulative value of the count variable a or b reaches or exceeds 2, the operating condition of the vertical mill gearbox is adjusted based on the optimal control operating condition parameters until the absolute error between the predicted life of the vertical mill gearbox gear and the current expected remaining life is reduced to within 10 hours. At this time, the control operation is stopped and the counts a and b are cleared to zero.

[0067] Extending the actual service life of the gears in the vertical mill gearbox to near the expected service life ensures the working efficiency and long-term stability of the gear transmission system, and reduces economic losses and safety risks caused by accidental gear damage.

[0068] The present invention also provides an online intelligent control system for the gear life of a vertical mill gearbox based on the method described in the present invention, including a data acquisition module, a LabVIEW host computer unit and an algorithm engine module;

[0069] The data acquisition module includes an encoder and a vibration sensor, which are used to acquire the gear speed signal and the vibration acceleration signal of the main test gearbox, respectively.

[0070] The LabVIEW host computer unit is used to receive the collected gear speed signal and gearbox vibration acceleration signal of the main test gearbox, and to determine whether the gear has entered the degradation stage.

[0071] The algorithm engine module is connected to the LabVIEW host computer unit. The algorithm engine module is used to predict the remaining life of the gears in the vertical mill gearbox and send it back to the LabVIEW host computer unit. The LabVIEW host computer unit is equipped with a model-free adaptive control (MFAC) module. The model-free adaptive control (MFAC) module outputs the optimal control condition parameters based on the remaining life of the gears in the vertical mill gearbox and the current expected remaining life. The LabVIEW host computer unit performs gear condition control based on the optimal control condition parameters.

[0072] This system utilizes various unit modules to achieve accurate prediction and online intelligent control of the remaining service life of the gears in the vertical mill gearbox. It successfully extends the actual service life of the gears in the vertical mill gearbox to close to the expected service life, ensuring the working efficiency and long-term stability of the gear transmission system, and reducing economic losses and safety risks caused by accidental gear damage. Attached Figure Description

[0073] Figure 1 This is a schematic diagram of the online intelligent control system for the gear life of the vertical mill gearbox of the present invention;

[0074] Figure 2 This is a schematic diagram of the process for constructing a life prediction model for the online intelligent control method for the gear life of a vertical mill gearbox according to the present invention.

[0075] Figure 3 This is a schematic diagram of the process for regulating the operating conditions of the gearbox gears in the vertical mill gearbox using the online intelligent control method for gear life of the vertical mill gearbox of the present invention. Detailed Implementation

[0076] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0077] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0078] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0079] This invention discloses an online intelligent control method for the gear life of a vertical mill gearbox. This method intelligently optimizes gear operating conditions online, effectively extending the actual service life of the gears in the vertical mill gearbox to be closer to their expected service life, thus ensuring the long-term efficient operation of the equipment. The online intelligent control method for the gear life of a vertical mill gearbox includes the following steps:

[0080] Every 5 minutes, the gear speed signal and the vibration acceleration signal of the main test gearbox are collected for 1 second (intermittently).

[0081] Based on the 3σ criterion, it is determined whether the gear has entered the degradation stage. If the gear has entered the degradation stage, a life prediction model is constructed and used to predict the remaining life of the gearbox of the vertical mill.

[0082] Combining the predicted remaining life of the gearbox of the vertical mill with the current expected remaining life, the optimal control torque is calculated based on the model-free adaptive control MFAC theory and by introducing pseudo-partial derivative PPD and compact dynamic linearized CFDL model.

[0083] The operating conditions of the gearbox gears in the vertical mill are adjusted based on the optimal control parameters.

[0084] In a preferred embodiment of the present invention, such as Figure 2 As shown, the method for constructing the lifetime prediction model is as follows:

[0085] S1. Select the statistical features with the best trend and monotonicity of the gear speed signal and the vibration acceleration signal of the main test gearbox throughout the whole life test as the health indicators of the gear. Form a health indicator sequence throughout the whole life test. Construct life samples on the health indicator sequence with a window length of 512 and a step size of 1 using the sliding window method.

[0086] S2, construct a multi-dimensional fusion feature extractor, extract multi-dimensional feature information related to the degradation state and remaining lifespan in the lifetime samples in parallel, fuse the multi-dimensional feature information, and generate the final output of the multi-dimensional fusion feature extractor;

[0087] S3, Construct a prediction ensemble module. The prediction ensemble module includes multiple domain-specific predictors. Each domain-specific predictor integrates multiple prediction heads and predicts the remaining lifetime results of the multivariate fusion features of the corresponding specific source domain samples. The remaining lifetime results of multiple prediction heads are fused through an ensemble learning strategy to generate the remaining lifetime prediction output of the gearbox gears.

[0088] S4. Design a multi-level, multi-source domain adaptive strategy to perform local and global alignment of feature distributions in the source and target domains, as well as alignment of the output boundary of the predictor, thereby reducing the distribution differences between data features in different domains and improving the domain-invariant representation ability of features.

[0089] S5, construct the target network function using multi-level multi-source domain adaptive loss and prediction losses of source domain and target domain;

[0090] The target network function is optimized using backpropagation and gradient descent algorithms to obtain the target network.

[0091] In a preferred embodiment of the present invention, the method used in step S1 to select the statistical feature with the best trend and monotonicity between the gear speed signal and the vibration acceleration signal of the main test gearbox is linear regression analysis, specifically:

[0092] y i (t)=β i,0 +β i,1 t,

[0093]

[0094] Where βi,0,βi,1 and y i (t) represents the intercept, slope, and corresponding statistical characteristic value of the i-th characteristic trend at time point t; It is the average time. K represents the average value of the i-th characteristic trend; K represents the length of the statistical characteristic data.

[0095] Preferably, the health indicator with the best trend and monotonicity selected in S1 is the pulse factor in the speed signal, which is:

[0096]

[0097] Among them, s f x represents the impulse factor. iLet n represent the i-th sample point, and n represent the number of samples. It represents the absolute value of the mean of the sample data.

[0098] In a preferred embodiment of the present invention, the final output of the multivariate fusion feature extractor in step S2 is:

[0099]

[0100] Among them, X n Let F represent the nth input sample. resnet (·)and These represent the ResNet branches of the residual network and their corresponding outputs, F and F respectively. bigru (·)and These represent the bidirectional gated recurrent unit (BIGRU) branches and their corresponding outputs, respectively; fn represents the final output of the multivariate fusion feature extractor.

[0101] More preferably, the predictor in step S3 is:

[0102]

[0103] Where H represents the number of prediction heads integrated by each predictor. Sn represents the prediction result of the domain-specific predictor for the sample. j This indicates the predicted remaining service life of the gearbox gears.

[0104] In a preferred embodiment of the present invention, the target network function is:

[0105] L = L y +k p ·L p +k w ·L w +k d ·L d ,

[0106] Where L represents the overall network optimization objective function, L y k represents the predicted total loss for the remaining useful life. p k w and k d L respectively P L W and L d The trade-off parameter; L P L represents the alignment loss at the predictor level. W L represents the Wasserstein distance loss. d This represents the loss of the domain discriminator during adversarial training.

[0107] In a preferred embodiment of the present invention, the control parameters are dynamically adjusted based on the model-free adaptive control (MFAC) theory, and pseudo-partial derivatives (PPD) and a compact dynamic linearized CFDL model are introduced to calculate the optimal control torque. The specific steps are as follows:

[0108] For a noiseless, ideal, single-input, single-output discrete-time nonlinear system, it can be expressed as:

[0109] h(k+1)=f(h(k),...,h(kn i ),l(k),...,l(kn j ))

[0110] Where h(k) represents the system input at time k, i.e., the remaining service life prediction model based on monitoring data, and l(k) represents the system output at time k, i.e., the load on the gear during operation. i n j Let f(·) be two unknown positive numbers, and f(·) be an unknown nonlinear function.

[0111] When the function f(·) is continuously differentiable and the system satisfies the Lipschitz condition, i.e., for any k1≠k2, k1,k2≥0 and

[0112] If l(k1)≠l(k2), then:

[0113] |h(k1+1)-h(k2+1)|≤b||l(k1)-l(k2)||

[0114] Where b is a constant greater than 0;

[0115] For nonlinear systems that satisfy the above conditions, by introducing a time-varying parameter called PPD, they can be converted into CFDL models, i.e.:

[0116] h(k+1)=h(k)+φ(k)Δl(k)

[0117] Wherein, φ(k) is the time-varying parameter of PPD;

[0118] To avoid negative impacts or over-adjustment on system stability while predicting and adjusting the control input, the MFAC control input criterion function introduces a λ-weighting factor to limit the range of control input variation, thus maintaining the expected value of the system output. It can be expressed as:

[0119] J(l(k))=|h * (k+1)-h(k+1)| 2 +λ|l(k)-l(k-1)| 2

[0120] Where J(l(k)) represents the control input criterion function;

[0121] Substituting the linear model of the system into the criterion function and differentiating it with respect to the control input l(k), setting the derivative to zero and introducing a step size factor ρ, the final control algorithm can be obtained, namely:

[0122]

[0123] Where l(k) represents the output load at time k, ρ represents the step size factor, λ represents the weighting factor, and φ c (k) represents the time-varying parameter of PPD at time k, h * (k+1) represents the expected remaining lifetime value of the system input at time k+1, and h(k) represents the remaining lifetime prediction value (PPD) output by the lifetime prediction model at time k. The estimation algorithm and algorithm reset mechanism are expressed as follows:

[0124]

[0125] in, Indicates the relationship between PPD and φ at time k. c The estimated value of (k), yes The initial value of η represents the introduced step size factor, μ represents the penalty factor for the PPD estimate, Δl(k) represents the difference in output torque between time k and time k-1, Δh(k) represents the difference in remaining service life prediction between time k and time k-1, and ε is a sufficiently small difference.

[0126] In a preferred embodiment of the present invention, such as Figure 3 As shown, the method for adjusting the operating conditions of the gearbox gears in a vertical mill, based on the optimal control parameters, is as follows:

[0127] When the difference between the predicted remaining life of the vertical mill gearbox and the current expected remaining life exceeds 10 hours, count variables a and b are added according to the direction of the deviation:

[0128] If the predicted life of the gearbox gears of the vertical mill exceeds the current expected remaining life by more than 10 hours, the value of the count variable a increases by 1; conversely, if the predicted life of the gearbox gears of the vertical mill is less than the current expected remaining life by more than 10 hours, the value of the count variable b increases by 1.

[0129] If the cumulative value of the count variable a or b reaches or exceeds 2, the operating condition of the vertical mill gearbox is adjusted based on the optimal control operating condition parameters until the absolute error between the predicted life of the vertical mill gearbox gear and the current expected remaining life is reduced to within 10 hours. At this time, the control operation is stopped and the counts a and b are cleared to zero.

[0130] This invention also provides an online intelligent control system for the gear life of a vertical mill gearbox based on the method described in this invention, such as... Figure 1 As shown, it includes a data acquisition module, a LabVIEW host computer unit, and an algorithm engine module. The vertical mill gearbox includes a drive motor, a main test gearbox, a secondary test gearbox, and a load motor. A control cabinet (such as a Siemens S7-1200 PLC) is set up to control the drive motor and the load motor. A thin oil station is set up to lubricate the main test gearbox and the secondary test gearbox through two independent oil circuits.

[0131] The data acquisition module includes an encoder and a vibration sensor, which are used to acquire the gear speed signal and the vibration acceleration signal of the main test gearbox, respectively.

[0132] The LabVIEW host computer unit operates based on the DAQmx driver data acquisition module provided by NI, receiving and storing raw data from the encoder and vibration sensor in real time, and monitoring whether the gear has entered the degradation stage in real time based on the 3σ criterion.

[0133] The LabVIEW host computer receives and stores raw data in real time via a data acquisition card, then packages and sends it to the algorithm engine module deployed on the local server. If degradation is detected, the LabVIEW host computer will package the stored data into JSON format via HTTP and send it to the algorithm engine module deployed on the local server.

[0134] The algorithm engine module is a self-developed algorithm engine module that can run on a local server. This module allows the embedding of Python-based algorithm models and provides a variety of input and output interfaces to enable real-time communication with various data acquisition software.

[0135] The algorithm engine module is electrically connected to the LabVIEW host computer unit. The algorithm engine module predicts the remaining life of the gears in the vertical mill gearbox and transmits this information back to the LabVIEW host computer unit via HTTP communication protocol. The LabVIEW host computer unit includes a Model-Free Adaptive Control (MFAC) module. This MFAC module uses MATLAB script nodes within LabVIEW to output optimal control parameters based on the remaining life of the gears in the vertical mill gearbox and the current expected remaining life. The LabVIEW host computer unit then adjusts the gear operating conditions based on these optimal control parameters.

[0136] After receiving the remaining service life result and expected remaining service life from the algorithm engine, the MFAC module outputs the optimal operating parameters of the vertical mill gearbox gears under the current actual remaining service life through the control algorithm. Furthermore, considering the stability issue of the service life prediction algorithm, a secondary judgment mechanism is introduced to further optimize the control algorithm and ensure system stability.

[0137] After the LabVIEW host computer unit obtains the control parameters output by the MFAC module, it communicates with the programmable logic controller (PLC) module in the control cabinet via Ethernet cable to realize real-time control of the gearbox gear running status.

[0138] This system utilizes various unit modules to achieve accurate prediction and online intelligent control of the remaining service life of the gears in the vertical mill gearbox. It successfully extends the actual service life of the gears in the vertical mill gearbox to close to the expected service life, ensuring the working efficiency and long-term stability of the gear transmission system, and reducing economic losses and safety risks caused by accidental gear damage.

[0139] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0140] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for online intelligent control of gear life in a vertical mill gearbox, characterized in that, Includes the following steps: Intermittently acquire gear speed signals and gearbox vibration acceleration signals of the main test gearbox; Based on the 3σ criterion, it is determined whether the gear has entered the degradation stage. If the gear has entered the degradation stage, a life prediction model is constructed and used to predict the remaining life of the gearbox of the vertical mill. Combining the predicted remaining life of the gearbox of the vertical mill with the current expected remaining life, the optimal control torque is calculated based on the model-free adaptive control MFAC theory and by introducing pseudo-partial derivative PPD and compact dynamic linearized CFDL model. The operating conditions of the gearbox gears in the vertical mill are adjusted based on the optimal control parameters.

2. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 1, characterized in that, The method for constructing a lifetime prediction model is as follows: S1. Select the statistical features with the best trend and monotonicity of the gear speed signal and the vibration acceleration signal of the main test gearbox throughout the whole life test as the health indicators of the gear. Form a health indicator sequence throughout the whole life test. Construct life samples on the health indicator sequence with a window length of 512 and a step size of 1 using the sliding window method. S2, construct a multi-dimensional fusion feature extractor, extract multi-dimensional feature information related to the degradation state and remaining lifespan in the lifetime samples in parallel, fuse the multi-dimensional feature information, and generate the final output of the multi-dimensional fusion feature extractor; S3, Construct a prediction ensemble module. The prediction ensemble module includes multiple domain-specific predictors. Each domain-specific predictor integrates multiple prediction heads and predicts the remaining lifetime results of the multivariate fusion features of the corresponding specific source domain samples. The remaining lifetime results of multiple prediction heads are fused through an ensemble learning strategy to generate the remaining lifetime prediction output of the gearbox gears. S4. Design a multi-level, multi-source domain adaptive strategy to perform local and global alignment of feature distributions in the source and target domains, as well as alignment of the output boundary of the predictor, thereby reducing the distribution differences between data features in different domains and improving the domain-invariant representation ability of features. S5, construct the target network function using multi-level multi-source domain adaptive loss and prediction losses of source domain and target domain; The target network function is optimized using backpropagation and gradient descent algorithms to obtain the target network.

3. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 2, characterized in that, In step S1, the method used to select the statistical feature with the best trend and monotonicity between the gear speed signal and the vibration acceleration signal of the main test gearbox is linear regression analysis, specifically: y i (t)=β i,0 +b i,1 t, Where, β i,0 ,β i,1 and y i (t) represents the intercept, slope, and corresponding statistical characteristic value of the i-th characteristic trend at time point t; It is the average time. K represents the average value of the i-th characteristic trend; K represents the length of the statistical characteristic data.

4. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 2, characterized in that, The health indicator selected in S1 is the pulse factor in the speed signal, which is: Among them, s f x represents the impulse factor. i Let n represent the i-th sample point, and n represent the number of samples. It represents the absolute value of the mean of the sample data.

5. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 2, characterized in that, The final output of the multivariate fusion feature extractor in step S2 is: Among them, X n Let F represent the nth input sample. resnet (·)and These represent the ResNet branches of the residual network and their corresponding outputs, F and F respectively. bigru (·)and These represent the branches of the bidirectional gated loop unit (BIGRU) and their corresponding outputs; f n This represents the final output of the multi-element fusion feature extractor.

6. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 2, characterized in that, The predictor in step S3 is: Where H represents the number of prediction heads integrated by each predictor. S represents the prediction result of the domain-specific predictor for the sample. n,j This indicates the predicted remaining service life of the gearbox gears.

7. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 2, characterized in that, The target network function is: L=L y +k p ·L p +k w ·L w +k d ·L d , Where L represents the overall network optimization objective function, L y k represents the predicted total loss for the remaining useful life. p k w and k d L respectively P L W and L d The trade-off parameter; L P L represents the alignment loss at the predictor level. W L represents the Wasserstein distance loss. d This represents the loss of the domain discriminator during adversarial training.

8. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 1, characterized in that, Based on the model-free adaptive control (MFAC) theory and by introducing pseudo-partial derivative (PPD) and compact dynamic linearization (CFDL) models, the optimal control torque is calculated. The specific steps are as follows: For a noiseless, ideal, single-input, single-output discrete-time nonlinear system, it can be expressed as: h(k+1)=f(h(k),...,h(k-n i ),l(k),...,l(k-n j )) Where h(k) represents the system input at time k, i.e., the remaining service life prediction model based on monitoring data, and l(k) represents the system output at time k, i.e., the load on the gear during operation. i n j Let f(·) be two unknown positive numbers, and f(·) be an unknown nonlinear function. When the function f(·) is continuously differentiable and the system satisfies the Lipschitz condition, i.e., for any k1≠k2, k1,k2≥0 and If l(k1)≠l(k2), then: |h(k1+1)-h(k2+1)|≤b||l(k1)-l(k2)|| Where b is a constant greater than 0; For a nonlinear system that meets the conditions, it can be transformed into a CFDL model by introducing a time-varying parameter called PPD, i.e.: h(k+1)=h(k)+φ(k)Δl(k) Wherein, φ(k) is the time-varying parameter of PPD; To avoid negative impacts or over-adjustment on system stability while predicting and adjusting the control input, the MFAC control input criterion function introduces a λ-weighting factor to limit the range of control input variation, thus maintaining the expected value of the system output. This factor is expressed as: J(l(k))=|h * (k+1)-h(k+1)| 2 +λ|l(k)-l(k-1)| 2 Where J(l(k)) represents the control input criterion function; Substituting the linear model of the system into the criterion function and differentiating it with respect to the control input l(k), setting the derivative to zero and introducing a step size factor ρ, yields the final control algorithm, namely: Where l(k) represents the output load at time k, ρ represents the step size factor, λ represents the weighting factor, and φ c (k) represents the time-varying parameter of PPD at time k, h * (k+1) represents the expected remaining lifetime value of the system input at time k+1, and h(k) represents the remaining lifetime prediction value output by the lifetime prediction model at time k. The estimation algorithm and algorithm reset mechanism of PPD are expressed as follows: in, Indicates the relationship between PPD and φ at time k. c The estimated value of (k), yes The initial value of η represents the introduced step size factor, μ represents the penalty factor for the PPD estimate, Δl(k) represents the difference in output torque between time k and time k-1, Δh(k) represents the difference in remaining service life prediction between time k and time k-1, and ε is a sufficiently small difference.

9. The online intelligent control method for gear life of a vertical mill gearbox as described in claim 1, characterized in that, The method for regulating the operating conditions of the gearbox gears in a vertical mill, based on the optimal control parameters, is as follows: When the difference between the predicted remaining life of the vertical mill gearbox and the current expected remaining life exceeds 10 hours, count variables a and b are added according to the direction of the deviation: If the predicted life of the gearbox gears of the vertical mill exceeds the current expected remaining life by more than 10 hours, the value of the count variable a increases by 1; conversely, if the predicted life of the gearbox gears of the vertical mill is less than the current expected remaining life by more than 10 hours, the value of the count variable b increases by 1. If the cumulative value of the count variable a or b reaches or exceeds 2, the operating condition of the vertical mill gearbox is adjusted based on the optimal control operating condition parameters until the absolute error between the predicted life of the vertical mill gearbox gear and the current expected remaining life is reduced to within 10 hours. At this time, the control operation is stopped and the counts a and b are cleared to zero.

10. An online intelligent control system for the gear life of a vertical mill gearbox based on the method of any one of claims 1-9, characterized in that, It includes a data acquisition module, a LabVIEW host computer unit, and an algorithm engine module; The data acquisition module includes an encoder and a vibration sensor, which are used to acquire the gear speed signal and the vibration acceleration signal of the main test gearbox, respectively. The LabVIEW host computer unit is used to receive the collected gear speed signal and gearbox vibration acceleration signal of the main test gearbox, and to determine whether the gear has entered the degradation stage. The algorithm engine module is connected to the LabVIEW host computer unit. The algorithm engine module is used to predict the remaining life of the gears in the vertical mill gearbox and send it back to the LabVIEW host computer unit. The LabVIEW host computer unit is equipped with a model-free adaptive control (MFAC) module. The model-free adaptive control (MFAC) module outputs the optimal control condition parameters based on the remaining life of the gears in the vertical mill gearbox and the current expected remaining life. The LabVIEW host computer unit performs gear condition control based on the optimal control condition parameters.

Images