Axis straightening method based on empirical value learning

Abstract

The invention discloses a shaft straightening method based on empirical value learning, which adopts a computer machine learning algorithm to optimize a straightening strategy, full-automatic straightening equipment comprises a casting horizontal measuring table and a gantry pressing device, one end of the equipment is provided with a pinch roller rotating mechanism, and the other end of the equipment is provided with a driven pinch roller. Firstly, a to-be-straightened shaft is placed on a V-shaped lifting sliding table of a casting platform, a single-side pressing wheel mechanism drives a rotating shaft to operate in a servo mode, meanwhile, the radial run-out value of the shaft at each distance is measured, the shaft bending degree is determined, the shaft adjusting sequence is generally from the middle of the shaft, the shaft is adjusted rightwards firstly and then adjusted leftwards from the middle, and after the requirement is basically met, the two ends of the shaft are adjusted; rough adjustment and straightening are conducted firstly, fine adjustment and straightening are conducted finally, and adjustment is conducted from one end of the shaft to the other end of the shaft. Through machine learning, the planning efficiency of the straightening sequence can be improved, and the overall performance of the straightening process can be improved. Through continuous learning and adaptation, the machine learning model can bring significant improvement and optimization to the straightening strategy planning stage.

Description

Technical Field

[0001] The invention belongs to the technical field of shaft alignment, and in particular relates to a shaft alignment method based on experience value learning. Background Art

[0002] During the forging, turning, heat treatment and other processes in the manufacturing process of shaft parts, factors such as uneven force or cutting heat may lead to uneven internal stress distribution, thus causing deformation of the shaft. As modern mechanical equipment develops towards high speed, precision and heavy load, the accuracy requirements for shaft parts are getting higher and higher, and shaft straightening technology needs to be continuously improved to meet this demand. Solving the geometric shape error problems such as bending and twisting of shaft parts caused by various reasons is a necessary repair process technology in the field of mechanical engineering and manufacturing. As a key rotating component in mechanical equipment, the quality indicators of shaft parts such as straightness, cylindricity and dynamic balance directly affect the working performance and service life of the entire equipment.

[0003] In the production process of shafts for submersible electric pumps, shaft parts with excessive runout after processing usually need to be straightened. The traditional shaft straightening method requires first measuring the high point of the shaft runout on the deflection meter, then placing the shaft on the press for straightening, and after straightening, placing the shaft on the deflection meter again to measure the runout value, and repeating this operation until the runout values ​​at all positions of the shaft meet the specified requirements. The above traditional straightening method requires moving the shaft back and forth between the deflection meter and the press, which is very cumbersome, and when the shaft parts are heavy, it is necessary to frequently use lifting equipment, which is time-consuming and labor-intensive, and requires high labor intensity. At the same time, it requires skilled operators and is highly dependent on personnel. Repeated operations can easily lead to incomplete data recording and data consistency during the straightening process, which invisibly increases labor intensity, training costs, and the complexity of human resource management. Summary of the invention

[0004] The problem to be solved by the present invention is to provide a shaft straightening method based on experience value learning, which ensures high precision and consistency in the straightening process through a system automation process, reduces the time and effort of manual operation, and thus significantly improves production efficiency.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a shaft straightening method based on experience value learning, comprising the following steps:

[0006] S1: Place the axis to be straightened on the straightening machine, and the servo drive of the rotating axis by the single-side pressure wheel mechanism;

[0007] S2: measuring the radial runout value of each distance of the shaft to be straightened to determine the bending degree of the shaft to be straightened;

[0008] S3: adjusting and straightening the axis to be straightened.

[0009] Further, the S2 comprises the following steps:

[0010] S21: measuring the radial runout value of the circumference of the axis to be straightened at every 300 mm spacing position, and the machine learning algorithm system marks and records each rotation high point, and records the data value of each high point at the same time;

[0011] S22: After the full-length inspection, the machine learning algorithm system automatically calculates whether the bending points of the shaft to be straightened are in a plane;

[0012] S23: When the bending points of the shaft to be straightened are not in the same plane, the machine learning algorithm system makes a special mark at this point, and after checking the entire length, the special point is counted.

[0013] Further, the S3 comprises the following steps:

[0014] S31: straightening the middle section of the shaft to be straightened;

[0015] S32: Straightening the two end sections of the shaft to be straightened.

[0016] Further, the S31 includes the following steps:

[0017] S311: aligning the special point marked on the shaft to be straightened toward the center of the shaft to be straightened, and after the special point is straightened, aligning from the middle of the shaft to be straightened toward the right end of the shaft to be straightened;

[0018] S312: The machine learning algorithm system calculates the bending degree of the shaft to be straightened, and automatically moves the shaft support of the straightening machine to the middle of two consecutive high points;

[0019] S313: moving the clamping device of the straightening machine to the high point of the shaft to be straightened, determining the exact position of the highest point, the shaft to be straightened automatically rotates to the point where the raised part faces upward, the clamping device presses on this point, and applies pressure according to the bending condition of the shaft to be straightened;

[0020] S314: Straightening is performed repeatedly to complete the straightening of all high points so that the runout of the axis to be straightened meets the requirements of the drawing.

[0021] Furthermore, in S313, after the pressure is applied, the pressure is maintained for 3-5 minutes, and after the axis to be straightened is calibrated to the required reading, pressure is applied again to bend it by 0.05-0.07 mm.

[0022] Further, the S32 includes the following steps:

[0023] S321: The machine learning algorithm system finds the lowest point of the shaft end of the shaft to be straightened and marks it;

[0024] S322: The low point of the axis to be straightened faces upwards, and the two movable bearings of the straightening machine support it, and the clamping device of the straightening machine applies force to the centers of the two bearings;

[0025] S323: Measure after pressing, and repeat this process until the shaft head runout meets the requirements of the drawing.

[0026] Furthermore, the machine learning algorithm system includes a historical data analysis unit, a prediction model building unit, an optimization algorithm application unit, a real-time feedback adjustment unit, and a multi-objective optimization unit.

[0027] Furthermore, the present invention also provides a device for running the above data processing method.

[0028] Furthermore, the present invention also provides a device, comprising a memory, a processor, and an algorithm stored in the memory and executable on the processor, wherein the processor implements the above-mentioned data processing method when executing the computer program.

[0029] Furthermore, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer algorithm, and the computer algorithm implements the above-mentioned data processing when executed by a processor.

[0030] The advantages and positive effects of the present invention are:

[0031] The present invention reduces the time and effort of manual operation through the system automation process, realizes fast and continuous straightening operation, and thus significantly improves production efficiency. Accurate sensors and sophisticated control systems reduce human errors and ensure high accuracy and consistency during the straightening process. The automated system can optimize the straightening strategy based on historical data and machine learning algorithms to improve the quality and efficiency of straightening. The automated system can automatically record the straightening data for subsequent analysis and quality improvement.

[0032] Through these methods, machine learning can not only improve the planning efficiency of the straightening sequence, but also improve the overall performance of the straightening process. Through continuous learning and adaptation, machine learning models can bring significant improvements and optimization to the straightening strategy planning stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is an overall flow chart of an embodiment of the present invention.

[0034] Figure 2 This is a diagram showing the position of the shaft support during shaft straightening according to an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0037] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0038] The embodiments of the present invention are further described below in conjunction with the accompanying drawings:

[0039] like Figure 1 As shown, a shaft straightening method based on experience value learning includes the following steps:

[0040] S1: Place the axis to be straightened on the straightening machine. Specifically, the straightening machine uses a computer machine learning algorithm to optimize the straightening strategy of the fully automatic straightening equipment. The straightening machine has a casting horizontal measuring table and a gantry clamping device. There is a pressure wheel rotating mechanism at one end of the straightening machine, and a pressure wheel driven mechanism at the other end. When in use, first place the axis to be straightened on the V-shaped lifting slide of the casting platform. The unilateral pressure wheel mechanism servo drives the rotating shaft to run, and at the same time measures the radial runout value of each distance of the axis to be straightened. Specifically, the sensor of this embodiment measures the radial runout value of the circumference of each 300mm spacing position of the axis to be straightened, and records the rotation angle corresponding to each runout value to determine the degree of bending of the axis to be straightened. The order of adjusting the axis is generally to start from the middle of the axis, first adjust to the right, and then adjust the axis from the middle to the left. After basically meeting the requirements, adjust the two ends of the axis, first coarsely adjust and straighten, and finally finely adjust and straighten, from one end of the axis to the other end of the axis.

[0041] S2: Measure the radial runout value of each distance of the shaft to be straightened to determine the bending degree of the shaft to be straightened. Specifically, S2 includes the following steps:

[0042] S21: Automatically measure the runout of the axis to be straightened and record the rotation angle corresponding to each runout value. The machine learning algorithm system marks and records each high point of rotation and records the data value of each high point.

[0043] S22: After the full-length inspection, the machine learning algorithm system automatically calculates whether the bending points of the shaft to be straightened are in the same plane.

[0044] S23: When the bending points of the shaft to be straightened are not in the same plane, the machine learning algorithm system makes a special mark at this point, and after checking the entire length, the special points are counted.

[0045] S3: Adjust and straighten the shaft to be straightened. During straightening, after the press reaches above the pressure point, the pressure wheel mechanism rotates the shaft to rotate the highest point of the runout value to just above it, and then moves the shaft support to the positions on both sides of the lower pressure point. The downward pressure and downward pressure holding time of the press are determined by previous computer historical data. The computer algorithm is supported by a large amount of basic data, and records the radial runout values ​​of the points before and after straightening corresponding to the downward pressure of different series of products. In the subsequent straightening process, the data of the entire shaft is first measured and recorded. The computer system automatically matches and searches for the corresponding radial runout value data before straightening according to the algorithm, and automatically selects the corresponding automatic downward pressure and downward pressure holding time for the high point of the shaft. Specifically, S3 includes the following steps,

[0046] S31: Straightening the middle section of the shaft to be straightened. Specifically, S31 includes the following steps:

[0047] S311: The special points marked on the shaft to be straightened are all aligned to the center of the shaft to be straightened. After the special points are aligned, the alignment is then performed from the middle of the shaft to be straightened to the right end of the shaft to be straightened.

[0048] S312: The machine learning algorithm system calculates the bending degree of the shaft to be straightened and automatically moves the shaft support of the straightening machine to the middle of two consecutive high points, such as Figure 2 shown.

[0049] S313: Move the clamping device of the straightening machine to the high point of the shaft to be straightened, determine the exact position of the highest point, and the shaft to be straightened automatically rotates to the raised part facing upwards, and the clamping device is pressed to this point, and pressure is applied according to the bending of the shaft to be straightened. Specifically, the clamping device should not be released immediately, and it is best to maintain pressure for 3-5 minutes. Considering that the shaft itself has rebound, the greater the shaft hardness, the greater the rebound. Generally, the shaft rebound is within the range of 0.05-0.075mm. After the shaft is calibrated until the required reading is reached, add (0.05-0.07mm) and repeat the straightening process until all high points are straightened.

[0050] S314: During the process of straightening the shaft, the bending of the shaft at the fulcrum should be measured repeatedly. If there is bending, the shaft should be straightened in the same way. Measure once after each pressing until the runout of the shaft body meets the requirements of the drawing.

[0051] S32: Straighten the two end sections of the shaft to be straightened. The bending of the two ends of the shaft must be performed after the entire shaft is straightened. Specifically, S32 includes the following steps:

[0052] S321: The machine learning algorithm system finds the low point of the shaft end to be straightened and marks it;

[0053] S322: The low point of the shaft to be straightened faces upwards, and is supported by two movable bearings of the straightening machine. The clamping device of the straightening machine applies force to the center of the two bearings. Specifically, the first bearing is 75mm away from the end, and the rear bearing is placed 400mm away from the end.

[0054] S323: Measure once after each pressing until the shaft head runout meets the requirements of the drawing.

[0055] Preferably, the machine learning algorithm system has the function of straightening strategy planning optimization. Specifically, it includes a historical data analysis unit, a prediction model building unit, an optimization algorithm application unit, a real-time feedback adjustment unit, and a multi-objective optimization unit, wherein:

[0056] Historical data analysis: Using historical alignment data, the machine learning model can analyze the best alignment sequence for different shaft types, sizes, and deviations. The model can identify which shaft characteristics are associated with alignment difficulty and success rate, thus providing a reference for the sequence arrangement of new alignment tasks.

[0057] Predictive model building: By training machine learning models, such as decision trees, random forests, or neural networks, the output results under different straightening sequences can be predicted. The model can predict the best straightening sequence based on the input axis parameters to minimize the straightening time and maximize the straightening quality.

[0058] Application of optimization algorithms: Use optimization algorithms (such as genetic algorithms, simulated annealing or ant colony algorithms) to search for the optimal straightening sequence. These algorithms are able to find the best solution among a large number of possible combinations, taking into account the various constraints and objectives of the straightening process.

[0059] Real-time feedback adjustment: During the straightening process, the machine learning model can dynamically adjust the straightening order based on real-time feedback data. If the straightening effect of a certain step is not as good as expected, the model can adjust the order of subsequent steps in time to compensate for the deviation.

[0060] Multi-objective optimization: The optimization of the straightening sequence is not only to reduce time, but also to consider multiple objectives such as cost, energy consumption and quality. The machine learning model can find a balance between these objectives and achieve multi-objective optimization.

[0061] The advantages and positive effects of the present invention are:

[0062] The present invention reduces the time and effort of manual operation through the system automation process, realizes fast and continuous straightening operation, and thus significantly improves production efficiency. Accurate sensors and sophisticated control systems reduce human errors and ensure high accuracy and consistency during the straightening process. The automated system can optimize the straightening strategy based on historical data and machine learning algorithms to improve the quality and efficiency of straightening. The automated system can automatically record the straightening data for subsequent analysis and quality improvement.

[0063] Through these methods, machine learning can not only improve the planning efficiency of the straightening sequence, but also improve the overall performance of the straightening process. Through continuous learning and adaptation, machine learning models can bring significant improvements and optimization to the straightening strategy planning stage.

[0064] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A shaft straightening method based on experience value learning, characterized in that: The following steps are included: S1: Place the axis to be straightened on the straightening machine, and the servo drive of the rotating axis by the single-side pressure wheel mechanism; S2: measuring the radial runout value of each distance of the shaft to be straightened to determine the bending degree of the shaft to be straightened; S3: adjusting and straightening the axis to be straightened.

2. The shaft straightening method based on experience value learning according to claim 1, characterized in that: The S2 comprises the following steps, S21: measuring the radial runout value of the circumference of the axis to be straightened at every 300 mm spacing position, and the machine learning algorithm system marks and records each rotation high point, and records the data value of each high point at the same time; S22: After the full-length inspection, the machine learning algorithm system automatically calculates whether the bending points of the shaft to be straightened are in a plane; S23: When the bending points of the shaft to be straightened are not in the same plane, the machine learning algorithm system makes a special mark at this point, and after checking the entire length, the special point is counted.

3. The shaft straightening method based on experience value learning according to claim 2, characterized in that: The S3 comprises the following steps, S31: straightening the middle section of the shaft to be straightened; S32: Straightening the two end sections of the shaft to be straightened.

4. The shaft straightening method based on experience value learning according to claim 3 is characterized in that: The S31 The following steps are included: S311: aligning the special point marked on the shaft to be straightened toward the center of the shaft to be straightened, and after the special point is straightened, aligning from the middle of the shaft to be straightened toward the right end of the shaft to be straightened; S312: The machine learning algorithm system calculates the bending degree of the shaft to be straightened, and automatically moves the shaft support of the straightening machine to the middle of two consecutive high points; S313: moving the clamping device of the straightening machine to the high point of the shaft to be straightened, determining the exact position of the highest point, the shaft to be straightened automatically rotates to the point where the raised part faces upward, the clamping device presses on this point, and applies pressure according to the bending condition of the shaft to be straightened; S314: Straightening is performed repeatedly to complete the straightening of all high points so that the runout of the axis to be straightened meets the requirements of the drawing.

5. The shaft straightening method based on experience value learning according to claim 4, characterized in that: In S313, after the pressure is applied, the pressure is maintained for 3-5 minutes, and after the axis to be straightened is calibrated to the required reading, pressure is applied again to bend it by 0.05-0.07 mm.

6. A shaft straightening method based on experience value learning according to any one of claims 3 to 5, characterized in that: The S32 The following steps are included: S321: The machine learning algorithm system finds the lowest point of the shaft end of the shaft to be straightened and marks it; S322: The low point of the axis to be straightened faces upwards, and the two movable bearings of the straightening machine support it, and the clamping device of the straightening machine applies force to the centers of the two bearings; S323: Measure after pressing, and repeat this process until the shaft head runout meets the requirements of the drawing.

7. A shaft straightening method based on experience value learning according to any one of claims 2 to 5, characterized in that: The machine learning algorithm system includes a historical data analysis unit, a prediction model building unit, an optimization algorithm application unit, a real-time feedback adjustment unit, and a multi-objective optimization unit.

8. A device, characterized in that: Run the data processing method according to any one of claims 1 to 7.

9. A device comprising a memory, a processor, and an algorithm stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the data processing method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer algorithm, characterized in that: When the computer algorithm is executed by a processor, the data processing according to any one of claims 1 to 7 is implemented.

Images