Multi-zone temperature balance regulation and control method for carbonitriding heat treatment of bearing

Abstract

The invention relates to the technical field of temperature balance regulation and control, in particular to a multi-zone temperature balance regulation and control method for carbonitriding heat treatment of a bearing, which comprises the following steps of: sequentially calculating the temperature deviation between a corresponding temperature threshold and an actual temperature under the same heating zone and the deviation change rate; matching the temperature deviation, and a fuzzy state vector and a membership degree vector corresponding to the deviation change rate; integrating the temperature deviation and the deviation change rate corresponding fuzzy state vector under each heating area into a fuzzy state combination, and matching a final activation rule corresponding to each heating area in the control rule base; integrating the fuzzy state vector and the membership degree vector of the final activation rule in each heating area in the step S2 to obtain a fuzzy set; defuzzifying the fuzzy set into a correction coefficient by adopting a centroid method; and adjusting the reference coefficient by adopting the correction coefficient to obtain a dynamic control parameter.

Description

Technical Field

[0001] This invention relates to the field of temperature equalization control technology, and more specifically, to a multi-zone temperature equalization control method for carbonitriding heat treatment of bearings. Background Technology

[0002] Bearings, as key components of mechanical equipment, are widely used in automobile manufacturing, aerospace and other fields. They play a core supporting and transmission role in equipment such as electronically controlled compressors, and need to withstand complex working conditions such as pressure fluctuations, axial forces and impact loads. Therefore, in order to avoid failure problems such as contact fatigue spalling, friction wear, corrosion and dimensional deformation of bearings under complex service conditions such as high speed and heavy load, the bearing carbonitriding heat treatment process is used to modify and strengthen the bearing surface. By co-diffusing carbon and nitrogen elements at high temperature to form a composite diffusion layer, it combines the advantages of deep carburizing layer and high hardness and corrosion resistance of nitriding. It can significantly improve the surface hardness of bearings (up to HV2000 and above), wear resistance and contact fatigue life (3-8 times higher than conventional heat-treated samples). Moreover, the diffusion layer has a gentle gradient and high residual compressive stress, which can delay crack propagation. The low treatment temperature can also reduce deformation and ensure dimensional accuracy and core toughness. During the carbonitriding process, the furnace is divided into multiple independent heating zones, each equipped with a dedicated heating element to achieve differentiated temperature control, adapting to the temperature requirements of the corresponding stage. Each heating zone is equipped with 3-5 high-precision thermocouples with an accuracy of ±0.5℃ for real-time temperature measurement, used to capture temperature fluctuations in each heating zone and accurately report temperature deviations. Based on the preset temperature thresholds corresponding to the carbonitriding process of each heating zone, combined with the real-time temperature detection and temperature threshold deviation, it is determined whether the heating power or circulating fan speed of the corresponding heating zone needs to be adjusted, thereby controlling the overall temperature inside the furnace and keeping the thermal deformation of the bearing rings within 0.05mm. This avoids local overheating / underheating deformation caused by uneven temperature, ensuring uniform bearing diffusion layer structure, reasonable hardness gradient, and stable assembly clearance with the electronically controlled compressor. However, during the carbonitriding heat treatment of bearings, the inherent heat transfer delay characteristics of the furnace structure, differences in the distribution of heating elements, uneven flow of the atmosphere inside the furnace, and differences in the placement of parts can all contribute to temperature deviations in different heating zones within the furnace. Although basic temperature anomalies can be identified and initially controlled by setting preset temperature thresholds, this method only allows for discrete "meeting / not meeting standards" judgments. It cannot accurately capture the gradual trend of temperature deviations or the coordinated fluctuation patterns across multiple zones, nor can it provide fine-grained compensation for temperature deviations. Therefore, if dynamic and balanced temperature control across multiple zones cannot be achieved during the carbonitriding process of bearings, a series of process problems will arise, as follows: On the one hand, uneven temperature can lead to inconsistent diffusion depth and uneven distribution of carbon and nitrogen concentration in bearing parts, resulting in defects such as non-martensitic structure and coarse network carbides on the surface, which significantly reduces the core mechanical properties of the parts, such as surface hardness, wear resistance and fatigue resistance. On the other hand, local overheating or underheating can significantly increase the thermal deformation of the bearing rings. In order to ensure the assembly accuracy of the parts, it is necessary to increase the subsequent grinding process, which will not only increase production energy consumption and processing costs, but may also damage the integrity of the surface structure of the parts due to excessive grinding, affecting the dimensional stability and service life of the parts. In view of this, we propose a multi-zone temperature equalization control method for carbonitriding heat treatment of bearings. Summary of the Invention

[0003] The purpose of this invention is to solve the process control problems in the traditional bearing carbonitriding heat treatment process, such as the inability of single-zone unified temperature control or conventional zone temperature control to adapt to multi-stage temperature requirements and accurately compensate for temperature deviations in multiple zones, which leads to uneven furnace temperature, inconsistent bearing diffusion layer depth and carbonitriding concentration distribution, excessive thermal deformation of bearing rings, decreased mechanical properties of parts, and increased processing costs.

[0004] To achieve the above objectives, the present invention provides a multi-zone temperature equalization control method for carbonitriding heat treatment of bearings, comprising the following steps: S1. Calculate the temperature deviation between the corresponding temperature threshold and the actual temperature in the same heating zone, and the rate of change of the deviation. S2. Preset the fuzzy universe of discourse corresponding to the temperature deviation and the rate of change of the deviation, divide each fuzzy universe of discourse into multiple fuzzy subsets including the universe values, and adjacent fuzzy subsets include partially overlapping universe values. Each fuzzy subset corresponds to a fuzzy state vector. The temperature deviation and the rate of change of deviation are converted into values ​​of the fuzzy universe of discourse. Based on the values ​​of the universe of discourse, the fuzzy subsets corresponding to the temperature deviation and the rate of change of deviation, as well as the fuzzy state vectors corresponding to the fuzzy subsets, are matched. After matching is completed, the membership degree between temperature deviation, deviation change rate and corresponding fuzzy subset is calculated to obtain the membership degree vector; S3. A control rule library containing multiple fuzzy state combinations and corresponding activation rules for the fuzzy state combinations is pre-set; the fuzzy state vectors corresponding to temperature deviation and deviation change rate in step S2 are integrated into fuzzy state combinations and matched with the corresponding final activation rules in the control rule library. If the temperature deviation and the rate of change of deviation correspond to multiple fuzzy state vectors, then the implication operator and the maximum operator are used to calculate the output intensity of each activation rule, and the final activation rule is selected. S4. Integrate the fuzzy state vector and membership vector from step S2 to obtain a fuzzy set; use the centroid method to defuzzify the fuzzy set as correction coefficients; and use the correction coefficients to adjust the benchmark coefficients to obtain dynamic control parameters.

[0005] As a further improvement to this technical solution, in step S1, the temperature deviation is the difference between the temperature threshold and the actual temperature; the deviation change rate is the ratio of the change in temperature deviation between two adjacent sampling times to the time interval between sampling times.

[0006] As a further improvement to this technical solution, the fuzzy universe structure corresponding to the temperature deviation and the rate of change of deviation in step S2 is completely consistent, and both are composed of multiple symmetrically distributed discrete universe values. When dividing the fuzzy universe into multiple fuzzy subsets in step S2, each fuzzy subset covers a specific interval within the fuzzy universe, forming a set of fuzzy subsets. After dividing the fuzzy subsets, multiple fuzzy subsets can cover the entire fuzzy universe without omission. Furthermore, each fuzzy subset represents a different fuzzy state, which is the fuzzy state vector corresponding to the fuzzy subset.

[0007] As a further improvement to this technical solution, in the process of dividing the fuzzy universe into multiple fuzzy subsets in step S2, there is partial overlap between adjacent fuzzy subsets. Specifically, a specific interval of adjacent fuzzy subsets shares at least one universe value.

[0008] As a further improvement to this technical solution, the universe of discourse values ​​corresponding to the temperature deviation and the rate of change of deviation in step S2 are: Calculate the quantization factor, and then convert the temperature deviation and the rate of change of deviation into universe values ​​within the fuzzy universe of discourse using the quantization factor: ; ; in: This is the temperature deviation quantification factor. The quantification factor for the rate of change of deviation. For the one-sided maximum value of the fuzzy universe of discourse, This represents the upper limit of the temperature deviation range. This represents the upper limit of the deviation rate range. Temperature deviation at sampling time k The corresponding universe value, The rate of change of the deviation at the k-th sampling time The corresponding universe value, , These are the quantized scaled values ​​of temperature deviation and deviation change rate, respectively, after being scaled by the corresponding quantization factors. This is the rounding function.

[0009] As a further improvement to this technical solution, step S2 sequentially uses the membership function to map the quantized scaling values ​​of temperature deviation and deviation change rate to the membership degrees of the fuzzy subsets corresponding to temperature deviation and deviation change rate, forming the membership vectors corresponding to temperature deviation and deviation change rate, specifically: Determine whether the temperature deviation and the rate of change of deviation correspond to the interval of a certain fuzzy subset. If so, retrieve the feature points corresponding to the fuzzy subset and use a piecewise formula to calculate the membership degree between the temperature deviation, the rate of change of deviation, and the corresponding fuzzy subset. If the temperature deviation and the rate of change of deviation correspond to the interval of the fuzzy subset, then... , bring up the blur sub The corresponding left endpoint a and right endpoint b of the universe of discourse interval; the quantization scaling value corresponding to the temperature deviation and the rate of change of the deviation is represented by x, then the quantization scaling value x and the fuzzy subset The membership degree between them is: .

[0010] As a further improvement to this technical solution, when step S3 integrates the fuzzy state vectors corresponding to temperature deviation and deviation change rate into a fuzzy state combination, if temperature deviation and deviation change rate each correspond to a fuzzy state vector, then they are integrated according to a fixed order pairing logic. If there are multiple fuzzy state vectors corresponding to temperature deviation and deviation change rate, then when integrating fuzzy state combinations, cross-integration is performed in a fixed order, and each fuzzy state combination includes a unique fuzzy state vector corresponding to temperature deviation and deviation change rate, so that each fuzzy state vector participates in integration, resulting in multiple fuzzy state combinations.

[0011] As a further improvement to this technical solution, when step S3 uses multiple fuzzy state vectors to match multiple activation rules, the specific steps are as follows: Step S3.1: Calculate the output intensity of each activation rule using the implication operator: Extract the membership degree corresponding to the temperature deviation fuzzy subset and the deviation change rate fuzzy subset in the fuzzy state combination respectively, take the minimum value of the two membership degrees as the output intensity of the activation rule, and finally obtain the output intensity corresponding to each activation rule; Step S3.2: Use the maximum operator to extract the maximum output intensity among all output intensities; define the activation rule corresponding to the maximum output intensity as the final activation rule.

[0012] As a further improvement to this technical solution, step S4 uses the centroid method to defuzzify the fuzzy set as the correction coefficient. Specifically, each universe value in the fuzzy set is regarded as a discrete particle, and its corresponding membership degree is used as the particle weight. The weighted centroid of the discrete particle system is calculated, and the obtained weighted centroid is the correction coefficient of the defuzzification output.

[0013] As a further improvement to this technical solution, the dynamic control parameters in step S4 are specifically as follows: The correction coefficients obtained from the defuzzification are: correction scaling coefficients. Corrected integral coefficients Corrected differential coefficients ; Using the correction scaling factor after defuzzification Corrected integral coefficients Corrected differential coefficients Adjusting the baseline proportional coefficient Kp0, baseline integral coefficient Ki, and baseline derivative coefficient Kd0, the updated dynamic control parameters are obtained as follows: ; Where Kp(k), Ki(k), and Kd(k) are the dynamic control parameters at time k.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: In this multi-zone temperature equalization control method for carbonitriding heat treatment of bearings, step S2 involves pre-setting the fuzzy universe corresponding to the temperature deviation and the rate of change of deviation, and then dividing the fuzzy universe into multiple fuzzy subsets. Adjacent fuzzy subsets partially overlap, further enabling a gradual representation of the continuous parameter states of temperature deviation and the rate of change of deviation, avoiding control discontinuities and abrupt changes in action caused by absolute division of parameter states. Furthermore, each fuzzy subset possesses a corresponding fuzzy vector during the fuzzy subset division. Subsequently, the fuzzy vectors corresponding to the temperature deviation and the rate of change of deviation are matched, and the membership degree between the temperature deviation, the rate of change of deviation, and the corresponding fuzzy subset is calculated based on the membership function, obtaining the membership vector corresponding to the temperature deviation and the rate of change of deviation, further providing quantitative data for subsequent rule matching. The fit of the fuzzy state is used as a basis for subsequent step S4 control, providing a fuzzy state basis that fits the actual working conditions and effectively making up for the shortcomings of traditional temperature control in accurately capturing dynamic temperature fluctuations. In step S3, by fusing the fuzzy vectors corresponding to temperature deviation and deviation change rate, and membership vectors, a fuzzy state combination covering the current working conditions is formed and matched with the corresponding activation rules again. In order to accurately select the control rule with the highest fit with the current carbonitriding temperature working conditions and avoid control conflicts caused by the parallel execution of multiple rules, the implication operator is used to calculate the output intensity of each activation rule (quantifying the rule matching degree), and the maximum operator is used to extract the rule with the highest output intensity as the final activation rule. This ensures that the final decision rule closely fits the process requirements for temperature stability and guarantees the accuracy of the control strategy. Step S4 then constructs a corresponding fuzzy output set based on the fuzzy state vector and membership vector corresponding to the final activation rule. The centroid method is then used to transform the fuzzy set into quantitative PID parameter correction coefficients. This transforms the fuzzy decision into precise control commands that can directly drive the actuator, enabling fine-tuning of heating power and fan speed in each heating zone. This effectively ensures balanced and stable temperature across multiple zones within the furnace, suppresses temperature fluctuations, and helps achieve core process indicators such as bearing diffusion layer uniformity and surface hardness. It avoids process problems such as uneven diffusion layer depth / carbonitrile concentration and excessive thermal deformation of the bearing race, ensuring that core indicators such as bearing diffusion layer uniformity and surface hardness meet standards. Simultaneously, it reduces subsequent grinding operations, lowers production costs, and improves part dimensional stability.

[0015] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the overall steps of the present invention. Detailed Implementation

[0017] The technical solutions in 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 embodiments of the present invention, and not all embodiments. 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.

[0018] refer to Figure 1 As shown, a multi-zone temperature equalization control method for carbonitriding heat treatment of bearings includes the following steps: S1. Calculate the temperature deviation between the corresponding temperature threshold and the actual temperature in the same heating zone, and the rate of change of the deviation. S2. Preset the fuzzy universe of discourse corresponding to the temperature deviation and the rate of change of the deviation, divide each fuzzy universe of discourse into multiple fuzzy subsets including the universe values, and adjacent fuzzy subsets include partially overlapping universe values. Each fuzzy subset corresponds to a fuzzy state vector. The temperature deviation and the rate of change of deviation are converted into values ​​of the fuzzy universe of discourse. Based on the values ​​of the universe of discourse, the fuzzy subsets corresponding to the temperature deviation and the rate of change of deviation, as well as the fuzzy state vectors corresponding to the fuzzy subsets, are matched. After matching is completed, the membership degree between temperature deviation, deviation change rate and corresponding fuzzy subset is calculated to obtain the membership degree vector; S3. A control rule library containing multiple fuzzy state combinations and corresponding activation rules for the fuzzy state combinations is pre-set; the fuzzy state vectors corresponding to temperature deviation and deviation change rate in step S2 are integrated into fuzzy state combinations and matched with the corresponding final activation rules in the control rule library. If the temperature deviation and the rate of change of deviation correspond to multiple fuzzy state vectors, then the implication operator and the maximum operator are used to calculate the output intensity of each activation rule, and the final activation rule is selected. S4. Integrate the fuzzy state vector and membership vector from step S2 to obtain a fuzzy set; use the centroid method to defuzzify the fuzzy set as correction coefficients; and use the correction coefficients to adjust the benchmark coefficients to obtain dynamic control parameters.

[0019] In the implementation process of the above embodiments, the dynamic control parameter S1 first calculates the temperature deviation and the rate of change of deviation in each heating zone, providing a precise operating condition data basis for subsequent regulation; the dynamic control parameter S2 achieves a gradual transformation of continuous temperature parameters into fuzzy states by presetting the fuzzy universe, dividing overlapping fuzzy subsets and matching fuzzy state vectors, and calculating membership vectors, thus solving the problem of parameter and rule adaptation; the dynamic control parameter S3 integrates state combinations based on the fuzzy state vector of the dynamic control parameter S2, and utilizes implication operators and maximum operators. The optimal activation rule is selected to avoid conflicts between multiple rules and ensure a high degree of fit between the rule and the operating conditions. Step S4, dynamic control parameters, integrates the fuzzy state and membership vector corresponding to the final rule and obtains the dynamic control parameters through the centroid method. Each step is connected and coordinated to achieve full-process coverage from operating condition data acquisition and fuzzification processing to rule matching and precise control. It also ensures the balanced and stable temperature of multiple zones during carbonitriding, effectively suppresses temperature fluctuations, and helps the bearing achieve core process indicators such as uniformity of the diffusion layer and surface hardness. At the same time, it reduces the thermal deformation of the bearing rings and subsequent processing costs, and improves the dimensional stability and performance consistency of bearing parts.

[0020] The more detailed work steps are as follows: S1. Obtain the preset temperature threshold for each heating zone, as well as the actual temperature collected at a fixed sampling frequency; The temperature deviation between the corresponding temperature threshold and each actual temperature in the same heating zone is calculated sequentially, along with the rate of change of the deviation. The rate of change of the deviation describes the change in temperature deviation between two adjacent sampling times, reflecting the trend of temperature deviation change (increasing or decreasing). More specifically: The temperature deviation calculated in step S1 accurately reflects the deviation and trend of the actual temperature of each heating zone relative to the temperature threshold, ensuring reliable data for subsequent adjustment in step S4. Specifically, a positive temperature deviation indicates that the actual temperature is higher than the temperature threshold, while a negative deviation indicates that the actual temperature is lower than the temperature threshold. The specific expression is as follows: ; in: Let K be the temperature deviation at the k-th sampling time. This represents the temperature threshold corresponding to the k-th sampling time. The actual temperature at the k-th sampling time; In step S1, the rate of change of deviation is calculated, and the specific expression is as follows: ; in: Let $\frac{k}{k}$ be the rate of change of the deviation at the $k$-th sampling time. The temperature deviation at the (k-1)th sampling time. Let K be the temperature deviation at the k-th sampling time. The sampling period (specifically, the sampling time) (Time interval between k).

[0021] This example further considers that after the control command is output in step S1, since the temperature deviation and deviation change rate output in step S1 are continuous physical parameters, they cannot be directly matched with discrete fuzzy control rules. Moreover, a single numerical judgment is prone to cause the control action to undergo abrupt switching between one and the other, which is difficult to adapt to the requirements of bearing carbonitriding process for temperature control smoothness. Therefore, step S2 is executed. Specifically, by pre-setting the fuzzy universe of discourse corresponding to the temperature deviation and deviation change rate under each heating zone, each fuzzy universe of discourse is divided into multiple fuzzy subsets including universe values, and adjacent fuzzy subsets include partially overlapping universe values. Each fuzzy subset corresponds to a fuzzy state vector. The temperature deviation and the rate of change of deviation are converted into universe values ​​within the fuzzy universe of discourse, respectively. Based on the universe values, the fuzzy subsets corresponding to the temperature deviation and the rate of change of deviation under each heating zone are matched, as well as the fuzzy state vectors corresponding to the fuzzy subsets. The membership degree between the temperature deviation, the rate of change of deviation and the corresponding fuzzy subsets is calculated to form the membership degree vectors corresponding to the temperature deviation and the rate of change of deviation. Step S2 presets the fuzzy universe of discourse corresponding to the temperature deviation and the rate of change of the deviation. The structures of the fuzzy universe of discourse corresponding to the temperature deviation and the rate of change of the deviation are completely consistent, and both are composed of multiple symmetrically distributed discrete universe values. Each value within the fuzzy universe of discourse is used to quantify the direction and extent of parameter deviation; specifically, the fuzzy universes of discourse corresponding to temperature deviation and the rate of deviation change are as follows: ; in, For the fuzzy universe of discourse corresponding to the temperature deviation, Let n be the fuzzy universe of discourse corresponding to the rate of change of deviation, and n be the one-sided maximum value of the fuzzy universe of discourse (considering the requirements of bearing carbonitriding control, the one-sided maximum value n=3 is usually used). The total number of universe values ​​in the fuzzy universe of discourse is... In the fuzzy universe of discourse, the positive and negative signs of the universe values ​​correspond to the temperature deviation, the deviation rate of change relative to the temperature threshold in step S1, the direction of deviation (positive sign means too high, negative sign means too low), and the magnitude of the absolute value corresponds to the strength of the deviation (the larger the absolute value, the stronger the deviation). To accurately adapt to the continuous fluctuations in temperature deviation and deviation change rate during the carbonitriding heat treatment of bearings, the abstract parameter deviations (temperature deviation and deviation change rate) are transformed into identifiable and decision-making concrete fuzzy states. Simultaneously, this ensures that parameter deviations under all normal operating conditions can be effectively captured without any control blind spots, meeting the core requirements of the process for precise temperature control, uniform diffusion layer structure, and ring thermal deformation ≤0.05mm. Therefore, the fuzzy universe is divided into multiple fuzzy subsets, each covering a specific interval within the fuzzy universe, forming fuzzy subsets. After dividing the fuzzy subsets, multiple fuzzy subsets can cover the entire fuzzy universe without omission. Each fuzzy subset represents a different fuzzy state. The specific fuzzy state can quantify the deviation direction (represented by the positive or negative sign of the universe value, with a positive sign indicating a higher deviation and a negative sign indicating a lower deviation) and the degree of deviation (represented by the absolute value of the universe value, with a larger absolute value indicating a stronger deviation). The specific representations that correspond one-to-one with the fuzzy subsets and contain information on the deviation direction and degree are the fuzzy state vectors corresponding to the fuzzy subsets. The fuzzy universe is divided into multiple fuzzy subsets, and the specific expression for the set of fuzzy subsets is as follows: .

[0022] In step S2, the corresponding universe of discourse values ​​for temperature deviation and deviation change rate are: calculate the quantization factor, and then convert the temperature deviation and deviation change rate into universe of discourse values ​​within the fuzzy universe of discourse using the quantization factor. ; ; in: This is the temperature deviation quantification factor. The quantification factor for the rate of change of deviation. For the one-sided maximum value of the fuzzy universe of discourse, This represents the upper limit of the temperature deviation range. This represents the upper limit of the deviation rate range. Temperature deviation at sampling time k The corresponding universe value, The rate of change of the deviation at the k-th sampling time The corresponding universe value, , These are the quantized scaled values ​​of temperature deviation and deviation change rate, respectively, after being scaled by the corresponding quantization factors. This is the rounding function; The rounding function mentioned above is used to avoid the problem that continuous temperature deviations and deviation change rates cannot be directly matched with discrete universe values ​​within the fuzzy universe. Specifically, since the temperature deviations and deviation change rates detected during the bearing carbonitriding process are continuous physical quantities, the quantization scaling values ​​after quantization factor conversion may still be non-integers (such as 1.2, 1.8), while the universe values ​​of the fuzzy universe are discrete integers such as {-3, -2, -1, 0, 1, 2, 3}. If the non-integer quantization scaling values ​​are not converted using the rounding function, they will not be able to establish a correspondence with the discrete universe values, and thus cannot be matched with the corresponding fuzzy subset.

[0023] Furthermore, considering that the furnace temperature naturally fluctuates continuously during the bearing carbonitriding process, the temperature deviation and deviation change rate detected in real time by high-precision thermocouples with an accuracy of ±0.5℃ are both continuous physical quantities, when matching the fuzzy subsets corresponding to the temperature deviation and deviation change rate, although the continuous quantization scaling values ​​can be converted into discrete universe values ​​within the fuzzy universe using a rounding function, subsequent step S3 needs to activate the corresponding control rule base based on the fuzzy subset matching results, and step S4 needs to perform defuzzification based on membership information to correct the coefficients. However, rounding will cause an absolute division between adjacent discrete universe values ​​corresponding to the quantization scaling values ​​(e.g., 1.4℃). (Rounded to 1, 1.6℃ rounded to 2). For absolute division, the control rule activated in step S3 will jump directly from the rule corresponding to slightly high to the rule corresponding to moderately high, ultimately leading to abrupt jumps and lack of transition gradients in subsequent analysis and correction coefficients. This can cause sudden and significant changes in the control actions such as heating power or fan speed, disrupting the temperature stability required by the process (e.g., furnace temperature oscillation exceeding ±1℃ control accuracy). Therefore, to avoid the above situation, in the process of dividing the fuzzy universe into multiple fuzzy subsets in step S2, partial overlap needs to be set between adjacent fuzzy subsets. Specifically, a specific interval of adjacent fuzzy subsets shares at least one universe value. Specifically: Let k represent an integer from 1 to n, and assign a continuous universe of discourse value to each fuzzy subset. Specifically, the negative fuzzy subset in the fuzzy subset set... cover Zero-fuzzy subset cover positive fuzzy subset cover .

[0024] Meanwhile, since step S2 has already converted the temperature deviation and deviation change rate into quantized scaling values ​​within the fuzzy universe of discourse through quantization factors when calculating the universe values, but to avoid absolute attribution judgments due to rounding, which could lead to rigid matching of control rules and abrupt changes in the output of correction coefficients, the parameter states corresponding to different quantized scaling values ​​do not absolutely belong to a single fuzzy subset (they may simultaneously be close to multiple adjacent fuzzy subsets). Since rule matching and control decisions in step S3 require precise characterization of the degree of fit of the quantized scaling values ​​to each fuzzy subset, step S2 sequentially uses membership functions to map the quantized scaling values ​​of temperature deviation and deviation change rate to the membership degrees of the corresponding fuzzy subsets (temperature deviation, deviation change rate), forming a membership vector corresponding to the temperature deviation and deviation change rate, specifically: Determine whether the temperature deviation and the rate of change of deviation correspond to the interval of a certain fuzzy subset. If so, retrieve the feature points corresponding to the fuzzy subset and use a piecewise formula to calculate the membership degree between the temperature deviation, the rate of change of deviation, and the corresponding fuzzy subset. This provides a quantitative basis for the integration of fuzzy state combinations and the activation of fuzzy control rules in subsequent step S3. Specifically: if the temperature deviation and the rate of change of deviation correspond to the interval of the fuzzy subset... Request a fuzzy subset The corresponding left endpoint 'a' and right endpoint 'b' of the universe of discourse interval; the quantization scaling value corresponding to the temperature deviation and the rate of change of the deviation is represented by 'x', then the quantization scaling value x and the fuzzy subset The membership degree between them is: .

[0025] In this example, in order to accurately match the temperature control requirements of bearing carbonitriding, step S3 presets a control rule library based on the carbonitriding process target. The control rule library includes multiple fuzzy state combinations and corresponding activation rules for the fuzzy state combinations (the activation rules are specifically the clearly matched control parameter adjustment strategies (such as synchronously adjusting the control parameters such as slightly reducing the heating power in each heating zone and increasing the fan speed at a medium speed). In step S2, the fuzzy state vectors corresponding to the temperature deviation and the rate of change of deviation in each heating zone are received and integrated to form the fuzzy state combination corresponding to the current working condition. The fuzzy state combination is sequentially input into the control rule library, and the unique activation rule corresponding to each heating zone in the fuzzy state combination is matched to obtain the final activation rule for each heating zone. Specifically: Since there is interval overlap between adjacent fuzzy subsets in step S2, the quantization scaling value will show non-zero membership degree on multiple adjacent fuzzy subsets after being substituted into the membership function. Therefore, the temperature deviation and the rate of change of deviation may correspond to multiple fuzzy state vectors. Therefore, when integrating the fuzzy state vectors corresponding to temperature deviation and deviation change rate into a fuzzy state combination in step S3, if temperature deviation and deviation change rate each correspond to a fuzzy state vector, then integration is performed according to a fixed order pairing logic. Specifically, integration can be performed in a fixed order with the fuzzy state vector of temperature deviation first and the fuzzy state vector of deviation change rate last, to obtain a unique fuzzy state combination, and the activation rule corresponding to the fuzzy state combination in the control rule base is matched as the final activation rule. If there are multiple fuzzy state vectors corresponding to temperature deviation and deviation change rate, then when integrating fuzzy state combinations, cross-integration is performed in a fixed order, and each fuzzy state combination includes a unique fuzzy state vector corresponding to temperature deviation and deviation change rate, so that each fuzzy state vector participates in integration to obtain multiple fuzzy state combinations. When matching multiple fuzzy state vectors with multiple activation rules, the prerequisite for activation rules must simultaneously satisfy the fuzzy states corresponding to temperature deviation and deviation change rate (these are core co-factor parameters affecting the temperature stability of carbonitriding furnaces; to avoid triggering unsuitable control strategies due to excessively high membership of a single parameter (temperature deviation, deviation change rate), resulting in excessive adjustments to heating power, fan speed, etc.), therefore, an implication operator is used to calculate the output strength of each activation rule. Specifically: the membership degrees corresponding to the fuzzy subset of temperature deviation and the fuzzy subset of deviation change rate in the fuzzy state combination are extracted respectively, and the minimum value of the two membership degrees is taken as the output strength of the activation rule (the degree to which the current input state meets the rule prerequisite); finally, the output strength corresponding to each activation rule is obtained. (m represents the number of output intensities corresponding to the activation rule), where the output intensity of the t-th activation rule is as follows: ; in, Let represent the membership degree of the fuzzy state corresponding to the temperature deviation. The membership degree of the fuzzy state corresponding to the rate of change of deviation.

[0026] For all activation rules and their corresponding output strengths obtained after integration and matching, By utilizing the extreme value characteristic of the maximum operator, the maximum output intensity is extracted from all output intensities. Since the output intensity is calculated by the minimum membership value of the fuzzy states corresponding to temperature deviation and deviation change rate, its magnitude directly reflects the degree of matching between the activation rule and the current operating condition (the higher the output intensity, the stronger the fit between the fuzzy state combination corresponding to the rule and the actual furnace temperature deviation and deviation change rate). Therefore, the activation rule corresponding to the maximum output intensity is the activation rule with the highest fit to the current carbonitriding operating condition (fuzzy state combination of temperature deviation and deviation change rate), and the maximum output intensity is defined as the maximum value of the activation rule corresponding to the current carbonitriding condition. The corresponding activation rule is the final activation rule, which can accurately select the control strategy that best matches the current furnace temperature fluctuation state from multiple sets of different activation rules. This avoids control conflicts that may be caused by the parallel execution of multiple rules, and ensures that the final activation rule closely matches the temperature stability requirements of the carbonitriding process. By matching the rule with the highest degree of fit, the adjustment of control parameters such as heating power and fan speed can be more accurately adapted to the current working conditions. This helps to ensure the stable achievement of uniformity of the infiltrated layer structure and hardness gradient, reduce the occurrence of process defects such as local overheating and underheating, and ultimately support the improvement of core performance indicators such as bearing contact fatigue life.

[0027] Step S4 is used to defuzzify the final activation rule output from step S3 into a correction coefficient that allows the actuator (heating element, circulating fan) to execute precisely. Specifically: When filtering the final activation rules in step S3, the output intensity is the membership degree between different fuzzy subsets corresponding to the temperature deviation and the rate of change of the deviation. If the maximum output intensity is obtained by using the maximum operator as the membership degree of a certain fuzzy subset corresponding to the temperature deviation, then the fuzzy state vector of the temperature deviation in step S2 (i.e., the universe value of different fuzzy subsets of the fuzzy subset set corresponding to the temperature deviation) and the membership vector (i.e., the membership degree of different fuzzy subsets of the fuzzy subset set corresponding to the temperature deviation) are retrieved. By integrating the fuzzy state vectors and membership vectors of each heating zone, a fuzzy set corresponding to all heating zones is obtained (the universe values ​​and membership degrees in the fuzzy set correspond one-to-one through fuzzy subsets). The centroid method is used to defuzzify the fuzzy set as the correction coefficient. Specifically, each universe value in the fuzzy set (each partition corresponds to a unique universe value) is regarded as a discrete particle, and its corresponding membership degree (each partition corresponds to a unique membership degree) is used as the particle weight. The weighted centroid of the discrete particle system is calculated, and the obtained weighted centroid is the correction coefficient of the defuzzification output. A more detailed expression is as follows: Let A be a fuzzy set, where the universe of discourse is... The membership degree corresponding to each universe value is Set the membership threshold to Then the correction coefficient after defuzzification (General symbol, can be replaced with correction ratio) Corrected integral coefficients Corrected differential coefficients )as follows: ; The preset benchmark parameters for the process are: benchmark proportional coefficient Kp0, benchmark integral coefficient Ki, and benchmark differential coefficient Kd0 (determined according to the carbonitriding process manual or trial production experience, such as benchmark proportional coefficient Kp0=5, benchmark integral coefficient Ki0=0.2, and benchmark differential coefficient Kd0=1 for the strong infiltration stage). The correction coefficients obtained from defuzzification are: correction scaling factor Corrected integral coefficients Corrected differential coefficients ; Using the correction scaling factor after defuzzification Corrected integral coefficients Corrected differential coefficients Adjusting the baseline proportional coefficient Kp0, baseline integral coefficient Ki, and baseline derivative coefficient Kd0, the updated dynamic control parameters are obtained as follows: ; Where Kp(k), Ki(k), and Kd(k) are the dynamic control parameters at time k (updated in real time according to the operating conditions).

[0028] For the specific fuzzy control rules of the bearing carbonitriding process, the specific examples of steps S1, S2, S3, and S4 in this example are as follows: Taking four heating zones as an example, the functions of the four heating zones are preheating, strong infiltration, and diffusion temperature equalization, respectively; in step S2, the fuzzy universe of discourse for temperature deviation (E) and deviation change rate (EC) is set to {-3,-2,-1,0,1,2,3}, and then the universe of discourse is gradually represented by dividing it into 7 adjacent overlapping fuzzy subsets. Finally, the fuzzy state of temperature deviation and deviation change rate are obtained through step S3; and the fuzzy state of temperature deviation, deviation change rate, and process stage are used as three-dimensional inputs. The degree of matching is quantified by membership vector, and the output intensity is calculated by the implication operator. The final activation rule corresponding to the control rule library is selected by the maximum operator. This is the precise control strategy for each heating zone that adapts to the current fuzzy state of temperature deviation, the fuzzy state of deviation change rate, and the corresponding process stage. Specifically, it clarifies the differentiated adjustment scheme of control parameters such as the adjustment range of heating power and the adjustment value of fan speed for each heating zone. For example, for the fuzzy state combination of slightly high and slightly increased temperature in a certain heating zone during the strong infiltration stage, the final activation rule is the specific execution instruction of slightly reducing the heating power (ΔP=-4%) and rapidly increasing the fan speed (+300r / min) in that heating zone. After receiving the specific execution instructions, step S4 integrates the fuzzy state vectors and membership vectors of each heating zone. This process involves mapping the universe value (quantified by temperature deviation and deviation change rate) of each heating zone to its membership degree (characterizing the degree of matching of the universe value), and summarizing them into a fuzzy set covering the working conditions of all heating zones. Specifically, the universe value of each heating zone is first extracted (e.g., the universe value of the preheating zone is -2, and the strong infiltration zone is 1), and then its corresponding membership degree is associated (e.g., the membership degree of the preheating zone is 0.9, and the membership degree of the strong infiltration zone is 0.8), so that each universe value in the fuzzy set is bound to a unique membership degree, forming a mapping set of "universe value - membership degree". Taking the carbonitriding stage of bearings as an example, the universe of discourse values ​​of the four heating zones are -2 for the preheating zone, 1 for the strong infiltration zone, -1 for the diffusion zone, and 0 for the isothermal zone, with corresponding membership degrees of 0.9, 0.8, 0.7, and 0.9, respectively. The membership degree threshold μ0 is set to 0.6, and the correction proportionality coefficient is calculated using the formula. , Molecular part: ; Denominator: ; Then correct the proportionality coefficient Similarly, the corrected integral coefficients can be calculated. Corrected differential coefficients .

[0029] In summary: Step S4 uses the fuzzy state vector and membership vector output in step S2, and the final activation rule determined in step S3. The core logic revolves around process adaptability and control precision. The key point is in step S2, where, after dividing the fuzzy universe of discourse into multiple fuzzy subsets, the partially overlapping universe values ​​in adjacent fuzzy subsets can break the absolute division of continuous parameters such as temperature deviation and deviation change rate, achieving a gradual representation of parameter states. Specifically: Because the temperature inside the bearing carbonitriding furnace is naturally continuously fluctuating due to the effects of furnace structure delay and uneven atmosphere flow, traditional discretization can easily lead to abrupt changes in control actions, which are either one or the other. Overlapping domain values ​​can allow the parameter state to be close to multiple adjacent fuzzy subsets at the same time, avoiding control discontinuities. Its role is to provide a basis for subsequent smooth control and adapt to the stringent requirements of the process for temperature stability. Furthermore, in step S3, during the process of further determining the final activation rule, the most suitable collaborative activation rule for the current temperature conditions of each heating zone is accurately selected based on the degree of fit between the membership vector quantization parameter and each fuzzy subset. Step S4 constructs a complete fuzzy set by integrating the fuzzy state vector and membership vector corresponding to the final activation rule under each heating zone, and then obtains the correction coefficient by defuzzification using the centroid method. This fully integrates the quantitative basis of the working condition information and rule matching under each heating zone, so that the final correction coefficient can achieve small-scale gradual change and precise adaptation: it avoids furnace temperature oscillation caused by excessive control amplitude (such as the uneven carbon and nitrogen concentration in the infiltration layer caused by a sudden rise and fall in temperature during the strong infiltration stage), and can specifically compensate for the temperature deviation of multiple zones, ensuring that the temperature of each heating zone synchronously approaches the temperature threshold of the corresponding different partitions in step S1. As can be seen from the above, in this example, steps S1, S2, S3, and S4 are for the carbonitriding heat treatment of bearings. When multiple different heating zones are set up by dividing them into independent heating zones, in order to avoid the temperature of each heating zone affecting each other through heat transfer and airflow in the furnace during the analysis and control of control parameters such as temperature and atmosphere concentration, resulting in uneven temperature and excessive gradient in multiple zones, which would lead to inconsistent diffusion layer depth, uneven carbonitriding concentration distribution, and excessive thermal deformation of bearing rings, the process avoids the consequences of decreased core performance such as surface hardness and wear resistance of parts and increased subsequent processing costs. While existing methods can employ PID control to analyze and adjust key control parameters such as temperature and atmosphere flow in the carbonitriding heat treatment of bearings, traditional PID control has significant limitations in this process scenario. The core issue is that the proportional, integral, and derivative correction coefficients are fixed after initial tuning (e.g., through trial and error or the Ziegler-Nichols method), making it impossible to dynamically adapt to different heating zones. For example, consider four heating zones: preheating, strong infiltration, and diffusion homogenization. During preheating, rapid heating is required to shorten the process cycle; a fixed small integral correction coefficient leads to insufficient heating rate and lag. During strong infiltration, maintaining a high and stable temperature of 850-900℃ is crucial; a fixed proportional correction coefficient easily causes temperature overshoot, leading to localized overheating. During diffusion, slow cooling is necessary to promote uniform diffusion of carbon and nitrogen atoms; a fixed coefficient makes it difficult to balance cooling rate and temperature stability. During homogenization, ensuring temperature balance across multiple zones is essential; a fixed coefficient cannot specifically compensate for regional differences. Meanwhile, although existing PID control methods also include fuzzy PID control, they still haven't broken free from the design framework of general scenarios and differ fundamentally from this example: While conventional fuzzy PID control methods achieve dynamic adjustment of coefficients through fuzzy logic, most lack an overlapping structure for adjacent fuzzy subsets. The division of continuous parameters such as temperature deviation and deviation change rate remains a discrete characteristic with no clear distinction, making it unsuitable for the naturally continuous temperature fluctuations within the bearing carbonitriding furnace, easily leading to abrupt changes in control actions. Furthermore, its fuzzy rules are mostly general logic, not related to the specific process requirements of carbonitriding for uniformity of the infiltration layer and thermal deformation of the bearing rings, and often employ... The single-input-single-output structure, without further integration of operating information from multiple heating zones, is prone to inter-zone control conflicts when applied to multi-zone temperature control (e.g., the superposition of cooling in heating zone A and heating in heating zone B disrupts the stability of the furnace atmosphere), making it difficult to achieve coordinated and balanced control across multiple zones. In contrast, step S2 in this example uses the design of adjacent overlapping fuzzy subsets to achieve a gradual representation of continuous parameters, matches the multi-zone coordination requirements with a process-specific rule library, and integrates the operating information of each heating zone with the membership quantification basis to generate correction coefficients. This fully adapts to the stringent process requirements of bearing carbonitriding, completely overcoming the limitations of conventional fuzzy PID control methods in coefficient representation, rule design, and multi-zone coordination.

[0030] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-zone temperature equalization control method for carbonitriding heat treatment of bearings, characterized in that... This includes the following steps: S1. Calculate the temperature deviation between the corresponding temperature threshold and the actual temperature in the same heating zone, and the rate of change of the deviation. S2. Preset the fuzzy universe of discourse corresponding to the temperature deviation and the rate of change of deviation under each heating zone, divide each fuzzy universe of discourse into multiple fuzzy subsets including universe values, and adjacent fuzzy subsets include partially overlapping universe values. Each fuzzy subset corresponds to a fuzzy state vector. The temperature deviation and the rate of change of deviation under each heating zone are transformed into the universe of discourse value within the fuzzy universe of discourse. Based on the universe of discourse value, the fuzzy subsets corresponding to the temperature deviation and the rate of change of deviation, as well as the fuzzy state vectors corresponding to the fuzzy subsets, are matched. After matching is completed, the membership degree between the temperature deviation, the rate of change of deviation and the corresponding fuzzy subset under each heating zone is calculated to obtain the membership degree vector; S3. A control rule library containing multiple fuzzy state combinations and corresponding activation rules for the fuzzy state combinations is pre-set; the fuzzy state vectors corresponding to the temperature deviation and deviation change rate under each heating zone in step S2 are integrated into fuzzy state combinations, and the final activation rules corresponding to each heating zone in the control rule library are matched. If the temperature deviation and the rate of change of deviation correspond to multiple fuzzy state vectors, then the implication operator and the maximum operator are used to calculate the output intensity of each activation rule, and the final activation rule under each heating zone is selected. S4. Integrate the fuzzy state vector and membership vector of each heating zone in step S2 to obtain a fuzzy set; use the centroid method to defuzzify the fuzzy set as the correction coefficient; and use the correction coefficient to adjust the benchmark coefficient to obtain the dynamic control parameters.

2. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 1, characterized in that: In step S1, the temperature deviation is the difference between the temperature threshold and the actual temperature; the deviation change rate is the ratio of the change in temperature deviation between two adjacent sampling times to the time interval between sampling times.

3. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 1, characterized in that: In step S2, the fuzzy universe structures corresponding to the temperature deviation and the rate of change of deviation are completely identical, both consisting of multiple symmetrically distributed discrete universe values. When dividing the fuzzy universe into multiple fuzzy subsets in step S2, each fuzzy subset covers a specific interval within the fuzzy universe, forming a set of fuzzy subsets. After dividing the fuzzy subsets, multiple fuzzy subsets can cover the entire fuzzy universe without omission. Furthermore, each fuzzy subset represents a different fuzzy state, which is the fuzzy state vector corresponding to the fuzzy subset.

4. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 3, characterized in that: In step S2, when dividing the fuzzy universe of discourse of each heating zone into multiple fuzzy subsets, there is partial overlap between adjacent fuzzy subsets. Specifically, a specific interval of an adjacent fuzzy subset shares at least one universe of discourse value.

5. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 4, characterized in that: In step S2, the corresponding universe of discourse values ​​for temperature deviation and the rate of change of deviation are: Calculate the quantization factor, and then convert the temperature deviation and the rate of change of deviation into universe values ​​within the fuzzy universe of discourse using the quantization factor: ； ； in: This is the temperature deviation quantification factor. The quantification factor for the rate of change of deviation. For the one-sided maximum value of the fuzzy universe of discourse, This represents the upper limit of the temperature deviation range. This represents the upper limit of the deviation rate range; Temperature deviation at sampling time k The corresponding universe value, The rate of change of the deviation at the k-th sampling time The corresponding universe value, , These are the quantized scaled values ​​of temperature deviation and deviation change rate, respectively, after being scaled by the corresponding quantization factors. This is the rounding function.

6. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 5, characterized in that: Step S2 sequentially uses membership functions to map the quantized scaling values ​​of temperature deviation and deviation change rate under each heating zone to the membership degree of the fuzzy subset corresponding to the temperature deviation and deviation change rate, forming a membership vector corresponding to the temperature deviation and deviation change rate, specifically: Determine whether the universe of discourse corresponding to the temperature deviation and the rate of change of deviation falls within the interval of a certain fuzzy subset. If it does, retrieve the feature points corresponding to the fuzzy subset and use a piecewise formula to calculate the membership degree between the temperature deviation, the rate of change of deviation, and the corresponding fuzzy subset. If the universe of discourse corresponding to the temperature deviation and the rate of change of deviation falls within the interval of the fuzzy subset... Request a fuzzy subset The corresponding left endpoint a and right endpoint b of the universe of discourse interval; the quantization scaling value corresponding to the temperature deviation and the rate of change of the deviation is represented by x, then the quantization scaling value x and the fuzzy subset The membership degree between them is: 。 7. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 6, characterized in that: When step S3 integrates the fuzzy state vectors corresponding to the temperature deviation and the rate of change of deviation in each heating zone into a fuzzy state combination, if the temperature deviation and the rate of change of deviation each correspond to a fuzzy state vector, then the integration is performed according to a fixed order pairing logic. If there are multiple fuzzy state vectors corresponding to the temperature deviation and the rate of change of deviation under each heating zone, then when integrating the fuzzy state combinations, cross-integration is performed in a fixed order, and each fuzzy state combination under the heating zone includes a unique fuzzy state vector corresponding to the temperature deviation and the rate of change of deviation, so that each fuzzy state vector participates in the integration and multiple fuzzy state combinations are obtained.

8. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 7, characterized in that: When step S3 uses multiple fuzzy state vectors to match multiple activation rules, the specific steps are as follows: Step S3.1: Calculate the output intensity of each activation rule using the implication operator: Extract the membership degree corresponding to the temperature deviation fuzzy subset and the deviation change rate fuzzy subset in the fuzzy state combination respectively, take the minimum value of the two membership degrees as the output intensity of the activation rule, and finally obtain the output intensity corresponding to each activation rule; Step S3.2: Use the maximum operator to extract the maximum output intensity from all output intensities; Define the activation rule corresponding to the maximum output intensity as the final activation rule for each heating zone.

9. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 8, characterized in that: Step S4 uses the centroid method to defuzzify the fuzzy set as the correction coefficient. Specifically, each universe value in the fuzzy set is regarded as a discrete particle, and its corresponding membership degree is used as the particle weight. The weighted centroid of the discrete particle system is calculated, and the obtained weighted centroid is the correction coefficient of the defuzzification output.

10. The multi-zone temperature equalization control method for bearing carbonitriding heat treatment according to claim 9, characterized in that: The dynamic control parameters in step S4 are specifically as follows: The correction coefficients obtained from the defuzzification are: correction scaling coefficients. Corrected integral coefficients Corrected differential coefficients ; Using the correction scaling factor after defuzzification Corrected integral coefficients Corrected differential coefficients Adjusting the baseline proportional coefficient Kp0, baseline integral coefficient Ki, and baseline derivative coefficient Kd0, the updated dynamic control parameters are obtained as follows: ； Where Kp(k), Ki(k), and Kd(k) are the dynamic control parameters at time k.

Images