Intelligent control method and system for sintering process based on powder metallurgy metal materials

Abstract

The present invention relates to the field of metal powder processing, and in particular to a method and system for intelligently controlling the sintering process of powder metallurgy metal materials. The method comprises: obtaining a temperature curve of the powder metallurgy sintering process, determining the offset of each curve using a cross-correlation function and aligning them, determining the segment length based on the segment suitability of the local area, analyzing the abnormality reference degree based on the correlation between the Mahalanobis distance of each segment and the sintering quality, obtaining the overall abnormality degree of each curve through normalization processing and weighted calculation, determining whether an abnormal temperature curve exists based on the overall abnormality degree and a preset abnormality degree threshold, and if the abnormality degree exceeds the preset threshold, issuing an early warning of abnormal sintering process and prompting adjustment of control. The present invention optimizes temperature control by accurately detecting local abnormalities, reducing production costs, and improving product quality and production efficiency.

Description

Technical Field

[0001] The present invention relates to the field of metal powder processing, and in particular to an intelligent control method and system for a sintering process based on powder metallurgy metal materials. Background Art

[0002] Powder metallurgy is an advanced metalworking process that manufactures metal products through the mixing, pressing, and sintering of metal powders. Compared to traditional processes, it can directly form complex precision parts, reducing processing costs and producing materials with special properties for applications in the automotive, machinery, electronics, and aerospace industries.

[0003] The sintering process of powder metallurgy (PM) metal materials is a crucial step, crucial for the final product's performance. During sintering, the powder particles undergo physical and chemical changes upon heating, such as diffusion bonding and densification. Improper sintering control can lead to problems such as high porosity and abnormal grain growth, seriously impacting product performance indicators such as strength, toughness, and wear resistance. Good sintering control ensures product density, enhances mechanical and corrosion resistance, and ensures stable chemical properties.

[0004] However, current methods for detecting anomalies in the sintering process's temperature curve present challenges. Existing tests primarily focus on the overall curve, making it difficult to accurately detect changes in internal characteristics. The sintering process is complex, and temperature variations at different stages significantly impact quality. However, current methods are unable to detect even the smallest internal anomalies, resulting in inaccurate temperature control, impacting product quality and hindering the development of powder metallurgy technology. Summary of the Invention

[0005] In order to solve the problem that the abnormal temperature curve detection of powder metallurgy metal materials during the sintering process is mostly targeted at the whole, making it difficult to capture small internal abnormalities, resulting in inaccurate temperature control and affecting product quality, the present invention provides solutions in the following aspects.

[0006] In a first aspect, an intelligent control method for a sintering process based on powder metallurgy metal materials includes: obtaining a temperature curve in a powder metallurgy sintering process; determining the offset of each temperature curve according to a cross-correlation function, aligning the curves, and judging the segment length according to the segment suitability of the local area of ​​each temperature curve; analyzing the abnormal reference degree of each segment based on the Mahalanobis distance of each segment and the correlation with the sintering quality, using the ratio between the abnormal reference degree of each segment and the sum of the abnormal reference degrees of all segments of the corresponding temperature curve for normalization processing, and weighting based on the normalized abnormal reference degree to obtain the overall abnormality degree of each temperature curve; if the overall abnormality degree is less than or equal to the abnormality degree threshold, then there is no abnormality in the sintering process; otherwise, if it is greater than the abnormality degree threshold, it is considered that there is an abnormal temperature curve in the sintering process, and then the staff is warned to adjust and control the sintering process.

[0007] Preferably, the offset alignment of the curves comprises the steps of:

[0008] Taking the first temperature change curve as the reference curve, calculate the cross-correlation function between each temperature change curve and the reference curve, obtain the offset value corresponding to the extreme value of the cross-correlation function of each temperature change curve, and take the difference between the offset value corresponding to the extreme value of the cross-correlation function and the average value of all offset values ​​as the actual offset value of each temperature curve;

[0009] According to the actual offset value of each temperature curve, each temperature curve is corrected horizontally to align all temperature curves in time.

[0010] Preferably, the offset alignment of the curves further includes:

[0011] Discretize the corrected temperature curve, and based on the discretization result, select the temperature curve with the largest number of discrete data points as the reference benchmark for the unified number of data points;

[0012] For temperature curves with fewer discrete data points than the reference benchmark, perform difference supplementation, starting from the last discrete data point of the temperature curve and supplementing data points in sequence until the number of discrete data points of the temperature curve reaches the reference benchmark;

[0013] For temperature curves with more discrete data points than the reference benchmark, truncation processing is performed. Starting from the last discrete data point of the temperature curve, the redundant data points are removed in sequence until the number of discrete data points of the temperature curve is reduced to the reference benchmark; the number of discrete data points of each temperature curve is unified.

[0014] Preferably, obtaining the segment length comprises the steps of:

[0015] Preset the initial length of the segment and obtain the discrete data points of the initial length in the discrete data for each temperature curve;

[0016] Traverse each temperature curve separately, divide each temperature curve according to the preset initial segment length, take any discrete data point as the target data, extract the discrete data point corresponding to the position of each target data from the current segment of the discrete data of all temperature curves, and construct the data sequence corresponding to the target data position of each temperature curve under the initial segment length;

[0017] Calculate the gradient between adjacent data points in the current segment of each temperature curve to obtain a gradient sequence, take any gradient data as target gradient data, and obtain a target gradient data sequence at a position corresponding to the target gradient data;

[0018] According to the information entropy of the data sequence at the corresponding position of the target data and the information entropy of the target gradient data sequence, the segmental fitness calculation formula is used to obtain the segmental fitness of each segment;

[0019] In response to the segmentation suitability being greater than the preset segmentation threshold, the current segmentation length is increased by one, the above steps are iterated, and the segmentation suitability is recalculated until the segmentation suitability is less than or equal to the preset segmentation threshold, then the current segmentation length is reduced by one, and a segmentation is completed; the remaining data is still iterated according to the initial segmentation length until all data are segmented.

[0020] Preferably, the calculation method of the segmented applicability includes the following steps:

[0021] The average of the sum of the information entropy of the data sequence at the corresponding position of the target data and the information entropy of the target gradient sequence is taken as the average information entropy, and the average information entropy is squared. The average of the squares of all average information entropies in the length of the current segment is calculated and square rooted to obtain the segment suitability of the current segment length.

[0022] Preferably, the calculation method of the abnormal reference degree includes:

[0023] Take any segment as the reference segment, calculate the average Mahalanobis distance of the reference segment, use a negative exponential function to perform exponential decay on the average Mahalanobis distance of all temperature curves corresponding to the reference segment, and use the result after subtracting the decay from 1 as the adjustment factor;

[0024] Calculate the difference between the average Mahalanobis distance of the reference segment and the average Mahalanobis distance of all temperature curves corresponding to the reference segment to obtain the Mahalanobis distance deviation of the reference segment; calculate the difference between the sintering quality of each temperature curve and the sintering quality of all temperature curves to obtain the sintering quality deviation; calculate the average value of the product between the Mahalanobis distance deviation and the sintering quality deviation, and divide it by the product of the standard deviation of the Mahalanobis distance of the reference segment and the standard deviation of the sintering quality to obtain the correlation coefficient; multiply the adjustment factor by the correlation coefficient as the abnormal reference degree of the reference segment.

[0025] Preferably, the calculation method of the overall abnormality degree includes:

[0026] Taking any temperature curve as the curve to be analyzed, the product sum of the abnormal reference degree of each segment of the curve to be analyzed and the average Mahalanobis distance of the corresponding segment is calculated respectively to obtain the overall abnormal degree of the curve to be analyzed.

[0027] In the second aspect, an intelligent control system for a sintering process based on powder metallurgy metal materials includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned intelligent control method for a sintering process based on powder metallurgy metal materials is implemented.

[0028] The present invention has the following effects:

[0029] 1. The present invention segments the temperature curve and calculates the anomaly reference degree based on the correlation between the Mahalanobis distance of each segment and the sintering quality, thereby achieving accurate identification of local anomalies and significantly improving the accuracy of anomaly detection. This helps to promptly discover and address potential problems in the sintering process and ensure product quality.

[0030] 2. Through precise anomaly detection and timely adjustment and control, the present invention can effectively reduce product defects and scrap rates caused by sintering anomalies, thereby reducing production costs. At the same time, the application of intelligent control systems reduces the workload and time of manual inspection, further improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a method flow chart of steps S1 to S4 in the intelligent control method for the sintering process of powder metallurgy metal materials according to an embodiment of the present invention.

[0032] Figure 2 It is a structural block diagram of an intelligent control system for a sintering process of powder metallurgy metal materials according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0034] Reference Figure 1 The intelligent control method for the sintering process of powder metallurgy metal materials includes steps S1 to S4, which are specifically as follows:

[0035] S1: Obtain the temperature curve during the powder metallurgy sintering process.

[0036] Select multiple batches of sintering processes, and accurately record the temperature changes from the beginning to the end of sintering to obtain the corresponding temperature change curves. The total number of collected temperature curves is recorded as .

[0037] For each temperature curve, the sintering results corresponding to each temperature curve are obtained by subjective evaluation to evaluate the quality. The sintering quality corresponding to the temperature curve is recorded as The sintering quality score is based on a percentage system, and the passing line for sintering quality is set as , which is used as the basis for measuring whether the sintering quality is qualified.

[0038] By collecting multiple sets of temperature curves and evaluating their corresponding quality, it can provide workers with a comprehensive and reliable reference for formulating intelligent control strategies for the sintering process, thereby ensuring the stability of the sintering process and the controllability of the quality of sintered products.

[0039] Further analysis shows that the temperature curve in the powder metallurgy sintering process contains a wealth of information. By comprehensively considering the trend changes of all temperature curves and segmenting them, the characteristics and potential problems of different stages in the sintering process can be captured in more detail. The abnormal reference degree of each segment reflects the degree of deviation of the segment from the ideal sintering state, and weighting the segments takes into account the differences in the importance of different segments on the sintering quality. The overall abnormality of the curve obtained by weighting can more accurately reflect the abnormality of the entire temperature curve, thereby providing a more reliable basis for the optimization and quality control of the sintering process. This method helps to promptly detect abnormalities in the sintering process, improve the stability and consistency of the sintering quality, and thus improve the overall performance of powder metallurgy products. The specific steps are as follows:

[0040] S2: Determine the offset of each temperature curve according to the cross-correlation function, perform offset alignment on the curves, and determine the segment length according to the segment suitability of the local area of ​​each temperature curve.

[0041] To offset align a curve, the following steps are included:

[0042] Taking the first temperature change curve as the reference curve, calculate the cross-correlation function between each temperature change curve and the reference curve, obtain the offset value corresponding to the extreme value of the cross-correlation function of each temperature change curve, and take the difference between the offset value corresponding to the extreme value of the cross-correlation function and the average value of all offset values ​​as the actual offset value of each temperature curve;

[0043] According to the actual offset value of each temperature curve, each temperature curve is corrected horizontally to align all temperature curves in time.

[0044] Discretize the corrected temperature curve, and based on the discretization result, select the temperature curve with the largest number of discrete data points as the reference benchmark for the unified number of data points;

[0045] For temperature curves with fewer discrete data points than the reference benchmark, perform difference supplementation, starting from the last discrete data point of the temperature curve and supplementing data points in sequence until the number of discrete data points of the temperature curve reaches the reference benchmark;

[0046] For temperature curves with more discrete data points than the reference benchmark, truncation processing is performed, starting from the last discrete data point of the temperature curve, and removing redundant data points in sequence until the number of discrete data points of the temperature curve is reduced to the reference benchmark;

[0047] Unify the number of discrete data points of each temperature curve.

[0048] It should be noted that the sampling frequency is set to 1Hz, that is, the temperature data is collected once per second. Starting from the starting time point of the temperature curve, the temperature data is collected in sequence. The temperature change curve The temperature value of a discrete data point is recorded as ,in is the temperature curve number, is the serial number of discrete data points on the temperature curve, and the difference supplement method can be linear interpolation;

[0049] It should also be added that during the truncation process, it is necessary to ensure that the main features and trends of the curve are retained. Generally, it is possible to prioritize retaining the data points at the beginning and middle parts of the curve, because these parts often contain the key information of the curve, while the tail of the curve may contain some unstable fluctuations or noise.

[0050] Through discretization processing and the method of unifying the number of data points, the corrected temperature curve can be kept consistent in data format, which facilitates subsequent analysis.

[0051] Get the segment length, including the following steps:

[0052] Preset the initial length of the segment and obtain the discrete data points of the initial length in the discrete data for each temperature curve;

[0053] Traverse each temperature curve separately, divide each temperature curve according to the preset initial segment length, take any discrete data point as the target data, extract the discrete data point corresponding to the position of each target data from the current segment of the discrete data of all temperature curves, and construct the data sequence corresponding to the target data position of each temperature curve under the initial segment length;

[0054] Calculate the gradient between adjacent data points in the current segment of each temperature curve to obtain a gradient sequence, take any gradient data as target gradient data, and obtain a target gradient data sequence at a position corresponding to the target gradient data;

[0055] According to the information entropy of the data sequence at the corresponding position of the target data and the information entropy of the target gradient data sequence, the segmental fitness calculation formula is used to obtain the segmental fitness of each segment;

[0056] In response to the segmentation suitability being greater than the preset segmentation threshold, the current segmentation length is increased by one, the above steps are iterated, and the segmentation suitability is recalculated until the segmentation suitability is less than or equal to the preset segmentation threshold, then the current segmentation length is reduced by one, and a segmentation is completed; the remaining data is still iterated according to the initial segmentation length until all data are segmented.

[0057] That is, the preset segmentation threshold is , which can be adjusted according to specific circumstances.

[0058] For example, suppose there are three temperature curves, each of which is divided into multiple segments. In the current analysis phase, the segment length of interest is set to 5 (i.e. ). For these three curves, the data in their current segments are as follows:

[0059] The current segment data of the first temperature curve is: ;

[0060] The current segment data of the second temperature curve is: ;

[0061] The current segment data of the third temperature curve is: ;

[0062] For each data point location , extract the data of the corresponding position from the current segment of all temperature curves, and then construct the corresponding data sequence. The specific data sequence constructed is as follows:

[0063] No. Data series: ;

[0064] No. Data series: ;

[0065] No. Data series: ;

[0066] No. Data series: ;

[0067] No. Data series: ;

[0068] Through the above steps, for each data point position , successfully constructed a data sequence containing the data points of all temperature curves at corresponding positions in the current segment.

[0069] For the current segment of each temperature curve, the gradient (i.e., the difference) between adjacent data points is calculated.

[0070] For the first curve: Gradient sequence: ;

[0071] For the second curve: Gradient sequence: ;

[0072] For the 3rd curve: Gradient sequence: ;

[0073] Build All Gradient sequence:

[0074] No. Gradient sequence: ;

[0075] No. Gradient sequence: ;

[0076] No. Gradient sequence: ;

[0077] No. Gradient sequence: .

[0078] It should be noted that in the process of segmenting discrete data, data with greater trend similarity among all discrete data can be distinguished by longer segments, while segments with smaller trend similarity should be segmented by smaller data to retain the detailed characteristics of the data in the segment.

[0079] The average of the sum of the information entropy of the data sequence at the corresponding position of the target data and the information entropy of the target gradient sequence is taken as the average information entropy, and the average information entropy is squared. The average of the squares of all average information entropies in the length of the current segment is calculated and square rooted to obtain the segment suitability of the current segment length.

[0080] Specifically, the segmented applicability satisfies the following relationship:

[0081] ;

[0082] in, Indicates that the current segment length is Yes, segmented applicability. Indicates the length of the current segment. Indicates the The information entropy of the data sequence, Indicates the Information entropy of gradient sequence.

[0083] It should be noted that information entropy measures the complexity of a data sequence; lower values ​​indicate more stable data and lower complexity. Therefore, when the segmentation suitability value is small, it means that at that segment length, the average information entropy of the data sequence and its gradient sequence is low, indicating that the data is relatively stable and has low complexity. This indicates that the current segment length can better reflect the data characteristics and that the segmentation is reasonable.

[0084] S3: Based on the Mahalanobis distance of each segment and the correlation with the sintering quality, the abnormal reference degree of each segment is analyzed, and weighted based on the abnormal reference degree to obtain the overall abnormal degree of each temperature curve.

[0085] Take any segment as the reference segment, calculate the average Mahalanobis distance of the reference segment, use a negative exponential function to perform exponential decay on the average Mahalanobis distance of all temperature curves corresponding to the reference segment, and use the result after subtracting the decay from 1 as the adjustment factor;

[0086] The Mahalanobis distance deviation of the reference segment is obtained by calculating the difference between the average Mahalanobis distance of the reference segment and the average Mahalanobis distance of all temperature curves corresponding to the reference segment.

[0087] Calculate the difference between the sintering quality of each temperature curve and the sintering quality of all temperature curves to obtain the sintering quality deviation;

[0088] The correlation coefficient was obtained by calculating the average value of the product between the Mahalanobis distance deviation and the sintering quality deviation and dividing it by the product of the standard deviation of the Mahalanobis distance of the reference segment and the standard deviation of the sintering quality.

[0089] The adjustment factor is multiplied by the correlation coefficient to obtain the abnormal reference degree of the reference segment.

[0090] It should be noted that for each segment of the discrete data of the temperature curve, its Mahalanobis distance represents the possibility of anomaly, and then weighted according to the anomaly reference degree is used to obtain the overall anomaly degree of each temperature curve.

[0091] The ratio of the abnormal reference degree of each segment to the sum of the abnormal reference degrees of all segments of the corresponding temperature curve is used for normalization processing.

[0092] Specifically, the abnormal reference degree satisfies the following relationship:

[0093] ;

[0094] in, Indicates the The abnormal reference degree of each segment, The first part of the whole temperature curve The average of the segmented Mahalanobis distances, Indicates the The temperature curve The average Mahalanobis distance of the segments, Indicates the The sintering quality of the temperature curve, It represents the average value of the sintering quality corresponding to all temperature curves. Indicates the The standard deviation of the Mahalanobis distance of each segment, represents the standard deviation of sintering quality, Indicates the total number of temperature curves, Represented by natural numbers An exponential function with base .

[0095] That is, the formula Part indicates the The correlation between segmentation and sintering quality is that the higher the correlation, the more significant the effect of the change of Mahalanobis distance on the sintering quality. Therefore, these segments are more likely to contain abnormal conditions that have a significant impact on the sintering quality, and thus should have a higher abnormal reference degree. It plays the role of adjusting the abnormal reference degree. When the average Mahalanobis distance of the segment is large, it indicates that the temperature change of the segment is more complicated and there may be more abnormal conditions. The value of is closer to 1, thereby improving the abnormal reference degree of the segments with higher abnormality, ensuring that the segments with higher abnormality can obtain higher abnormal reference degrees, making these segments easier to identify and focus on in subsequent analysis.

[0096] Taking any temperature curve as the curve to be analyzed, the product sum of the abnormal reference degree of each segment of the curve to be analyzed and the average Mahalanobis distance of the corresponding segment is calculated respectively to obtain the overall abnormal degree of the curve to be analyzed.

[0097] Specifically, the overall abnormality degree satisfies the following relationship:

[0098] ;

[0099] Where, Indicates the The overall abnormality of the temperature curve, Indicates the number of segments of the temperature curve, Indicates the The abnormal reference degree of each segment, Indicates the The temperature curve The average Mahalanobis distance of the segments.

[0100] S4: If the overall abnormality level is less than or equal to the abnormality level threshold, then there is no abnormality in the sintering process. On the contrary, if it is greater than the abnormality level threshold, then it is considered that there is an abnormal temperature curve in the sintering process, and then the staff is warned to adjust and control the sintering process.

[0101] For example, the abnormality threshold is , collect and analyze the temperature curve data during the sintering process, calculate the average value and standard deviation of the overall abnormality of the temperature curve during the normal sintering process, select the confidence interval coefficient, verify through experiments, and repeatedly adjust and confirm that the abnormality threshold can effectively distinguish between normal and abnormal sintering processes. It can also be adjusted according to specific circumstances.

[0102] The present invention also provides an intelligent control system for the sintering process based on powder metallurgy metal materials. Figure 2 As shown, the system includes a processor and a memory. The memory stores computer program instructions. When executed by the processor, the computer program instructions implement the intelligent control method for a sintering process based on powder metallurgy metal materials according to the first aspect of the present invention. The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are well known in the art and are not described in detail here.

[0103] It should be noted that those skilled in the art may make various modifications and improvements without departing from the scope of the present invention, and these modifications and improvements fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be based on the appended claims.

Claims

1. An intelligent control method for the sintering process of powder metallurgy metal materials, characterized in that: include: Obtain the temperature curve during powder metallurgy sintering; Determine the offset of each temperature curve based on the cross-correlation function, perform offset alignment on the curves, and determine the segment length based on the segment suitability of the local area of ​​each temperature curve; Based on the correlation between the Mahalanobis distance of each segment and the sintering quality, the abnormal reference degree of each segment is analyzed, and weighted based on the abnormal reference degree to obtain the overall abnormal degree of each temperature curve; If the overall abnormality level is less than or equal to the abnormality threshold, then there is no abnormality in the sintering process. On the contrary, if it is greater than the abnormality threshold, then it is considered that there is an abnormal temperature curve in the sintering process, and the staff is warned to adjust and control the sintering process. The calculation method of the abnormal reference degree includes: taking any segment as the reference segment, calculating the average Mahalanobis distance of the reference segment, using a negative exponential function to perform exponential decay on the average Mahalanobis distance of all temperature curves corresponding to the reference segment, and subtracting the decay from 1 as the adjustment factor; Calculate the difference between the average Mahalanobis distance of the reference segment and the average Mahalanobis distance of all temperature curves corresponding to the reference segment to obtain the Mahalanobis distance deviation of the reference segment; calculate the difference between the sintering quality of each temperature curve and the sintering quality of all temperature curves to obtain the sintering quality deviation; calculate the average value of the product between the Mahalanobis distance deviation and the sintering quality deviation, and divide it by the product of the standard deviation of the Mahalanobis distance of the reference segment and the standard deviation of the sintering quality to obtain the correlation coefficient; multiply the adjustment factor by the correlation coefficient as the abnormal reference degree of the reference segment.

2. The intelligent control method for sintering process based on powder metallurgy metal materials according to claim 1, characterized in that: The offset alignment of the curves comprises the following steps: Taking the first temperature change curve as the reference curve, calculate the cross-correlation function between each temperature change curve and the reference curve, obtain the offset value corresponding to the extreme value of the cross-correlation function of each temperature change curve, and take the difference between the offset value corresponding to the extreme value of the cross-correlation function and the average value of all offset values ​​as the actual offset value of each temperature curve; According to the actual offset value of each temperature curve, each temperature curve is corrected horizontally to align all temperature curves in time.

3. The intelligent control method for sintering process based on powder metallurgy metal materials according to claim 2, characterized in that: The offset alignment of the curves further includes: Discretize the corrected temperature curve, and based on the discretization result, select the temperature curve with the largest number of discrete data points as the reference benchmark for the unified number of data points; For temperature curves with fewer discrete data points than the reference benchmark, perform difference supplementation, starting from the last discrete data point of the temperature curve and supplementing data points in sequence until the number of discrete data points of the temperature curve reaches the reference benchmark; For temperature curves with more discrete data points than the reference benchmark, truncation processing is performed. Starting from the last discrete data point of the temperature curve, the redundant data points are removed in sequence until the number of discrete data points of the temperature curve is reduced to the reference benchmark; the number of discrete data points of each temperature curve is unified.

4. The intelligent control method for sintering process based on powder metallurgy metal materials according to claim 1, characterized in that: Obtaining the segment length includes the steps of: Preset the initial length of the segment and obtain the discrete data points of the initial length in the discrete data for each temperature curve; Traverse each temperature curve separately, divide each temperature curve according to the preset initial segment length, take any discrete data point as the target data, extract the discrete data point corresponding to the position of each target data from the current segment of the discrete data of all temperature curves, and construct the data sequence corresponding to the target data position of each temperature curve under the initial segment length; Calculate the gradient between adjacent data points in the current segment of each temperature curve to obtain a gradient sequence, take any gradient data as target gradient data, and obtain a target gradient data sequence at a position corresponding to the target gradient data; According to the information entropy of the data sequence at the corresponding position of the target data and the information entropy of the target gradient data sequence, the segmental fitness calculation formula is used to obtain the segmental fitness of each segment; In response to the segmentation fitness being greater than the preset segmentation threshold, the current segmentation length is increased by one, the above steps are iterated, and the segmentation fitness is recalculated until the segmentation fitness is less than or equal to the preset segmentation threshold, then the current segmentation length is reduced by one, and one segmentation is completed; The remaining data is iterated according to the initial segment length until all data are segmented.

5. The intelligent control method for sintering process based on powder metallurgy metal materials according to claim 4, characterized in that: The calculation method of the segmented applicability includes the following steps: The average of the sum of the information entropy of the data sequence at the corresponding position of the target data and the information entropy of the target gradient sequence is taken as the average information entropy, and the average information entropy is squared. The average of the squares of all average information entropies in the length of the current segment is calculated and square rooted to obtain the segment suitability of the current segment length.

6. The intelligent control method for sintering process based on powder metallurgy metal materials according to claim 1, characterized in that: The calculation method of the overall abnormality degree includes: Taking any temperature curve as the curve to be analyzed, the product sum of the abnormal reference degree of each segment of the curve to be analyzed and the average Mahalanobis distance of the corresponding segment is calculated respectively to obtain the overall abnormal degree of the curve to be analyzed.

7. The intelligent control system for the sintering process of powder metallurgy metal materials is characterized by: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent control method for the sintering process based on powder metallurgy metal materials according to any one of claims 1 to 6 is implemented.

Images