EIGA pulverizing collaborative control method and system based on multi-source data fusion

Abstract

The present application relates to the technical field of milling control system, and more particularly to a kind of EIGA milling collaborative control method and system based on multi-source data fusion, wherein the EIGA milling collaborative control method based on multi-source data fusion comprises: obtaining the sensor monitoring data of each sensor in EIGA milling process;For current process, extract the change characteristics of each sensor monitoring value;Determine the process fluctuation resistance factor of current process;Obtain the change information of upstream process key parameters, determine the production adaptation degree of current process;Determine the collaborative control compensation, and optimize the control parameter of current process based on collaborative control compensation.The present application extracts the change characteristics of each sensor monitoring value to represent the data fluctuation state under different sampling rates, eliminates the time dislocation between each dimension inside the feature vector, so that control decision can be based on the data reflecting the current real milling state, and improves the accuracy and reliability of collaborative control.

Description

Technical Field

[0001] This invention relates to the field of pulverizing control system technology, specifically to an EIGA pulverizing collaborative control method and system based on multi-source data fusion. Background Technology

[0002] EIGA (Electrode Induction Melting Gas Atomization) powder production technology is a key process for preparing high-quality metal powders. Its production process involves multiple interconnected steps, including electrode feeding, induction melting, gas atomization, vacuum environment control, and powder collection. To achieve end-to-end state awareness and coordinated control, existing powder production processes typically deploy various types of sensors at each of these steps to monitor physical quantities such as temperature, pressure, electrode displacement, oxygen concentration, and particle size distribution in real time. Due to the fundamental differences in the dynamic response characteristics of different physical quantities, the sampling frequencies of various sensors also exhibit significant variations. For example, electrode displacement monitoring can reach the hundreds of hertz range, while oxygen concentration monitoring may only be a few tenths of a hertz. When constructing a coordinated control strategy based on multi-source data fusion, it is necessary to uniformly process and assess the state of the aforementioned heterogeneous sensor data.

[0003] However, existing EIGA pulverizing control systems generally employ a forced time alignment approach when fusing multi-source data. This involves unifying all sensor data onto the same time axis and directly constructing a feature vector, which is then used to determine and decide on the current pulverizing state. Since the current sampled value of a slowly varying sensor actually reflects the historical state from several seconds ago, while rapidly varying data has been updated multiple times within that time window, this forced alignment method results in significant time misalignment between different dimensions within the feature vector. Control decisions based on this misaligned feature vector cannot accurately describe the current true pulverizing state, easily leading to misjudgments (such as underreporting sudden increases in oxygen concentration or misjudging particle size trends), thus limiting the accuracy and reliability of collaborative control. Summary of the Invention

[0004] This invention provides a collaborative control method and system for EIGA powder production based on multi-source data fusion to solve the above-mentioned problems.

[0005] The present invention provides a collaborative control method and system for EIGA powder production based on multi-source data fusion, which adopts the following technical solution: In a first aspect, one embodiment of the present invention provides an EIGA pulverizing collaborative control method based on multi-source data fusion. The method includes: acquiring sensor monitoring data from various sensors in the EIGA pulverizing process; extracting variation characteristics of the monitoring values ​​of each sensor for the current process; wherein the variation characteristics are used to characterize the data fluctuation state of each sensor under different sampling rates; determining a process fluctuation resistance factor for the current process based on the variation characteristics; wherein the process fluctuation resistance factor is used to quantify the resistance capability of the current process to parameter fluctuations; acquiring variation information of key parameters in upstream processes, and determining the production adaptability of the current process based on the variation information and the process fluctuation resistance factor; wherein the production adaptability is used to quantify the rhythm matching degree between upstream and downstream processes; determining a collaborative control compensation amount based on the production adaptability, and optimizing the control parameters of the current process based on the collaborative control compensation amount.

[0006] Furthermore, the step of extracting the change characteristics of each sensor monitoring value for the current process includes: maintaining a cache queue for each sensor in the current equipment group corresponding to the current process; wherein the cache queue is used to store the historical sampling values ​​of each sensor and their corresponding timestamps; determining the rate of change of each sensor monitoring value based on the difference between the current sampling value and the previous sampling value in the cache queue and the corresponding sampling time interval; and determining the range of each sensor monitoring value based on the maximum and minimum sampling values ​​in the cache queue.

[0007] Furthermore, determining the process fluctuation resistance factor of the current process based on the change characteristics includes: determining the comprehensive difference degree of the monitoring data of each sensor in the current equipment group based on the change rate and the range of each sensor; wherein the comprehensive difference degree is used to quantify the overall fluctuation difference degree of the monitoring data of each sensor in the current equipment group; and determining the process fluctuation resistance factor of the current process based on the comprehensive difference degree.

[0008] Further, determining the comprehensive difference degree of the monitoring data of each sensor in the current device group based on the rate of change and the range of each sensor includes: for each sensor in the current device group, calculating the difference between the monitoring value of the sensor and the monitoring values ​​of other sensors of the same attribute at the current time and the previous time, respectively, to obtain the monitoring value difference value; determining the mean value of the monitoring values ​​of sensors of the same attribute and the maximum monitoring difference value of sensors of the same attribute; determining the fluctuation reference range of each sensor based on the monitoring value difference value, the mean value of the monitoring values ​​of sensors of the same attribute, and the maximum monitoring difference value; and determining the comprehensive difference degree of the monitoring data of each sensor in the current device group based on the fluctuation reference range, the rate of change, and the range.

[0009] Further, determining the process fluctuation resistance factor of the current process based on the comprehensive difference degree includes: for each sensor in the current equipment group, determining the average monitoring value of the sensor and obtaining the fluctuation benchmark value of the sensor; calculating the difference between the monitoring values ​​of the sensor and other sensors of the same attribute at the current time and the previous time to obtain the monitoring value difference value; determining the maximum monitoring difference value of the sensors of the same attribute and the extreme value range of the monitoring values ​​of the sensors of the same attribute; determining the fluctuation benchmark range of each sensor based on the fluctuation benchmark value, the extreme value range, and the ratio of the maximum monitoring difference value to the monitoring value difference value; and determining the process fluctuation resistance factor of the current process based on the maximum and minimum values ​​among the fluctuation benchmark ranges of all sensors in the current equipment group and the comprehensive difference degree.

[0010] Furthermore, the step of obtaining the change information of key parameters of the upstream process includes: determining the parameter change rate based on the sampled value of the key parameter of the upstream process at the current moment and the sampled value at the previous moment; determining the change direction of the key parameter of the upstream process; and determining the fluctuation amplitude ratio based on the parameter change rate and the parameter change rate of the key parameter of the upstream process in the previous production process.

[0011] Further, determining the production adaptability of the current process based on the change information and the process fluctuation resistance factor includes: determining the expected absorption capacity index of the current process based on the parameter change rate, the change direction, the fluctuation amplitude ratio, and the process fluctuation resistance factor; and determining the production adaptability of the current process based on the expected absorption capacity index and the parameter change rate.

[0012] Furthermore, the step of determining the collaborative control compensation amount based on the production adaptability and optimizing the control parameters of the current process based on the collaborative control compensation amount includes: determining the cumulative mismatch deviation based on the production adaptability of the current process and the production adaptability of historical processes; determining the mismatch correction degree based on the cumulative mismatch deviation and the recent mismatch degree change trend; determining the collaborative control compensation amount based on the mismatch correction degree; and optimizing the control parameters of the current process based on the collaborative control compensation amount.

[0013] Further, determining the collaborative control compensation amount based on the mismatch correction degree includes: determining a basic proportional compensation amount based on the mismatch correction degree; determining a directional compensation amount based on the change direction of key parameters in the upstream process and the mismatch correction degree; determining a cumulative deviation compensation amount based on the cumulative mismatch deviation; and determining a collaborative control compensation amount based on the basic proportional compensation amount, the directional compensation amount, and the cumulative deviation compensation amount.

[0014] Secondly, another embodiment of the present invention provides an EIGA pulverizing collaborative control system based on multi-source data fusion, including a server and sensors and actuators respectively communicatively connected to the server, wherein: The sensor is used to acquire monitoring data in the EIGA powdering process and send the sensor monitoring data to the server; The actuator is used to receive optimized control parameters sent by the server and adjust the production parameters of the current process based on the optimized control parameters; The server is configured to acquire sensor monitoring data sent by the sensors; extract the change characteristics of each sensor monitoring value for the current process; wherein the change characteristics are used to characterize the data fluctuation state of each sensor under different sampling rates; determine the process fluctuation resistance factor of the current process based on the change characteristics; wherein the process fluctuation resistance factor is used to quantify the resistance capability of the current process to parameter fluctuations; acquire the change information of key parameters of upstream processes, and determine the production adaptability of the current process based on the change information and the process fluctuation resistance factor; wherein the production adaptability is used to quantify the rhythm matching degree between upstream and downstream processes; determine the collaborative control compensation amount based on the production adaptability, and optimize the control parameters of the current process based on the collaborative control compensation amount.

[0015] The beneficial effects of the technical solution of the present invention are: In this embodiment of the invention, sensor monitoring data of each sensor in the EIGA pulverization process are acquired; for the current process, the variation characteristics of the monitored values ​​of each sensor are extracted; wherein, the variation characteristics are used to characterize the data fluctuation state of each sensor under different sampling rates; based on the variation characteristics, the process fluctuation resistance factor of the current process is determined; wherein, the process fluctuation resistance factor is used to quantify the resistance capability of the current process to parameter fluctuations; the variation information of key parameters of upstream processes is acquired, and based on the variation information and the process fluctuation resistance factor, the production adaptability of the current process is determined; wherein, the production adaptability is used to quantify the rhythm matching degree between upstream and downstream processes; based on the production adaptability, the collaborative control compensation amount is determined, and the control parameters of the current process are optimized based on the collaborative control compensation amount.

[0016] This invention, by extracting the variation characteristics of each sensor's monitored values ​​to characterize data fluctuations at different sampling rates, replaces the existing method of forcibly aligning all sensor data to the same time axis. This eliminates temporal misalignment between dimensions within the feature vector, enabling control decisions to be based on data reflecting the current actual pulverizing state, thereby improving the accuracy and reliability of multi-source data fusion and collaborative control. Furthermore, by determining the process fluctuation resistance factor of the current process to quantify its resistance to parameter fluctuations, the tolerance ranges of different process stages to disturbances are differentiated, providing a quantitative basis for the transmission and absorption of disturbances between upstream and downstream processes. This provides a basis for understanding the underlying processes, thus avoiding chain oscillations caused by independent adjustments in a single stage. Furthermore, by determining the production fit based on changes in key parameters of upstream processes and the process fluctuation resistance factor of the current process, a quantitative assessment of the rhythm matching degree between upstream and downstream processes is achieved. This allows downstream processes to anticipate upstream disturbances and calculate pre-adjustment amounts, thereby achieving smooth adaptive adjustment of the overall production rhythm. Additionally, by determining the cumulative mismatch deviation and mismatch correction degree based on the production fit of current and historical processes, the drift accumulation of process fit status can be identified during long-term continuous pulverizing operations, triggering timely corrective actions to enhance the sensitivity to minor state deviations. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart of the EIGA powder preparation collaborative control method based on multi-source data fusion provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the architecture of the EIGA powder-making collaborative control system based on multi-source data fusion provided in an embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following detailed description, in conjunction with the accompanying drawings and preferred embodiments, provides a detailed account of the specific implementation methods, structures, features, and effects of the method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] The following description, in conjunction with the accompanying drawings, details the specific scheme of the EIGA powder-making collaborative control method and system based on multi-source data fusion provided by this invention.

[0021] like Figure 1 As shown, this embodiment of the invention provides a collaborative control method for EIGA powder production based on multi-source data fusion, comprising: Step S110: Obtain sensor monitoring data from each sensor in the EIGA powder preparation process.

[0022] In the EIGA pulverizing process, each production device is equipped with sensors for sensing the status of the production process. Each sensor transmits its collected monitoring data along with the corresponding timestamp to the server that executes the EIGA pulverizing collaborative control method based on multi-source data fusion. The server groups the sensors according to a preset sensor installation location table and assigns the monitoring data collected by each sensor and its timestamp to the corresponding device group to complete the initial acquisition and grouping storage of multi-source sensor monitoring data.

[0023] It should be noted that the above equipment groups are divided according to the process steps in which the sensors are deployed in the EIGA pulverization process. Each equipment group corresponds to a specific pulverization process, and all sensors deployed in the same process are grouped into the same equipment group so that the sensor monitoring data in each equipment group can centrally reflect the operating status of the corresponding process.

[0024] Step S120: For the current process, extract the change characteristics of the monitored values ​​of each sensor; wherein, the change characteristics are used to characterize the data fluctuation state of each sensor under different sampling rates.

[0025] It is understandable that forcibly aligning all sensor monitoring data to the same time axis to construct a feature vector would result in significant temporal misalignment between different dimensions within the feature vector. Control decisions based on this misaligned vector cannot accurately describe the current true powder-making state. Therefore, for each sensor, step S120 extracts the variation characteristics of its monitored values ​​to characterize the data fluctuation state of each sensor under different sampling rates. This provides information reflecting the true dynamic changes of each sensor for subsequent state assessment without relying on forced time alignment.

[0026] Optionally, step S120 includes: maintaining a cache queue for each sensor in the current equipment group corresponding to the current process; wherein the cache queue is used to store the historical sampling values ​​of each sensor and their corresponding timestamps; determining the rate of change of the monitored values ​​of each sensor based on the difference between the current sample value and the previous sample value in the cache queue and the corresponding sampling time interval; and determining the range of the monitored values ​​of each sensor based on the maximum and minimum sample values ​​in the cache queue.

[0027] In the implementation of the above scheme, a buffer queue can be maintained for each sensor. The buffer queue stores the historical sampling values ​​of the sensor in the most recent sampling and their corresponding timestamps. When the corresponding sensor monitoring value is needed in the future, the corresponding sampling data can be directly read from the buffer queue corresponding to the sensor.

[0028] The aforementioned rate of change of monitored values ​​is used to characterize the dynamic change trend of sensor monitored values ​​per unit time, and its calculation formula is as follows:

[0029] In the formula, For sensors The rate of change of the monitored values; For sensors At the present moment The sampled values; For sensors Previous sampling time The sampled values; This represents the time interval between two adjacent samples.

[0030] The range mentioned above is used to characterize the fluctuation range of sensor monitoring values ​​within the recent sampling window, and its calculation formula is as follows:

[0031] In the formula, For sensors The corresponding cache queue; and These represent the sensors. The maximum and minimum values ​​of all sampled values ​​in the corresponding cache queue.

[0032] The aforementioned rate of change and range are used to capture the dynamic change trend and recent fluctuation amplitude of the sensor, respectively, and can extract comparable state change information from data sources with different sampling rates without forcing a uniform time axis.

[0033] It is understood that the variation characteristics of the sensor monitoring values ​​mentioned above are not limited to the rate of change and range of the monitoring values, but may also include statistical characteristics such as variance, standard deviation, moving average, kurtosis or skewness. These statistical characteristics can reflect the fluctuation range or dynamic trend of the sensor monitoring values ​​within a specific time window.

[0034] Step S130: Based on the change characteristics, determine the process fluctuation resistance factor of the current process; wherein, the process fluctuation resistance factor is used to quantify the current process's ability to resist parameter fluctuations.

[0035] It is understandable that the powder objects and monitored data types processed at different stages of the EIGA powder preparation process are fundamentally different, resulting in varying sensitivities and tolerances to parameter fluctuations at each stage. If these stage-specific anti-disturbance characteristics are ignored and cross-process collaborative control is directly implemented, parameter fluctuations in upstream processes may be over- or under-responded to by downstream processes, leading to chain oscillations caused by independent adjustments in a single link. This invention, based on the changing characteristics of the monitored data from each sensor, determines the process fluctuation resistance factor for the current process. This quantifies the tolerance capability of the current process to allow fluctuations in various monitored quantities while maintaining powder quality and process stability. This provides a staged quantitative basis for assessing disturbance transmission between upstream and downstream processes and calculating collaborative control compensation, enabling downstream processes to make differentiated pre-adjustment responses based on the actual impact of upstream disturbances.

[0036] Optionally, step S130 includes: determining the overall difference of the monitoring data of each sensor in the current equipment group based on the rate of change and range of each sensor; wherein the overall difference is used to quantify the overall fluctuation difference of the monitoring data of each sensor in the current equipment group; and determining the process fluctuation resistance factor of the current process based on the overall difference.

[0037] Currently, due to differences in the physical quantities monitored and sampling rates of the sensors within the current equipment group, their monitoring data exhibit individualized fluctuation characteristics. If the fluctuation status of a single sensor is used to assess the fluctuation resistance capability of the current process, it cannot reflect the overall data fluctuation difference level of the process under multi-source data fusion conditions. This invention quantifies the overall fluctuation difference degree of the monitoring data from each sensor within the current equipment group through the aforementioned comprehensive difference measure, thereby providing a basic input characterizing the overall data fluctuation characteristics of the process for evaluating the process fluctuation resistance factor. This allows the process fluctuation resistance factor to be determined based on the multi-source data difference level at the process level. A scheme for determining the comprehensive difference degree is described below: Optionally, the above-mentioned determination of the comprehensive difference of the monitoring data of each sensor in the current equipment group based on the rate of change and range of each sensor includes: for each sensor in the current equipment group, calculating the difference between the monitoring value of the sensor and the monitoring values ​​of other sensors of the same attribute at the current time and the previous time, and obtaining the monitoring value difference value; determining the mean value of the monitoring values ​​of sensors of the same attribute and the maximum monitoring difference value of sensors of the same attribute; determining the fluctuation reference range of each sensor based on the monitoring value difference value, the mean value of the monitoring values ​​of sensors of the same attribute, and the maximum monitoring difference value; and determining the comprehensive difference of the monitoring data of each sensor in the current equipment group based on the fluctuation reference range, the rate of change, and the range.

[0038] The aforementioned attributes refer to the type and category of physical quantities monitored by the sensors. For example, temperature sensors, pressure sensors, and speed sensors correspond to different physical quantity monitoring attributes such as temperature, pressure, and speed, respectively. Sensors with the same attribute refer to a group of sensors within the same equipment group that monitor the same physical quantity of the same type and perform monitoring of the same physical quantity in the same process step. For example, all temperature sensors in the current equipment group belong to the same attribute category. Their monitoring data are used to reflect the state distribution of that process step in the corresponding physical quantity dimension, thus providing a basis for horizontal comparison between monitoring data of the same type to determine the fluctuation reference range.

[0039] For each sensor in the current equipment group, the difference between the monitored values ​​of this sensor and all other sensors of the same attribute at the current time and the previous time is calculated to obtain the monitoring value difference value. The average monitoring value of the sensor group of the same attribute to which this sensor belongs is determined as the fluctuation benchmark value, and the maximum monitoring difference value of this sensor group is determined. Furthermore, the extreme value range of the monitoring values ​​of this sensor group is determined. Based on the fluctuation benchmark value, the extreme value range, and the ratio of the maximum monitoring difference value to the aforementioned monitoring value difference value, the fluctuation benchmark range of each sensor is determined. The formula for calculating the fluctuation benchmark range is:

[0040] In the formula, For sensors The average value of the monitored values; The maximum and minimum range of values ​​monitored by sensors with the same attribute. This represents the maximum value among the other sensor readings of the same attribute at the current moment. This represents the minimum value among all other sensor readings of the same attribute at the current moment. and The differences quantify the distribution span of the monitoring values ​​of a group of sensors with the same attributes at the current moment; This represents the maximum difference in monitoring values ​​between sensors with the same properties. This represents the difference between the current sensor's monitored values ​​and those of other sensors with the same properties.

[0041] Based on the established fluctuation baseline range for each sensor, the comprehensive difference in monitoring data from each sensor within the current equipment group is further determined based on the rate of change and range of each sensor. The formula for calculating the comprehensive difference is as follows:

[0042] In the formula, For the current equipment group The total number of all sensors in the system; This represents the average number of sensors in each equipment group across all current process groups. For the current equipment group Medium sensor The number of sensors belonging to the attribute category; For sensors The absolute value of the rate of change of the monitored values; For sensors The range. In the above formula for calculating the overall difference, It can reflect the overall scale and relative level of the current process; Iterate through the fluctuation data of all sensors within the current device group, and sort the attribute value fluctuations by the number of sensors for that attribute. Suppression is implemented to ensure the stability of monitoring data of different attributes when compared horizontally.

[0043] It should be noted that the above-mentioned comprehensive difference determination scheme is not the only scheme for determining the comprehensive difference. The embodiments of the present invention can also use one of the following schemes to determine the comprehensive difference: (1) Replace the range in the above scheme with the standard deviation or mean absolute deviation of the buffer queue data to measure the statistical dispersion of the sensor monitoring values ​​and the average deviation of each sample value from the mean, and then multiply it with the rate of change and normalize it according to the number of attribute sensors to obtain the comprehensive difference. (2) Use the weighted sum of the rate of change and the range (or standard deviation), and perform linear aggregation by setting the weight coefficients of the dynamic change trend and the static fluctuation amplitude respectively, and then multiply it by the equipment group size factor to obtain the comprehensive difference.

[0044] Optionally, the above-mentioned determination of the process fluctuation resistance factor for the current process based on the comprehensive difference degree includes: for each sensor in the current equipment group, determining the average value of the sensor's monitoring value and obtaining the fluctuation benchmark value of the sensor; calculating the difference between the monitoring values ​​of the sensor and other sensors of the same attribute at the current time and the previous time, respectively, to obtain the monitoring value difference value; determining the maximum value of the monitoring difference of sensors of the same attribute and the extreme value range of the monitoring values ​​of sensors of the same attribute; determining the fluctuation benchmark range of each sensor based on the fluctuation benchmark value, the extreme value range, and the ratio of the maximum value of the monitoring difference to the monitoring value difference value; and determining the process fluctuation resistance factor for the current process based on the maximum and minimum values ​​among the fluctuation benchmark ranges of all sensors in the current equipment group and the comprehensive difference degree.

[0045] For each sensor in the current equipment group, the average monitoring value of that sensor at the current time and the previous time is used as the fluctuation baseline value. This fluctuation baseline value represents the center of the monitoring level of that sensor at adjacent sampling times. The differences between the monitoring values ​​of this sensor and all other sensors of the same attribute at the current time and the previous time are calculated to obtain the monitoring value difference value. This difference value quantifies the degree of individual deviation between the current sensor and the group of sensors of the same attribute. The maximum monitoring difference value among all monitoring value differences of sensors of the same attribute at the current time and the previous time is determined. This maximum monitoring difference value represents the extreme difference level exhibited by the individual with the greatest deviation in the group of sensors of the same attribute. The extreme value range of the monitoring values ​​of sensors of the same attribute is determined. This extreme value range is determined by the difference between the maximum and minimum values ​​of the monitoring values ​​of sensors of the same attribute, quantifying the distribution span of the monitoring values ​​of the group of sensors of the same attribute at the current time.

[0046] Based on the maximum and minimum values ​​and the overall difference among all sensors within the current equipment group's fluctuation reference range, the process fluctuation resistance factor for the current process is determined. The formula for calculating the process fluctuation resistance factor is:

[0047] In the formula, This is the maximum value among all the fluctuation reference ranges of all sensors in the current equipment group; This is the minimum value among all the fluctuation reference ranges of all sensors in the current equipment group; and The difference can characterize the range of variation of the monitoring data processed in the current process as a whole; The overall difference in monitoring data from various sensors within the current equipment group; For the current process (or the current equipment group) ( ) process fluctuation resistance factor.

[0048] Step S140: Obtain the change information of key parameters of upstream processes, and determine the production adaptability of the current process based on the change information and the process fluctuation resistance factor; wherein, the production adaptability is used to quantify the degree of rhythm matching between upstream and downstream processes.

[0049] It is understandable that in the continuous production process of EIGA powder preparation, there is a close parameter coupling and disturbance transmission relationship between upstream and downstream processes. Changes in the output parameters of the upstream process will generate expected disturbances in the downstream process. However, due to the differences in the powder objects and physical properties processed, the tolerance and absorption capacity of parameter fluctuations vary from process stage to process stage. If this cross-process disturbance transmission characteristic and the stage-differentiated anti-disturbance capability are ignored and single-stage independent adjustment is performed, it is easy to cause the downstream process to over-respond or under-respond to the upstream disturbance, thus leading to chain oscillations. The embodiments of the present invention obtain the change information of key parameters of the upstream process and determine the production adaptability by combining the process fluctuation resistance factor of the current process, so as to quantify the rhythm matching degree between upstream and downstream processes, thereby judging whether the upstream disturbance can be smoothly absorbed by the current process, and calculating the pre-adjustment amount of the downstream actuator in advance, so as to achieve smooth adaptive adjustment of the overall production rhythm.

[0050] Optionally, the above-mentioned acquisition of change information of key parameters of upstream processes includes: determining the parameter change rate based on the sampled value of the key parameters of upstream processes at the current moment and the sampled value at the previous moment; determining the change direction of the key parameters of upstream processes; and determining the fluctuation amplitude ratio based on the parameter change rate and the parameter change rate of the key parameters of upstream processes in the previous production process.

[0051] In the EIGA powder production process, key parameters of upstream processes refer to the core process parameters output by the process steps preceding the current process that directly affect the operating status of downstream processes and can characterize the output characteristics of upstream processes. For example, melting power or electrode feed rate in the melting process. Changes in these parameters will directly cause changes in the state of materials received by downstream processes or process environment conditions. Therefore, they can be extracted as change information to support the quantitative assessment of the magnitude and trend of upstream disturbances by downstream processes.

[0052] The rate of change of key parameters in upstream processes is used to quantify the relative degree of change of the parameter between adjacent sampling times, and its calculation formula is as follows:

[0053] In the formula, These are the current sampled values ​​of key parameters in the upstream process; These are the sampled values ​​of key parameters from the upstream process at the previous moment; The parameter change rate of the key parameters in the upstream process can characterize the relative change of the upstream parameters relative to the previous state.

[0054] The direction of change of key parameters in upstream processes is used to characterize the trend polarity of the parameter's change, which is determined by a sign function:

[0055] In the formula, The value can be +1, -1, or 0, corresponding to the three trends of the upstream parameter: rising, falling, or remaining flat, respectively.

[0056] The volatility ratio measures the degree of deviation of current upstream parameter changes from historical production levels, and its calculation formula is as follows:

[0057] In the formula, This refers to the rate of change of key parameters in the upstream process. This is the rate of change of the parameter in the previous production process, i.e., the historical baseline rate of change. The ratio of fluctuation amplitude, when A value greater than 1 indicates that the current disturbance exceeds historical norms, requiring enhanced downstream response; when A value less than or equal to 1 indicates that the current disturbance is within the historical normal range.

[0058] The following is a scheme for determining production adaptability: Optionally, the above-mentioned determination of the production adaptability of the current process based on change information and process fluctuation resistance factor includes: determining the expected absorption capacity index of the current process based on parameter change rate, change direction, fluctuation amplitude ratio and process fluctuation resistance factor; and determining the production adaptability of the current process based on the expected absorption capacity index and parameter change rate.

[0059] The aforementioned expected absorption capacity index is used to quantify the theoretical absorption level that the current process can provide against upstream disturbances based on its own and preceding process's resistance to process fluctuations. Its calculation formula is as follows:

[0060] In the formula, The average value of the process fluctuation resistance factor calculated for all processes up to the current process is used to characterize the overall resistance level of each process up to the current process to parameter fluctuations. The directional sensitivity coefficient is used to quantify the sensitivity of the current process to the trend type of upstream disturbances. The value ranges from 0 to 1 and can be configured differently according to the trend type of upstream disturbances. A larger value is assigned to disturbances with an upward trend, and a smaller value is assigned to disturbances with a downward trend or a stable state. This reflects the process characteristic that an increase in melt superheat is more likely to cause pulverization abnormalities than a decrease, so that the expected absorption capacity index can be adaptively adjusted according to the directional risk level of upstream disturbances. This indicates the direction of change for key parameters in the upstream process, with values ​​of +1, -1, or 0. absolute value The value can be 0 or 1, and is used to characterize whether there is a directional change in the upstream disturbance; The ratio of fluctuation amplitude; As an indicator of expected absorption capacity, in In the calculation formula, The direction and amplitude correction term is smaller when the upstream disturbance direction is upward and the fluctuation amplitude ratio is larger, so that the expected absorption capacity index is reduced accordingly, in order to reflect the sensitivity of the current process to abnormal upward disturbances.

[0061] Production fit is used to quantify the actual rhythm matching degree between upstream and downstream processes, and its calculation formula is as follows:

[0062] In the formula, This is an indicator of expected absorption capacity. This refers to the rate of change of key parameters in the upstream process. For production adaptability, the value ranges from 0 to 1. The closer the value is to 1, the better the matching of upstream and downstream rhythms. In the above production fit... In the calculation formula, when the expected absorption capacity index is much greater than the change rate of upstream parameters, the production adaptability approaches 1, indicating that the current process can fully absorb upstream disturbances; when the expected absorption capacity index is much less than the change rate of upstream parameters, the production adaptability approaches 0, indicating that the current process has difficulty in smoothly coping with upstream disturbances.

[0063] Step S150: Based on production adaptability, determine the collaborative control compensation amount, and optimize the control parameters of the current process based on the collaborative control compensation amount.

[0064] Optionally, step S150 includes: determining the cumulative mismatch deviation based on the production adaptability of the current process and the production adaptability of historical processes; determining the mismatch correction degree based on the cumulative mismatch deviation and the recent trend of mismatch degree changes; determining the collaborative control compensation amount based on the mismatch correction degree; and optimizing the control parameters of the current process based on the collaborative control compensation amount.

[0065] In continuous milling operations, the production fit deviation of each process accumulates over time. If adjustments are made independently segment by segment, it may lead to a systematic deviation. This invention introduces a cumulative mismatch deviation to quantify the cumulative deviation between the production fit of each process from the first process to the current process (or within a sliding window) and the ideal matching state. The calculation formula is as follows:

[0066] In the formula, For the first Production adaptability of each process step; This is the cumulative mismatch deviation, which is the sum of the mismatches of all processes from the first process to the current process; This represents the total number of processes from the first process to the current process. The cumulative mismatch deviation mentioned above... In the calculation formula, through the analysis of each process... The process involves accumulating and quantifying the mismatch states that were not fully absorbed from upstream disturbances in historical processes. The larger the value, the more serious the deviation of the overall process rhythm.

[0067] The recent adaptation trend is used to characterize the direction and magnitude of the change in the production adaptation of the current process relative to the previous process. Its calculation formula is as follows:

[0068] In the formula, For the production adaptability of the current process; The production compatibility of the previous process; express The sign can be +1, -1, or 0, corresponding to three trends: increasing fit, decreasing fit, or remaining flat, respectively.

[0069] The upstream volatility persistence index measures the stability of upstream parameter volatility in recent processes. Its calculation formula is as follows:

[0070] In the formula, For the most recent The upstream fluctuation range ratio corresponding to each process step; upstream fluctuation persistence index The larger the value, the more persistent the upstream disturbance is in time, and the more its influence on the current process has an inertial characteristic.

[0071] Decay weight Based on upstream volatility persistence indicators The calculation, used to adjust the contribution of recent fit trends to the correction calculation, is as follows:

[0072] In the formula, This is the persistence coefficient, used to adjust the influence of the upstream volatility persistence index on the attenuation weight, with a value ranging from 1 to 3. When the upstream volatility persistence index... At higher values, an increase in the denominator leads to A decrease in the correlation coefficient means that, under conditions of strong and persistent disturbances, the corrective effect of recent trends in fitness is weakened, thus avoiding overreaction to short-term fluctuations; conversely, when the upstream volatility persistence index... At lower levels, As the value approaches 1, the moderating effect of recent trends on correction is fully preserved. (Regarding the persistence coefficient...) When the value is large, the upstream volatility persistence indicator For decay weight The inhibitory effect is enhanced. As the cumulative mismatch bias decreases, its attenuating effect on the current correction degree intensifies; when When the value is small, the suppression effect weakens, and the sensitivity to the persistence index of upstream fluctuations decreases. This can be addressed by adjusting the persistence coefficient. The weighting of historical cumulative mismatch in the current correction calculation can be flexibly configured according to process characteristics to adapt to the correction needs under different fluctuation and persistence conditions.

[0073] The mismatch correction degree is used to assess the adequacy of the overall production process in correcting state drift, and serves as the basis for triggering early warnings and calculating collaborative control compensation. Its calculation formula is as follows:

[0074] In the formula, This is the cumulative attenuation coefficient, with a value ranging from 0.2 to 0.5; Production fit for the current process. Mismatch correction degree. The value ranges from 0 to 1, and its magnitude directly reflects the extent to which the current process flow has adequately corrected historical mismatches. The closer it is to 1, the better the current process is adapted to the current process, and the more it has the ability to correct and buffer against the cumulative disturbances from upstream. The lower the value, the more severe the cumulative mismatch and the insufficient current correction capability. (The above mismatch correction level...) In the calculation formula, the numerator term By coupling the recent trend of fitness changes with the decay weight, the current fitness is directionally corrected: the correction is increased when the fitness shows an improving trend, and decreased when the fitness shows a deteriorating trend; denominator term By accumulating mismatch deviations to suppress the attenuation of correction, the process segment with more severe historical mismatches can obtain a stronger characterization of correction requirements.

[0075] Optionally, the above-mentioned determination of the collaborative control compensation amount based on the mismatch correction degree includes: determining the basic proportional compensation amount based on the mismatch correction degree; determining the directional compensation amount based on the change direction of the key parameters of the upstream process and the mismatch correction degree; determining the cumulative deviation compensation amount based on the cumulative mismatch deviation; and determining the collaborative control compensation amount based on the basic proportional compensation amount, the directional compensation amount, and the cumulative deviation compensation amount.

[0076] The collaborative control compensation is used to convert the mismatch correction degree into specific adjustment commands for each actuator. It is composed of three parts: basic proportional compensation, directional compensation, and cumulative deviation compensation. Basic ratio compensation The calculation formula is:

[0077] In the formula, This is the proportional gain, with a value ranging from 0.2 to 0.5; This represents the degree of mismatch correction.

[0078] Directional compensation The calculation formula is:

[0079] In the formula, This represents the directional gain, with a value ranging from 0.1 to 0.3. This indicates the direction of change of key parameters in the upstream process, with values ​​of +1, -1, or 0, corresponding to an upward, downward, or flat trend, respectively.

[0080] Cumulative Deviation Compensation The calculation formula is:

[0081] In the formula, This is the integral gain, ranging from 0.05 to 0.15. This represents the cumulative mismatch bias.

[0082] Total collaborative compensation The algebraic sum of the three:

[0083] Under normal operating conditions, the total amount of coordinated compensation After limiting, the amplitude is constrained to between preset upper and lower limits. The limiting formula is:

[0084] In the formula, The maximum compensation is set at 0.3. To compensate for the lower limit, the value is set to -0.2.

[0085] The values ​​of the proportional gain, directional gain, and integral gain are differentiated based on the functional positioning and response characteristics of each component in the collaborative control compensation. The proportional gain ranges from 0.2 to 0.5 because its corresponding basic proportional compensation directly responds to the immediate degree of mismatch represented by the current mismatch correction and is the main contributor to the collaborative control compensation; therefore, it is given a relatively high gain weight to ensure a rapid response to the current mismatch state. The directional gain ranges from 0.1 to 0.3 because its corresponding directional compensation is only used to apply directional correction based on the upstream disturbance trend type. If the gain is too large, it may lead to oversensitivity to trend changes and trigger reverse oscillations; therefore, its value is lower than 0.3. The proportional gain is used to suppress the overreaction of directional correction; the integral gain ranges from 0.05 to 0.15, because its corresponding cumulative deviation compensation is used to continuously and slowly eliminate historical cumulative mismatch. If the gain is too large, it will cause integral saturation effect or cause the system to produce violent adjustment action when the cumulative deviation is large. Therefore, its value is the smallest among the three to achieve gradual correction of historical deviation. The values ​​of the above three gains are all limited to a range much less than 1, and after synthesis and limiting, the upper limit of the total compensation is 0.3 and the lower limit is -0.2, thereby ensuring that the optimization adjustment of the control parameters of multiple actuators is always within a safe and controllable adjustment range, avoiding actuator oscillation or process over-adjustment caused by excessive gain.

[0086] Based on the collaborative compensation amount after amplitude limiting Calculate the optimized control parameters for each actuator and output the corresponding control commands, where: Optimal value of electrode feed speed Calculate using the following formula:

[0087] In the formula, This represents the current electrode feed rate; Optimal value of atomizing air pressure Calculate using the following formula:

[0088] In the formula, This is the current atomizing pressure; Optimal value of induction power Calculate using the following formula:

[0089] In the formula, This represents the current induction power. The induction power and electrode feed speed are inversely related. When the cooperative compensation amount U is positive, the electrode feed speed increases while the induction power decreases to prevent the melt overheating from continuing to rise. The optimized control parameters are transmitted to the controllers of each actuator via the control bus, driving the actuators to complete the corresponding powder-making control operations.

[0090] Furthermore, when mismatch correction degree When the amplitude falls below a preset threshold (e.g., 0.2), the conventional limiting is insufficient to prevent over-adjustment. In this case, a safe gradual adjustment mode is forcibly entered, strictly limiting the amplitude of each step of the collaborative compensation to within the preset safe threshold range. This safety threshold is much smaller than the boundary value of the conventional limit, ensuring that the actuator can make parameter fine-tuning in very small steps within the high mismatch risk zone, avoiding system oscillation caused by excessive adjustment in a single step.

[0091] like Figure 2 As shown, based on the same inventive concept, this embodiment of the invention also provides an EIGA pulverizing collaborative control system 200 based on multi-source data fusion, including a server 210 and sensors 220 and actuators 230 respectively communicatively connected to the server 210, wherein: Sensor 220 is used to acquire monitoring data in the EIGA powder making process and send the sensor monitoring data to server 210; Actuator 230 is used to receive optimized control parameters sent by server 210 and adjust the production parameters of the current process based on the optimized control parameters; Server 210 is used to acquire sensor monitoring data sent by sensors; for the current process, it extracts the change characteristics of the monitoring values ​​of each sensor; the change characteristics are used to characterize the data fluctuation state of each sensor under different sampling rates; based on the change characteristics, it determines the process fluctuation resistance factor of the current process; the process fluctuation resistance factor is used to quantify the resistance of the current process to parameter fluctuations; it acquires the change information of key parameters of upstream processes, and based on the change information and the process fluctuation resistance factor, it determines the production adaptability of the current process; the production adaptability is used to quantify the rhythm matching degree between upstream and downstream processes; based on the production adaptability, it determines the collaborative control compensation amount, and optimizes the control parameters of the current process based on the collaborative control compensation amount.

[0092] It is understood that the server 210 described above can implement any one of the functions of the EIGA pulverizing collaborative control method based on multi-source data fusion provided in the embodiments of the present invention. The method implementation and working principle of each function can be found in the method embodiment section, and will not be repeated in the system embodiment section.

[0093] This invention is now complete.

[0094] In summary, in this embodiment of the invention, sensor monitoring data from each sensor in the EIGA pulverization process are acquired; for the current process, the variation characteristics of the monitored values ​​of each sensor are extracted; wherein, the variation characteristics are used to characterize the data fluctuation state of each sensor under different sampling rates; based on the variation characteristics, the process fluctuation resistance factor of the current process is determined; wherein, the process fluctuation resistance factor is used to quantify the resistance capability of the current process to parameter fluctuations; the variation information of key parameters in the upstream process is acquired, and based on the variation information and the process fluctuation resistance factor, the production adaptability of the current process is determined; wherein, the production adaptability is used to quantify the rhythm matching degree between upstream and downstream processes; based on the production adaptability, the collaborative control compensation amount is determined, and the control parameters of the current process are optimized based on the collaborative control compensation amount.

[0095] This invention extracts the variation characteristics of the monitored values ​​from each sensor to characterize the data fluctuation state under different sampling rates, replacing the existing technology's method of forcibly aligning all sensor data to the same time axis. This eliminates the time misalignment between different dimensions within the feature vector, enabling control decisions to be based on data reflecting the current actual pulverizing state, thereby improving the accuracy and reliability of multi-source data fusion and collaborative control. Furthermore, by determining the process fluctuation resistance factor of the current process to quantify the process's resistance to parameter fluctuations, the tolerance range of different process stages to disturbances is differentiated, providing a quantitative basis for the transmission and absorption of disturbances between upstream and downstream processes. Based on this, chain oscillations caused by independent adjustment of a single link are avoided; on the other hand, by determining the production adaptability based on the change information of key parameters of upstream processes and the process fluctuation resistance factor of the current process, the degree of rhythm matching between upstream and downstream processes is quantitatively assessed, enabling downstream processes to sense upstream disturbances in advance and calculate pre-adjustment amounts, thereby achieving smooth adaptive adjustment of the overall production rhythm; furthermore, by determining the cumulative mismatch deviation and mismatch correction degree based on the production adaptability of current and historical processes, the drift accumulation of process adaptability status can be identified in long-term continuous pulverization operations, and timely correction actions can be triggered to enhance the sensitivity to the identification of slight state deviations.

[0096] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A collaborative control method for EIGA milling based on multi-source data fusion, characterized in that, The method includes: Acquire sensor monitoring data from various sensors in the EIGA powder production process; For the current process, the variation characteristics of the monitored values ​​of each sensor are extracted; wherein, the variation characteristics are used to characterize the data fluctuation state of each sensor under different sampling rates, and the variation characteristics include the rate of change and range of the monitored values ​​of each sensor; Based on the aforementioned change characteristics, a process fluctuation resistance factor for the current process is determined; wherein, the process fluctuation resistance factor is used to quantify the current process's ability to resist parameter fluctuations; determining the process fluctuation resistance factor for the current process based on the aforementioned change characteristics includes: determining the comprehensive difference degree of the monitoring data of each sensor within the current equipment group corresponding to the current process based on the change rate and range of each sensor; wherein, the comprehensive difference degree is used to quantify the overall fluctuation difference degree of the monitoring data of each sensor within the current equipment group; for each sensor within the current equipment group, determining the average monitoring value of the sensor and obtaining the fluctuation baseline value of the sensor; calculating the difference between the monitoring values ​​of the sensor and other sensors of the same attribute at the current time and the previous time, respectively, to obtain the monitoring value difference value; determining the maximum monitoring difference value of the sensors of the same attribute and the extreme value range of the monitoring values ​​of the sensors of the same attribute; determining the fluctuation baseline value, the extreme value range, and the ratio of the maximum monitoring difference value to the maximum monitoring value difference value, determining the fluctuation baseline range of each sensor; determining the process fluctuation resistance factor for the current process based on the maximum and minimum values ​​among the fluctuation baseline ranges of all sensors within the current equipment group and the comprehensive difference degree; The process involves acquiring information on changes in key parameters of upstream processes, and determining the production adaptability of the current process based on this information and the process fluctuation resistance factor. The production adaptability quantifies the degree of rhythm matching between upstream and downstream processes. Acquiring information on changes in key parameters of upstream processes includes: determining the parameter change rate based on the sampled values ​​of the key parameters at the current moment and the sampled values ​​at the previous moment; determining the direction of change of the key parameters of the upstream process; and determining the fluctuation amplitude ratio based on the parameter change rate and the parameter change rate of the key parameters of the upstream process in the previous production process. The step of determining the production adaptability of the current process based on the change information and the process fluctuation resistance factor includes: determining the expected absorption capacity index of the current process based on the parameter change rate, the change direction, the fluctuation amplitude ratio, and the process fluctuation resistance factor; and determining the production adaptability of the current process based on the expected absorption capacity index and the parameter change rate. Based on the production adaptability, a collaborative control compensation amount is determined, and the control parameters of the current process are optimized based on the collaborative control compensation amount, including: determining the cumulative mismatch deviation based on the production adaptability of the current process and the production adaptability of historical processes; determining the mismatch correction degree based on the cumulative mismatch deviation and the recent mismatch degree change trend; determining the collaborative control compensation amount based on the mismatch correction degree; and optimizing the control parameters of the current process based on the collaborative control compensation amount.

2. The EIGA pulverizing collaborative control method based on multi-source data fusion according to claim 1, characterized in that, The methods for determining the rate of change and range of the monitored values ​​of each sensor include: For the current equipment group corresponding to the current process, maintain a cache queue for each sensor in the current equipment group; wherein, the cache queue is used to store the historical sampling values ​​of each sensor and their corresponding timestamps; Based on the difference between the current sampled value and the previous sampled value in the buffer queue and the corresponding sampling time interval, the rate of change of the monitored values ​​of each sensor is determined; The range of each sensor's monitored value is determined based on the maximum and minimum sampled values ​​in the buffer queue.

3. The EIGA pulverizing collaborative control method based on multi-source data fusion according to claim 1, characterized in that, The determination of the collaborative control compensation amount based on the mismatch correction degree includes: Based on the mismatch correction degree, determine the basic proportional compensation amount; Based on the direction of change of key parameters in upstream processes and the mismatch correction degree, the directional compensation amount is determined. Based on the cumulative mismatch deviation, determine the cumulative deviation compensation amount; Based on the basic proportional compensation amount, the directional compensation amount, and the cumulative deviation compensation amount, the collaborative control compensation amount is determined.

4. An EIGA pulverizing collaborative control system based on multi-source data fusion, characterized in that, It includes a server and sensors and actuators respectively communicatively connected to the server, wherein: The sensor is used to acquire monitoring data in the EIGA powdering process and send the sensor monitoring data to the server; The actuator is used to receive optimized control parameters sent by the server and adjust the production parameters of the current process based on the optimized control parameters; The server is used to execute the EIGA powder-making collaborative control method based on multi-source data fusion as described in any one of claims 1-3.

Images