CN-121998501-A - Workpiece quality on-line analysis and production decision optimization model

CN121998501ACN 121998501 ACN121998501 ACN 121998501ACN-121998501-A

Abstract

The invention discloses an on-line analysis and production decision optimization model for workpiece quality, which belongs to the field of industrial production management and comprises a spectrum generation module, a factor construction module and an instruction splitting module, wherein the spectrum generation module is used for generating a one-dimensional process energy density spectrum on line, the process energy density spectrum is formed by dynamically distributing weights for a plurality of process control variables and carrying out nonlinear superposition, the factor construction module is used for calculating stable factors of all production beat points based on the process energy density spectrum to construct a beat characteristic field, the stable factors are determined according to the deviation degree of a current spectrum value relative to a target standard, the spectrum value change gradient of an adjacent beat point and the state logic values of a downstream buffer zone and a downstream process, and the instruction splitting module is used for splitting an optimization decision instruction into a plurality of decision quantum packets. The method can realize accurate adaptation and dynamic resource allocation of decision instructions and production beats, avoid resource waste caused by the traditional fixed flow, improve the utilization efficiency of production resources such as equipment, materials and the like, reduce the production cost and enhance the flexibility of production management and control.

Inventors

LIN QIANG
WANG SIYUAN
LIN YILONG
ZHANG PEIXIANG
LIN CHAOWEI
HUANG GENGJIE

Assignees

福建科烨数控科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260126

Claims (10)

1. The workpiece quality online analysis and production decision optimization model is characterized by comprising the following steps of: the spectrum generation module is used for generating a one-dimensional process energy density spectrum on line, wherein the process energy density spectrum is formed by dynamically distributing weights to a plurality of process control variables and carrying out nonlinear superposition; The factor construction module calculates the stability factor of each production beat point based on the process energy density spectrum to construct a beat characteristic field, wherein the stability factor is determined according to the deviation degree of the current spectrum value relative to a target reference, the spectrum value change gradient of the adjacent beat points and the state logic values of a downstream buffer zone and a downstream process; the instruction splitting module is used for splitting the optimized decision instruction into a plurality of decision quantum packets and binding each decision quantum packet with future production beat points of which the stability factors in the beat characteristic field exceed a set threshold value; the reverse sequence simulation module is used for performing reverse sequence disturbance simulation on the decision quantum packet sequence bound with the beat points, reversely calculating the modification quantity of each decision quantum packet to the process energy density spectrum after the decision quantum packet takes effect from the tail end of the sequence, and checking whether the stability factor of any beat point is lower than a set threshold value; The agent game module is used for taking the decision quantum packet passing the inspection as an agent, and performing agent game in a beat point interval when the stability factor exceeds a set threshold value so as to adjust the binding beat point of the agent game module to form an optimized sequence; And the beat synchronization module is used for synchronizing the actual production beat and the optimization sequence, and executing the included adjustment instruction when the beat reaches the binding beat point of any decision quantum packet.
2. The workpiece quality online analysis and production decision optimization model of claim 1, wherein the spectrum generation module comprises: Calculating an instantaneous influence potential sequence changing along with time based on the position of the current value of each process control variable relative to the process constraint window and the change acceleration of the current value; constructing a time sequence entanglement matrix according to the instantaneous influence potential sequences of the process control variables, wherein each element of the matrix is determined based on the waveform similarity and the phase difference of the two-by-two instantaneous influence potential sequences; convolving the instantaneous influence potential sequence of each process control variable with a dynamic convolution kernel constructed by a row vector corresponding to the process control variable in a time sequence entanglement matrix to generate a primary density stream corresponding to each variable; At each moment, identifying the main guide flow with the largest gradient in all the primary density flows, focusing the energy of each primary density flow at the current moment on the characteristic frequency envelope of the main guide flow, and collapsing to generate a single-dimensional process energy density spectrum.
3. The workpiece quality online analysis and production decision optimization model of claim 1 or 2, wherein the factor construction module comprises: Calculating and generating an influence cone projection vector according to the difference value between the current value of the process energy density spectrum and the target reference value and combining the local frequency characteristic of the process energy density spectrum, wherein each component of the influence cone projection vector represents the expected influence amplitude of the current spectrum value deviation on the future continuous beat point; and calculating the inertia damping factor of the beat point based on the change curvature and the change direction persistence of the process energy density spectrum near the beat point.
4. The workpiece quality online analysis and production decision optimization model of claim 3, wherein the factor construction module is specifically configured to: Calculating a pressure conduction coefficient through a pressure conduction model according to the physical occupation state of the downstream buffer zone and the ready signal of the downstream process; the stable factor field distribution obtained in the previous iteration is used as initial field distribution, the influence cone projection vector is used as a disturbance source, the inertia damping factor is used as a field medium absorption coefficient, the pressure conduction coefficient is used as a boundary driving force, and the stable factor of the current beat point is obtained and updated through field iterative balance calculation.
5. The workpiece mass online analysis and production decision optimization model of any one of claims 1, 2,3, or 4, wherein the instruction splitting module comprises: analyzing the decision instruction according to different adjustment dimensions, determining an influence potential amplitude according to adjustment amplitude and parameter sensitivity coefficient of each dimension, and determining an action span by combining a physical time window required by each dimension to take effect, thereby generating a plurality of decision quantum packets with the influence potential amplitude and the action span attribute; And calculating the dynamic affinity of each decision quantum packet for each candidate beat point of which the stability factor exceeds a set threshold in the beat characteristic field, wherein the dynamic affinity is determined by the matching degree of the stability factor allowance of the beat point and the influence potential amplitude of the decision quantum packet and the coincidence degree of the action span of the decision quantum packet and the future process stability window of the beat point based on process energy density spectrum prediction.
6. The workpiece quality online analysis and production decision optimization model of claim 5, wherein the instruction splitting module is specifically configured to perform the following binding operations: And performing multiple rounds of two-way bidding matching by taking the influence potential amplitude of each decision quantum packet as bidding capability and the dynamic affinity of each beat point to each decision quantum packet as admittance, and determining the finally bound beat point for each decision quantum packet.
7. The workpiece quality online analysis and production decision optimization model of claim 6, wherein the reverse order simulation module comprises: generating characteristic disturbance waves for the decision quantum packet of each bound beat point, wherein the initial amplitude of the characteristic disturbance waves is determined by the influence potential amplitude of the decision quantum packet, the wavelength is determined by the action time span of the characteristic disturbance waves, and the bound beat points of the decision quantum packet are taken as emission origins to be attenuated and propagated along a time axis; And superposing the characteristic disturbance waves of all the decision quantum packets on a reverse sequence time axis from the tail end of the sequence, judging the interference effect of each characteristic disturbance wave according to the phase of each characteristic disturbance wave, and generating an interference superposition waveform of the comprehensive disturbance amplitude.
8. The workpiece quality online analysis and production decision optimization model of claim 7, wherein the reverse order simulation module is further configured to: Mapping the comprehensive disturbance amplitude of the interference superposition waveform into a modified risk field of the energy density spectrum of the corresponding beat point process through a risk transfer function, wherein the modified risk field comprises an expected amount and an uncertainty range of spectrum value modification; And substituting the modified risk field into the beat characteristic field model to perform prospective simulation, re-estimating the stability factor of the related beat point, and checking whether the re-estimated stability factor is lower than a set threshold value.
9. The workpiece quality online analysis and production decision optimization model of claim 6 or 8, wherein the proxy gaming module comprises: Calculating an intra-sequence conflict coefficient between each decision quantum packet which passes verification, determining the conflict coefficient based on the time proximity of the binding beat point and the effect interference degree of the reverse order disturbance simulation prediction, and calculating migration willingness for each decision quantum packet, wherein the migration willingness is positively correlated with the sum of the conflict coefficients and negatively correlated with the stability factor allowance of the current binding beat point; calculating the dynamic bargaining capacity of each decision agent according to migration wish and the influence potential amplitude of the decision quantum packet, and distributing game strategies for each decision agent based on the capacity; And in the beat point interval when the stability factor exceeds the set threshold, executing multiple rounds of negotiation according to the game strategy, calling reverse order disturbance simulation to perform local verification aiming at beat point variation proposal generated by each round of negotiation, recording the verified variation as a temporary binding protocol, and forming an optimized decision sequence based on a finally achieved protocol set.
10. The workpiece quality online analysis and production decision optimization model of claim 9, wherein the beat synchronization module comprises: the bridge takes the binding beat point of the bridge as a theoretical zeroing moment, dynamically adjusts the passing speed of the bridge according to the deviation rate of the monitored actual production beat rate and the theoretical beat rate, and simultaneously calibrates an execution critical window around the zeroing point; When the dynamic countdown bridge enters the execution critical window, the real-time stability factor of the target beat point and the current state deviation of related process equipment are checked, the adjustment instruction contained in the decision sub-packet is conditionally compiled according to the check result, and the compiled instruction is issued at the time when the bridge returns to zero.

Description

Workpiece quality on-line analysis and production decision optimization model Technical Field The invention relates to the field of industrial production management, in particular to an on-line analysis and production decision optimization model for workpiece quality. Background Currently, quality analysis in industrial production is mostly dependent on single variable monitoring or artificial experience judgment, systematic analysis capability on multivariate synergism is lacking, the traditional method is often used for treating the influence of each technological parameter in an isolated way, deep association between parameters and a recessive action mechanism on quality indexes cannot be effectively mined, so that when quality abnormality occurs, root causes are difficult to quickly locate, meanwhile, a production decision process is limited by artificial data processing efficiency and experience limitation, decision is often delayed from dynamic change of a production site according to lack of scientificity of data support, and quality risks caused by process fluctuation cannot be responded in time. In production management links such as resource scheduling and workflow planning, the traditional mode mostly adopts a fixed flow or static allocation strategy, and cannot be combined with real-time quality data to dynamically adjust resource allocation and technological parameters, when parameter drift, equipment state fluctuation and other conditions occur in the production process, the conventional decision support system is difficult to quickly generate a targeted optimization scheme, so that the stability of the production process is insufficient, the fluctuation of the product yield is large, in addition, the traditional abnormal cause positioning often needs manual link-by-link investigation, the time is long for several hours, the production efficiency is seriously influenced, and the production cost and the quality loss are increased. Disclosure of Invention Aiming at the problems existing in the prior art, the invention aims to provide an on-line analysis and production decision optimization model for workpiece quality, which can realize accurate adaptation of decision instructions and production beats and dynamic allocation of resources, avoid resource waste caused by the traditional fixed flow, improve the utilization efficiency of production resources such as equipment, materials and the like, reduce the production cost and enhance the flexibility of production management and control. In order to solve the problems, the invention adopts the following technical scheme: the workpiece quality online analysis and production decision optimization model comprises the following steps: the spectrum generation module is used for generating a one-dimensional process energy density spectrum on line, wherein the process energy density spectrum is formed by dynamically distributing weights to a plurality of process control variables and carrying out nonlinear superposition; The factor construction module calculates the stability factor of each production beat point based on the process energy density spectrum to construct a beat characteristic field, wherein the stability factor is determined according to the deviation degree of the current spectrum value relative to a target reference, the spectrum value change gradient of the adjacent beat points and the state logic values of a downstream buffer zone and a downstream process; the instruction splitting module is used for splitting the optimized decision instruction into a plurality of decision quantum packets and binding each decision quantum packet with future production beat points of which the stability factors in the beat characteristic field exceed a set threshold value; the reverse sequence simulation module is used for performing reverse sequence disturbance simulation on the decision quantum packet sequence bound with the beat points, reversely calculating the modification quantity of each decision quantum packet to the process energy density spectrum after the decision quantum packet takes effect from the tail end of the sequence, and checking whether the stability factor of any beat point is lower than a set threshold value; The agent game module is used for taking the decision quantum packet passing the inspection as an agent, and performing agent game in a beat point interval when the stability factor exceeds a set threshold value so as to adjust the binding beat point of the agent game module to form an optimized sequence; And the beat synchronization module is used for synchronizing the actual production beat and the optimization sequence, and executing the included adjustment instruction when the beat reaches the binding beat point of any decision quantum packet. Further, the spectrum generation module includes: Calculating an instantaneous influence potential sequence changing along with time based on the position of the current value of ea