CN-121998314-A - Production plan scheduling collaborative optimization method and system based on MOM and multi-source data

CN121998314ACN 121998314 ACN121998314 ACN 121998314ACN-121998314-A

Abstract

The invention relates to the technical field of intelligent manufacturing and industrial software, and discloses a production plan scheduling collaborative optimization method and a production plan scheduling collaborative optimization system based on MOM and multi-source data, wherein the method comprises the steps of constructing an initial time-varying capacity constraint envelope of equipment; extracting processing task demand characteristics and calculating processing effect residual vectors, gradually superposing the residual vectors to equipment states based on a state evolution equation to update an envelope in real time, executing inclusion detection of the demand characteristics and the envelope, identifying physical dimensions which lead to narrowing of the envelope when a detection result is not included, generating an active restorative virtual task insertion sequence to remodel equipment state tracks until the inclusion requirement is met, and finally generating machine instruction codes to execute closed-loop control. According to the invention, by establishing a bidirectional coupling mechanism of task execution and equipment physical state evolution, the on-line active restoration of equipment capacity is realized by utilizing an active restoration strategy, and the high precision and stability of heavy equipment manufacturing are ensured.

Inventors

YANG BO
AN DI
ZHANG LEI
CHEN WENQIANG
SU JINGQIN
XU QIANGQIANG
ZHANG FENG

Assignees

中煤科工集团信息技术有限公司

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (10)

1. The production plan scheduling collaborative optimization method based on MOM and multi-source data is characterized by comprising the following steps: S1, collecting real-time working condition data of manufacturing equipment, and establishing an initial time-varying capacity constraint envelope of the manufacturing equipment in a future scheduling window by combining a preset natural decay model, wherein the initial time-varying capacity constraint envelope defines a performance boundary set of the manufacturing equipment under no processing load interference; S2, analyzing to-be-arranged to generate a production order, extracting a processing task demand feature fingerprint representing the hard physical condition of the processing task for each processing task, and calculating a processing effect residual vector representing the reverse action of the processing task on the physical state of the manufacturing equipment; S3, in the process of generating a candidate task sequence, taking the initial time-varying capacity constraint envelope established in the step S1 as an evolution standard, and according to a state evolution equation, gradually overlapping the processing effect residual vector calculated in the step S2 corresponding to the preceding task in the candidate task sequence to the current state of the manufacturing equipment, and calculating and updating the time-varying capacity constraint envelope of the subsequent time step in real time; s4, performing inclusion detection, and judging whether the characteristic fingerprint of the processing task requirement extracted in the step S2 corresponding to the subsequent processing task to be arranged completely falls into the time-varying capability constraint envelope updated in the step S3; S5, when the inclusion detection result is not included, identifying physical dimensions which lead to narrowing of the updated time-varying capacity constraint envelope, generating active restorative virtual tasks which do not generate actual output, inserting the active restorative virtual tasks into the candidate task sequence, and reshaping a state evolution track of the manufacturing equipment by using the active restorative virtual tasks until the reshaped time-varying capacity constraint envelope can contain the processing task demand feature fingerprints of the subsequent processing tasks to be processed; And S6, converting the final optimized sequence containing the processing task and the active restorative virtual task into a machine instruction code, and transmitting the machine instruction code to the manufacturing equipment to execute closed-loop control.
2. The method of claim 1, wherein in step S1, the process of establishing an initial time-varying capacity constraint envelope of the manufacturing equipment within a future schedule window further comprises drift pre-determination logic, specifically comprising: calculating a time-varying gradient based on the continuous time sequence state vector of the manufacturing equipment, and constructing an equipment state evolution trend equation at the current moment; Superposing the processing effect residual vector corresponding to the processing task to be executed to the equipment state evolution trend equation to generate a synthetic track curve, calculating a geometric intersection point of the synthetic track curve and a dynamic safety early warning boundary inside the initial time-varying capacity constraint envelope, and marking the geometric intersection point as a drift critical time point; And when the drift critical time point is earlier than the planned execution time of the subsequent processing task, generating a trigger signal to activate the active restorative virtual task generation flow in the step S5.
3. The method for collaborative optimization of production plan scheduling based on MOM and multi-source data according to claim 1, wherein the step S2 specifically comprises: Analyzing the process attribute and technical index data of the workpieces contained in the production order to be arranged, and screening out key geometric features and quality constraint parameters of the workpieces to be processed; Quantitatively mapping the quality constraint parameters into multidimensional vectors, so as to construct the processing task demand feature fingerprint, wherein the dimensions of the multidimensional vectors at least comprise a form and position tolerance level, a surface roughness threshold value and a minimum dynamic stiffness demand; Calculating the machining effect residual vector by using a physical effect mapping algorithm, deducing the thermal deformation increment, the stress accumulation amount and the cutter abrasion increment in the machining process according to the workpiece material property and the cutting parameter, and integrating the thermal deformation increment, the stress accumulation amount and the cutter abrasion increment into the machining effect residual vector.
4. The method for collaborative optimization of MOM and multi-source data-based production planning scheduling according to claim 1, wherein in step S3, performing an effect complementary-based ordering strategy when generating a candidate task sequence specifically comprises: Establishing a task classification matrix, and dividing the processing task into a heat accumulation type task, a heat dissipation type task, a stress loading type task and a stress releasing type task according to the key component direction of the processing effect residual vector; Monitoring an offset vector of the current physical state of the manufacturing equipment relative to the initial time-varying capacity constraint envelope geometric center in real time; Preferably, a processing task for which the processing effect residual vector can be subtracted from the offset vector is selected as the execution task for the next time step, thereby maintaining the stability of the state trajectory integration within the initial time varying capability constraint envelope.
5. The method for collaborative optimization of MOM and multi-source data-based production planning scheduling according to claim 1, wherein the specific method for generating active restorative virtual tasks that do not produce actual yields and inserting the candidate task sequences in step S5 comprises: Determining a corresponding equipment state failure mode according to the identified physical dimension which leads to the narrowing, accessing a preset recovery strategy library and calling a matched active recovery virtual task template, wherein the active recovery virtual task template defines recovery action types including idle cooling, destressing operation or cutter reset; Calculating a state difference value between the updated current state of the manufacturing equipment and a target permission state boundary required by the subsequent processing task to be processed in the step S3, and reversely solving the necessary execution duration of the active restorative virtual task based on the state difference value by utilizing a preset physical response function; Inserting the active recovery virtual task generated based on the active recovery virtual task template into the candidate task sequence before the inclusion detection result is an unencluded processing task node, and carrying out forward processing on the unencluded processing task node and the planned starting time of all subsequent processing tasks according to the necessary execution time.
6. The method for collaborative optimization of production planning scheduling based on MOM and multi-source data according to claim 1, wherein in the step S5, when determining the execution duration of the active restorative virtual task, the step of further performing parameter correction based on a phase synchronization principle specifically includes: Constructing a discrete demand weight sequence of a processing task to be produced and a continuous capacity carrier function representing the performance fluctuation rule of the manufacturing equipment; calculating the time domain matching degree integral of the discrete demand weight sequence and the continuous capability carrier function; When the result of the time domain matching degree integration shows that the high-weight processing task falls in the trough section of the continuous capability carrier function, the duration of the active restorative virtual task is adjusted, and time domain phase shifting is carried out on the subsequent continuous capability carrier function until the time window of the high-weight processing task is aligned with the crest section of the continuous capability carrier function.
7. The method for collaborative optimization of production planning scheduling based on MOM and multisource data according to claim 1, wherein in the step S3, the specific method for constructing the state evolution equation based on the dynamic hysteresis loop geometric model comprises: Constructing a state phase plane coordinate system which adopts an external excitation load as an abscissa and physical state deviation as an ordinate; Fitting a loading response path of a processing task execution stage and an unloading recovery path of an intermittent stage, and defining a closed area surrounded by the loading response path and the unloading recovery path as a dynamic hysteresis loop; Calculating the geometric area of the dynamic hysteresis loop to characterize the dissipated energy density of a single machining cycle; Introducing a preset damage constitutive mapping coefficient, and calculating and converting the dissipated energy density into a state drift increment item representing irreversible degradation of equipment performance; And establishing a recursive mathematical expression describing the evolution of the equipment state along with the processing process according to the state accumulation principle of the discrete time steps, wherein the recursive mathematical expression enables the equipment state of the next time step to be equal to the vector sum of the current equipment state and the state drift increment term, so that the construction of the state evolution equation is completed.
8. The collaborative optimization method for production planning scheduling based on MOM and multi-source data according to claim 7, wherein in the step S1, the specific method for performing long-term constraint feedback correction on the natural decay model by using fatigue damage indexes comprises: Extracting the dissipation energy density calculated for each processing task in the step S3, performing weighted summation calculation to generate the fatigue damage index of the current scheduling window, and calculating the damage accumulation rate of the fatigue damage index along with time; comparing the damage accumulation rate to a full life cycle maintenance planning curve for the manufacturing equipment; When the damage accumulation rate exceeds a preset safety threshold, calculating the envelope estimated shrinkage of the next period according to the accumulation amount of the fatigue damage index, and inwards shrinking the boundary of the initial time-varying capacity constraint envelope by utilizing the envelope estimated shrinkage.
9. The method for collaborative optimization of MOM and multisource data-based production plan scheduling according to claim 1, further comprising an initial boundary condition confirmation step prior to step S1, comprising: Collecting residual physical state data of the manufacturing equipment after the last period is finished; Driving the manufacturing equipment to execute a standardized rapid calibration detection instruction, collecting real-time response data and comparing the real-time response data with theoretical reference data to generate a deviation matrix; And (2) combining the residual physical state data with the deviation matrix, calculating an initial capacity state vector at the current moment, and taking the initial capacity state vector as a time zero point boundary condition for establishing the initial time-varying capacity constraint envelope in the step (S1).
10. A co-optimizing system for production schedule based on MOM and multi-source data, for implementing the co-optimizing method for production schedule based on MOM and multi-source data according to any one of claims 1 to 9, the system comprising: the edge side multidimensional state sensing unit (100) is deployed in the manufacturing equipment and is used for collecting vibration, load and temperature field working condition data; The capacity envelope calculation and edge gateway subsystem (200) is connected with the edge side multidimensional state sensing unit (100) and is used for receiving the working condition data and processing and generating a time-varying capacity constraint envelope; The collaborative optimization and decision-making server (300) is connected with the capability envelope calculation and edge gateway subsystem (200) and is used for executing logical operations of task analysis, effect residual vector calculation, dynamic sequence deduction, envelope conflict detection and active restorative virtual task generation; And the scheduling instruction and equipment control cooperative interface (400) is used for converting a final optimization sequence containing the active restorative virtual task generated by the cooperative optimization and decision server (300) into a composite instruction and transmitting the composite instruction to the equipment controller.

Description

Production plan scheduling collaborative optimization method and system based on MOM and multi-source data Technical Field The invention relates to the technical field of intelligent manufacturing and industrial software, in particular to a production plan scheduling collaborative optimization method and system based on MOM and multi-source data. Background Heavy equipment manufacturing, taking hydraulic supports and coal mining machines as representative coal mining machines as examples, is a typical complex discrete manufacturing industry in the field of industrial manufacturing, and has the characteristics of complex product structure, high customization degree, long production period and multi-variety small-batch production. In such a manufacturing model, the manufacturing operations management platform serves as a core hub connecting the enterprise resource planning layer and the plant control layer, assuming the role of converting production orders into specific execution instructions. Through scientific production planning and scheduling, manufacturing resources such as personnel, equipment, materials and the like are reasonably configured, and the method has important roles in ensuring on-schedule delivery of large equipment, shortening manufacturing cycle and reducing operation cost. Existing production plan scheduling techniques follow hierarchical processing logic, namely, a main production plan and a bill of materials are acquired from an ERP system, and a workshop operation plan is generated by using an advanced plan scheduling engine in combination with inventory data of a warehouse management system. The working principle is based on a static resource capacity constraint model, the working procedures are ordered and assigned by adopting an operation research algorithm or heuristic rules, a production scheme meeting basic process constraints can be calculated under an ideal stable environment, and the production scheme is issued to a production site for execution in a Gantt chart form to try to construct a determined and predefined production order. The existing scheduling model is focused on static orders or process dimensions in multiple ways, a multi-layer linkage mechanism of a project level, an order level and a process level is lacked, so that macroscopic project milestones are difficult to effectively restrict microscopic process execution, an upper layer plan and a lower layer plan are easy to be disjointed, the existing system is often passive and lagged in data utilization, multi-source heterogeneous data such as an equipment internet of things and material logistics are difficult to fuse in real time, dynamic disturbance such as equipment sudden faults and material delays cannot be perceived by a scheduling scheme, the generated plan is optimal in theory but poor in practical executable, and time-consuming global rearrangement is often required once abnormality occurs. In view of the above, the present invention aims to provide a method and a system for collaborative optimization of production plan scheduling based on MOM and multi-source data, so as to solve the deficiencies in the prior art. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a production plan scheduling collaborative optimization method and a production plan scheduling collaborative optimization system based on MOM and multi-source data, which solve the problems that a planning level fracture exists in a production plan and scheduling system, macroscopic milestones and microscopic execution are difficult to collaborate, multi-source heterogeneous data cannot be effectively fused to cope with a dynamic production environment, a rapid self-adaptive decision response mechanism is lacking in the face of frequent disturbance, and artificial experience knowledge and algorithm models are difficult to perform closed loop iterative optimization, so that the performability of a scheduling scheme is poor and the intelligentization level is difficult to continuously promote. In order to achieve the above purpose, the invention is realized by the following technical scheme that the first aspect of the invention provides a production plan scheduling collaborative optimization method based on MOM and multi-source data, which comprises the following steps: S1, collecting real-time working condition data of manufacturing equipment, and establishing an initial time-varying capacity constraint envelope of the manufacturing equipment in a future scheduling window by combining a preset natural decay model, wherein the initial time-varying capacity constraint envelope defines a performance boundary set of the manufacturing equipment under no processing load interference; S2, analyzing to-be-arranged to generate a production order, extracting a processing task demand feature fingerprint representing the hard physical condition of the processing task for each processing task, and calculating a processi