A method for intelligent flow scheduling of high-performance oil circuits in sheet metal piping networks

Abstract

This invention relates to the field of data management technology, specifically to an intelligent flow scheduling method for high-performance oil circuits in sheet metal pipeline networks. The specific implementation process includes: establishing a distributed intelligent execution entity and a resource optimization model based on a sheet metal pipeline network containing schedulable units and production management information including historical performance baselines and production plans; the distributed intelligent execution entity acquires and evaluates the operating data of the schedulable units to generate real-time status data; combining the real-time status data and historical performance baselines to analyze drift trends and generate performance degradation data; based on the production plan and performance degradation data, quantifying the comprehensive operating cost as the objective function to calculate and generate resource allocation strategies and work instructions; and the distributed intelligent execution entity combining work instructions, real-time status data, and performance degradation data to execute dynamic compensation management and generate scheduling instructions. This invention optimizes resource allocation by evaluating the real-time and long-term status of equipment, achieving risk avoidance and dynamic cost optimization.

Description

Technical Field

[0001] This invention relates to the field of data management technology, specifically to a method for intelligent flow scheduling of high-performance oil circuits in sheet metal pipeline networks. Background Technology

[0002] In the production and operation management of sheet metal processing, the operational efficiency of production equipment is crucial for achieving on-time delivery and cost control. Among these factors, the scheduling of oil circuit flow affects production efficiency, processing accuracy, and equipment lifespan. Ensuring the economy and foresight of flow scheduling is a prerequisite for achieving lean production and intelligent operation management.

[0003] Existing operation management and scheduling methods have some problems. On the one hand, traditional scheduling methods use preset models based on fixed parameters, which cannot dynamically adjust resource allocation when faced with fluctuations in oil parameters during production, resulting in slow system response and energy waste. On the other hand, the data-driven intelligent algorithms that have been applied are usually static strategies, lacking the ability to adapt online to long-term changes in equipment performance, and cannot meet the needs of continuous optimization in resource scheduling and risk avoidance in industrial scenarios.

[0004] In summary, existing technologies are unable to cope with short-term operational changes and long-term performance variations, resulting in low overall efficiency. Therefore, a method for intelligent flow scheduling of high-performance oil circuits in sheet metal piping networks is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a method for intelligent flow scheduling of high-performance oil circuits in sheet metal pipeline networks for intelligent data management. To address the problems of existing technologies, this invention first establishes a distributed intelligent execution entity and a resource optimization model based on a sheet metal pipeline network containing schedulable units and production management information. The production management information includes historical performance baselines and production plans. Then, the operating data of the schedulable units is acquired, and real-time status data is generated through the distributed intelligent execution entity based on this data. The current output status is calculated for preset test instructions, and the drift trend of the current output status deviating from the historical performance baseline is analyzed to generate performance degradation data. Next, based on the production plan and the performance degradation data, a comprehensive operating cost is quantified and generated. Using the comprehensive operating cost as the objective function, the resource optimization model calculates and generates a resource allocation strategy. Work instructions are generated based on the resource allocation strategy. Finally, the distributed intelligent execution entity combines the work instructions, real-time status data, and performance degradation data to execute dynamic compensation management and generate scheduling instructions. The schedulable units dynamically track the work instructions based on the scheduling instructions. This invention achieves a collaborative scheduling effect of proactive risk avoidance and dynamic cost optimization by predicting the load demand of future production tasks and the evolution trend of their own performance status.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for intelligent flow scheduling of high-performance oil circuits in sheet metal piping networks includes:

[0008] Based on the sheet metal pipeline network containing schedulable units and production management information, a distributed intelligent execution entity and resource optimization model are established; the production management information includes historical performance baselines and production plans.

[0009] The system acquires the operational data of the schedulable unit, evaluates the operational data to generate real-time status data, calculates the current output status for a preset test instruction, analyzes the drift trend of the current output status deviating from the historical performance baseline, and generates performance degradation data of the schedulable unit.

[0010] Based on the production plan and the performance degradation data, the comprehensive operating cost is quantified and generated. The comprehensive operating cost is used as the objective function. The resource optimization model calculates and generates a resource allocation strategy. Based on the resource allocation strategy, work instructions are generated.

[0011] The distributed intelligent executor combines the job instructions, real-time status data, and performance degradation data to perform dynamic compensation management and generate scheduling instructions; the schedulable unit dynamically tracks the job instructions based on the scheduling instructions.

[0012] Preferably, the distributed intelligent actuator includes a state evaluation unit and a performance trend prediction unit; the state evaluation unit includes an information network model, which receives the operating data and generates real-time state data, and the information network model includes a fluid state model; the performance trend prediction unit includes an online learning algorithm, which analyzes the output state of preset test operation instructions and historical performance baselines to generate performance degradation data; based on the operation instructions and by fusing the real-time state data and the performance degradation data, the dynamic compensation management is executed, and the final control output is calculated.

[0013] Preferably, the resource optimization model is constructed based on a demand forecasting model and an optimization algorithm, and is implemented through data integration, optimization decision-making, and instruction generation stages. The data integration stage receives and integrates the production plan and the performance degradation data to generate a global decision view that includes the pipeline network status and future task load. The optimization decision-making stage, based on the global decision view, executes the forward-looking optimization solution to generate the resource allocation strategy, which includes the running roles and work tasks of schedulable units. The instruction generation stage parses the running roles and work tasks into work instructions that include timing and physical parameters.

[0014] Preferably, the historical performance baseline is the mapping relationship between the baseline job and the required resource consumption when the schedulable unit is in the baseline state.

[0015] Preferably, the process of generating real-time status data includes: using the operating data as input to the information network model, the operating data including temperature data and pressure data; the information network model performing a nonlinear transformation on the temperature data and pressure data to generate preliminary inference values ​​of real-time status data; the real-time status data including oil viscosity and density; calculating reference values ​​of real-time status data through a fluid state model, the fluid state model including a viscosity-temperature relationship model and a density-temperature-pressure relationship model; comparing the preliminary inference values ​​and the reference values ​​of real-time status data and calculating the difference; determining weights based on the difference; and obtaining the final status data by weighted summation of the preliminary inference values ​​and the reference values ​​of real-time status data.

[0016] Preferably, the process of generating performance degradation data includes: calculating a sequence signal and obtaining the required current output state based on a preset test instruction; comparing the current output state with a historical performance baseline to generate a real-time deviation value sequence; filtering out random noise from the real-time deviation value sequence to obtain a smooth curve, and extracting a dynamic drift trend signal by applying linear regression trend analysis to the smooth curve; extracting feature parameters to generate a feature vector based on the dynamic drift trend signal, the feature parameters including average bias, response delay, overshoot and oscillation amplitude, and drift rate; normalizing the feature parameters using the mapping model in the online learning algorithm and weighted summing to generate a current performance degradation index and a future degradation rate; matching the feature parameters with a preset fault rule base to identify potential fault signatures; calculating the predicted remaining service life based on the current performance degradation index, the future degradation rate, and a preset failure threshold; and integrating the current performance degradation index, the future degradation rate, the potential fault signature, and the predicted remaining service life into performance degradation data output.

[0017] Preferably, the process of generating the resource allocation strategy includes: parsing the production plan to generate specific process steps, obtaining the hydraulic demand curves for the specific process steps, generating a time-series load demand sequence and a predicted total energy consumption demand based on the hydraulic demand curves; combining the predicted total energy consumption demand and the performance degradation data, quantitatively evaluating the performance status, operating efficiency, and failure risk of the schedulable unit, determining resource loss costs, energy costs, and risk costs, and weighted summing the resource loss costs, energy costs, and risk costs to generate a quantitative comprehensive operating cost; using the total output of the schedulable unit satisfying the time-series load demand sequence as a constraint, constructing an objective function, the objective function taking minimizing the comprehensive operating cost as the optimization objective, and solving the objective function to generate the resource allocation strategy.

[0018] Preferably, the step of generating scheduling instructions through dynamic compensation management includes: parsing the job instruction to obtain the target trajectory; inputting the target trajectory into a simplified dynamic model to generate an initial control output sequence, wherein the simplified dynamic model includes the state parameters and response characteristics of the schedulable unit in the baseline state; performing feedforward compensation based on the deviation between the real-time state data and the baseline state system data to obtain a feedforward compensated control output sequence; using the potential fault signature of the performance degradation data to perform adaptive feedback correction on the feedforward compensated control output sequence; calculating the final control output sequence; and generating scheduling instructions.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0020] 1. This invention utilizes online learning algorithms and a pre-defined fault rule base to perform deep quantification and pattern matching on feature parameters extracted from the deviation between the current output state of schedulable units and historical performance baselines. In this way, the current real-time operating status of the equipment can be perceived, long-term performance evolution trends, potential fault modes, and remaining service life can be predicted, generating performance degradation data. This enables a shift from passive response to proactive prediction, effectively avoiding unplanned downtime caused by equipment aging and improving the predictability and reliability of equipment management.

[0021] 2. This invention establishes a resource optimization model and utilizes an optimization solver to perform forward-looking collaborative optimization of the global decision view, which includes production planning and performance degradation data. Simultaneously, it quantifies the comprehensive operating cost, including resource depletion, energy consumption, and potential failure risks, and uses the quantification results as the objective function for global optimization. This operation uses equipment status as the core of production scheduling decisions, proactively balancing the relationship between short-term production tasks and long-term equipment health, prioritizing the allocation of high-precision or heavy-load tasks to units with better health status, and achieving dynamic minimization of global operating costs while satisfying the constraints of the overall production plan.

[0022] 3. This invention utilizes a dynamic compensation management mechanism within a distributed intelligent execution system to perform closed-loop regulation on work instructions that combine real-time status data and performance degradation data. This regulation includes feedforward compensation and adaptive feedback correction, forming a control system with dual safeguards. Feedforward compensation offsets disturbances caused by short-term changes in status data in real time, while adaptive feedback correction continuously addresses the impact of long-term performance degradation. This ensures that work instructions can be accurately executed under the dual challenges of changing working conditions and continuous equipment performance degradation, guaranteeing the long-term stability of processing accuracy and product quality, and enhancing the robustness of the entire production system. Attached Figure Description

[0023] Figure 1The flowchart illustrates a method for intelligent flow scheduling of high-performance oil circuits in sheet metal pipelines, as proposed in an embodiment of this invention.

[0024] Figure 2 This is a schematic diagram of the resource scheduling process proposed in an embodiment of the present invention.

[0025] Figure 3 This is a schematic diagram of the intelligent decision-making process proposed in an embodiment of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Please see Figures 1 to 3 This invention provides an intelligent flow scheduling method for high-performance oil circuits in sheet metal pipeline networks, the technical solution of which is as follows:

[0028] A method for intelligent flow scheduling of high-performance oil circuits in sheet metal pipelines, comprising the following steps:

[0029] Based on the sheet metal pipeline network containing schedulable units and production management information, a distributed intelligent execution entity and resource optimization model are established; the production management information includes historical performance baselines and production plans.

[0030] The system acquires the operational data of the schedulable unit, evaluates the operational data to generate real-time status data, calculates the current output status for a preset test instruction, analyzes the drift trend of the current output status deviating from the historical performance baseline, and generates performance degradation data of the schedulable unit.

[0031] Based on the production plan and the performance degradation data, the comprehensive operating cost is quantified and generated. The comprehensive operating cost is used as the objective function. The resource optimization model calculates and generates a resource allocation strategy. Based on the resource allocation strategy, work instructions are generated.

[0032] The distributed intelligent executor combines the job instructions, real-time status data, and performance degradation data to perform dynamic compensation management and generate scheduling instructions; the schedulable unit dynamically tracks the job instructions based on the scheduling instructions.

[0033] Example 1

[0034] This embodiment provides a specific application of a flow intelligent scheduling method for high-performance oil circuits in sheet metal pipelines. Its typical application scenario is that Company A introduces a flow intelligent scheduling method to reduce the energy consumption of the hydraulic system and avoid unplanned shutdowns caused by equipment aging.

[0035] refer to Figure 1 The method includes:

[0036] S1. Based on the sheet metal pipeline network containing schedulable units and production management information, establish a distributed intelligent execution entity and resource optimization model; the production management information includes historical performance baselines and production plans;

[0037] S2. Obtain the operation data of the schedulable unit, and the distributed intelligent executor evaluates the operation data to generate real-time status data;

[0038] S3. Calculate the current output state for the preset test instructions, analyze the drift trend of the current output state deviating from the historical performance baseline, and generate performance degradation data of the schedulable unit.

[0039] S4. Based on the production plan and the performance degradation data, a comprehensive operating cost is generated in a quantitative manner. The comprehensive operating cost is used as the objective function. The resource optimization model calculates and generates a resource allocation strategy. Based on the resource allocation strategy, a work instruction is generated.

[0040] S5. The distributed intelligent executor combines the job instructions, real-time status data, and performance degradation data to perform dynamic compensation management and generate scheduling instructions; the schedulable unit dynamically tracks the job instructions based on the scheduling instructions.

[0041] Furthermore, based on the sheet metal pipeline network containing schedulable units and production management information, a resource optimization model and a distributed intelligent execution entity are established; the production management information includes historical performance baselines and production plans, corresponding to step S1 above, and the specific process includes:

[0042] Analyze the physical sheet metal piping network, identify core value units, and designate these core value units as independent schedulable units within the system, such as hydraulic presses and high-performance oil circuit stations. Assign unique identifiers to each schedulable unit and input its physical parameters into the system, including maximum pressure, rated flow rate, and cylinder size. Equip the defined schedulable units with edge computing devices, such as industrial PCs or high-performance PLCs. Deploy distributed intelligent actuators based on these edge computing devices. The deployment process specifically involves instantiating a software architecture that includes a state evaluation unit and a performance trend prediction unit. The state evaluation unit includes an information network model, which is constructed based on a physical information neural network and includes a fluid state model. For example, the information network model uses a fully connected feedforward neural network with three hidden layers. The input layer receives temperature and pressure data and uses ReLU as the activation function. The output layer outputs preliminary inferences of oil viscosity and density.

[0043] The loss function of the information network model is obtained by summing data loss and physical constraints. The data loss is the mean square error between the preliminary inferred value and the reference value, and the reference value is obtained by consulting the specification. The physical constraints are specifically: the viscosity and density values ​​of the oil inferred by the information network model are compared with the theoretical values ​​calculated by the fluid state model based on temperature and pressure data to obtain the absolute values ​​of viscosity residuals and density deviations. The physical constraints are obtained by the absolute weighted sum of viscosity residuals and density deviations. The weights are set based on the importance of the actual working conditions. For example, for tasks requiring precise pressure maintenance, pressure stability is crucial, and the weight of density deviation, which is severely affected by pressure, can be set to 0.6, and the weight of viscosity deviation can be set to 0.4. For auxiliary tasks with low requirements, the weights can be set to 0.5 for both.

[0044] The performance trend prediction unit is configured with an online learning algorithm that includes a mapping model, such as a recursive least squares algorithm based on online gradient descent, to update the mapping model in real time based on newly acquired performance characteristics. The distributed intelligent actuator is bound to the corresponding schedulable unit to form a minimum functional closed loop. In this way, the distributed intelligent actuator can infer oil state data in real time and adaptively update its internal information network model, ensuring high-precision execution of control commands under varying operating conditions, improving the system's state awareness and adaptive capabilities, and ensuring accurate and efficient production scheduling.

[0045] Through a standard data interface, the production management system is connected to a locally deployed resource optimization model server. Digital profiles are established for schedulable units, generating historical performance baselines. This process includes executing preset baseline operations based on the baseline state of the schedulable units and recording the mapping relationship between these operations and the required resource consumption. This mapping relationship is then stored to form the historical performance baseline. For example, it records the baseline control output required by the servo valve when a new hydraulic press executes operation instructions covering a range of common operating conditions. By establishing historical performance baselines through the control output of new equipment executing baseline operations, a data foundation is provided for calculating performance degradation indices and predicting remaining service life. This transforms the equipment aging process into quantifiable, concrete values, improving the accuracy of condition assessment and ensuring the reliability of resource scheduling.

[0046] A resource optimization model is constructed based on a demand forecasting model and optimization algorithms. The resource optimization model architecture is serial, including data integration, optimization decision-making, and instruction generation stages. The data integration stage receives production plans from the production management system and real-time performance degradation data from distributed intelligent executors through a configured data interface. It then merges the system production plan and real-time performance degradation data to generate a global decision view containing the current state of the pipeline network and future task loads. The optimization decision-making stage includes an optimization solver that employs intelligent optimization algorithms, such as genetic algorithms and simulated annealing algorithms. It uses the time-series load requirements stipulated in the production plan as constraints and the comprehensive operating cost as the objective function to perform forward-looking optimization, generating a resource allocation strategy that includes the schedulable unit operating roles and work tasks. The comprehensive operating cost is obtained by quantifying resource loss costs, energy costs, and failure risk costs based on performance degradation data and predicted energy consumption requirements, and then weighting and summing them. The instruction generation stage parses the resource allocation strategy generated by the optimization decision into work instructions containing time-series and physical parameters that can be directly executed by the distributed intelligent executors. By constructing a resource optimization model, performance degradation, energy consumption, and risks are quantified into comprehensive operating costs and used as optimization targets. This enables scheduling decisions to proactively balance short-term production with long-term equipment health, improving the scientific and economical nature of the decisions.

[0047] Further, the operational data of the schedulable unit is acquired, and the distributed intelligent executor evaluates the operational data to generate real-time status data. Corresponding to step S2 above, the specific process includes:

[0048] The distributed intelligent actuator's state assessment unit collects real-time sensor data installed on key nodes of the schedulable unit to obtain operational data, including real-time temperature and pressure data. This real-time temperature and pressure data is then used as input to an information network model within the state assessment unit to infer real-time state data, including oil viscosity and density. The inference process includes: using the real-time temperature and pressure data as input to the information network model, which processes the input data through nonlinear transformations to generate preliminary inferred values ​​for the real-time state data; and using a fluid state model, based on the real-time temperature and pressure data, calculating theoretical reference values ​​for the real-time state data. This fluid state model includes a viscosity-temperature relationship model and a density-temperature-pressure relationship model. The viscosity-temperature relationship model transforms the nonlinear viscosity-temperature relationship into a linear one by performing a double logarithmic mathematical transformation on viscosity and absolute temperature. The density-temperature-pressure relationship model first calculates the basic density under normal pressure using a polynomial function of temperature, and then superimposes the compression effect of high pressure using a correction term containing the natural logarithm. The oil characteristic constants in the model are pre-calibrated by consulting the product specifications. These oil characteristics include reference viscosity and density. For example, when hydraulic press A1 uses hydraulic oil a to perform continuous stamping tasks, and the oil temperature rises to 65°C and the pressure reaches 20MPa during the task, the calculation process of the viscosity-temperature relationship model is as follows: The system first consults the specifications of hydraulic oil a to obtain calibration data for two key points: "40°C, viscosity 46cSt" and "100°C, viscosity 6.8cSt". Through double logarithmic transformation, the two key points are connected by a straight line in the logarithmic coordinate system, thus defining a linear viscosity-temperature relationship. When the sensor measures a real-time temperature of 65°C, the system can obtain the corresponding theoretical viscosity value, which is 23.5cSt, through the straight line. The calculation process of the density-temperature-pressure relationship model is as follows: First, the system consults the specifications of hydraulic oil a and obtains its standard density of 877 kg / m³ at 15°C. Using a second-order polynomial function, the basic density at 65°C and normal pressure is calculated to be approximately 858 kg / m³. Next, based on the current pressure, the system calculates the density increase due to pressure using a correction term containing the natural logarithm, which is approximately 12.5 kg / m³. Finally, the basic density and the density increase are added together to obtain the final theoretical density under the current operating conditions.

[0049] The difference between the preliminary inferred value and the theoretical reference value is calculated, and weights are dynamically allocated based on this difference. For example, a trust threshold is set based on the theoretical reference value, such as 5% of the theoretical reference value. If the difference is within the trust threshold, the information network model and the fluid state model are considered reliable under the current operating conditions, and a weight of 0.8 can be assigned to the information network model. If the difference exceeds the trust threshold, the model prediction is considered to deviate from the physical constraints, and the weight of the information network model decreases from 0.8, linearly decreasing to 0.2 within the range where the difference increases from 5% to 15%. For example, when the difference is 10%, the weight can take the intermediate value of 0.5; if the difference exceeds 15%, the weight is fixed at the minimum value of 0.2. The preliminary inferred value and the reference value are summed to generate real-time state data. By fusing network inference and fluid state model verification to generate real-time state data, accurate perception of real-time state data is achieved, effectively improving risk resistance and providing a high-precision and robust data foundation for precise control and optimized scheduling.

[0050] Furthermore, the current output state is calculated based on the preset test instructions, the drift trend of the current output state deviating from the historical performance baseline is analyzed, and performance degradation data of the schedulable unit is generated. Corresponding to step S3 above, the specific process includes:

[0051] During non-production intervals in schedulable units, the performance trend prediction unit of the distributed intelligent actuator automatically executes standardized test instructions. For example, during the waiting period after a hydraulic press in company A completes a stamping task, it executes the instruction to "drive a hydraulic cylinder to linearly increase the pressure from 0MPa to 20MPa within 500ms." The system collects the actual pressure output status, compares the actual pressure output status with the stored historical performance baseline in real time, and generates a real-time deviation value sequence reflecting the current performance deviation. The real-time deviation value sequence is filtered by a low-pass filter algorithm to remove random noise and generate a smooth deviation curve. A linear regression algorithm is used to perform trend analysis on the smooth deviation curve to extract dynamic drift trend information. Based on the dynamic drift trend signal, quantified characteristic parameters are included, such as average bias, response delay, overshoot and oscillation amplitude, and drift rate. For example, the average bias is 0.15 MPa lower than the historical performance baseline when the current pressure output is on average, indicating possible internal leakage and decreased pump efficiency; the response delay is 60 ms later than the baseline when the time to reach 90% of the target pressure, which can be considered a decrease in system response performance; the overshoot and oscillation amplitude increases from 1% of the baseline to 3%, and the residual oscillation amplitude after stabilization increases, which can be considered related to the deterioration of the control valve's dynamic characteristics; the drift rate, based on historical diagnostic data, can be considered to be deteriorating at a rate of 0.01 MPa per week for the average bias.

[0052] The online learning algorithm within the performance trend prediction unit utilizes a mapping model to normalize and weightedly sum the feature parameters to generate a quantitative index. The mapping model is a multivariate linear model, and the current performance degradation index is obtained by weighted summation of the normalized feature parameters. The weighting is updated in real-time using a recursive least squares algorithm based on online gradient descent. Specifically, the initial weight coefficients are preset to equal values. The system obtains verification results of the equipment status by associating with specific maintenance events, and uses these verification results as the target index. For example, when a schedulable unit triggers unplanned maintenance due to performance issues, the system automatically marks the target index corresponding to the performance feature vector collected before maintenance as 0.9, providing supervised learning samples for the algorithm. The newly collected performance feature vector is used as input to the recursive least squares algorithm, with minimizing the error between the performance degradation index and the target index as the optimization objective. The weight coefficients of the multivariate linear model are iteratively adjusted and updated.

[0053] The current performance degradation index is obtained; for example, the current performance degradation index is calculated to be 0.58 based on comprehensive characteristic parameters. The performance degradation index is set to 0 for fully effective and 1 for failure. Secondly, based on the drift rate, the future degradation rate is predicted to be an increase of 0.015 per week. Simultaneously, the characteristic parameters are matched with a preset fault rule base. This preset fault rule base includes identifying typical failure modes of the hydraulic system using failure mode and impact analysis methods. For each failure mode, the correlation between the failure mode and the data characteristic parameters is collected by analyzing historical maintenance records. Based on these correlations, machine-readable logical rules are extracted and solidified to form a rule base. For example, the rule entries might be:

[0054] Rule ID01: If the average bias is greater than 0.1MPa and the response delay is greater than 50ms, the potential fault signature is "increased risk of internal leakage in the main valve"; Rule ID02: If the overshoot is greater than 5% and the steady-state oscillation amplitude is greater than 1.5%, the potential fault signature is "control loop parameter drift and deterioration of servo valve dynamic characteristics"; Rule ID03: If the response delay does not change significantly, but the overshoot and oscillation amplitude continue to increase, the potential fault signature is "PID parameter mismatch in the control system".

[0055] Next, based on the current degradation index, future degradation rate, and preset failure threshold, the predicted remaining service life is calculated. Finally, the current performance degradation index, future degradation rate, potential fault signature, and predicted remaining service life are integrated into structured performance degradation data. By analyzing the deviation between current performance and historical baseline, performance degradation data is generated, realizing a shift from passive fault response to proactive health prediction, improving the predictability and reliability of equipment management, and providing quantitative basis for risk assessment and cost optimization decisions.

[0056] Furthermore, during the generation of performance degradation data, a dynamic correction mechanism for predicted remaining useful life is introduced. The system matches the extracted performance feature vector with a preset fault rule base, identifies fault signatures, and calls the adjustment coefficients associated with the fault signatures to correct the initially predicted future degradation rate. Based on the corrected future degradation rate, the predicted remaining useful life is recalculated. Through this dynamic correction mechanism, the prediction of remaining useful life is transformed from general trend extrapolation to precise calculation based on specific physical fault modes, improving the accuracy and reliability of remaining useful life prediction. This provides a reliable basis for preventative maintenance decisions and reduces resource consumption and fault risk costs in overall operating costs.

[0057] Furthermore, based on the production plan and the performance degradation data, a comprehensive operating cost is quantified and generated. The resource optimization model uses the comprehensive operating cost as the objective function to calculate and generate a resource allocation strategy, and generates work instructions based on the resource allocation strategy. Corresponding to step S4 above, the specific process includes:

[0058] The resource optimization model receives production plans from Company A's production management system, such as producing 5,000 sheet metal parts of model A1 within the next 24 hours. Based on process documents, it obtains the hydraulic demand curve required for stamping A1 parts, generating a time-series load demand sequence and predicting total energy consumption. Combining performance degradation data, it proactively quantifies the cost of schedulable units performing tasks. For example, it performs cost analysis on hydraulic press A2, which has a performance degradation index of 0.58, a potential fault signature indicating "increased risk of internal leakage in the main pump or main valve," and a predicted remaining service life of 21 weeks.

[0059] The comprehensive operating cost is obtained by weighted summation of resource depletion cost, energy cost, and risk cost. The weighting coefficients are set based on the task; for example, under normal production tasks, they are set to 0.3, 0.5, and 0.2 respectively. When task priority or accuracy requirements change, the risk cost weight is increased to 0.7, and other weights are normalized. The resource depletion cost is proportional to the square of the performance degradation index. A penalty weight is assigned to the performance degradation index through non-linear amplification. For example, the penalty weight for healthy device A1 with a performance degradation index of 0.05 can be 0.0025, and the penalty weight for degraded device A2 with a performance degradation index of 0.58 can be 0.34. The energy cost is based on prediction. The energy consumption calculation consists of a baseline power and additional losses due to performance degradation. The risk cost is obtained by adding the equipment cost and the potential quality non-conformity cost. The equipment cost is inversely proportional to the predicted remaining service life, specifically by dividing the equipment replacement cost by the predicted remaining service life. The potential quality non-conformity cost is quantified by a risk multiplier, which is dynamically adjusted based on task accuracy and equipment status. For example, when A2 performs a high-precision task, based on the fault signature of "increased risk of internal leakage in the main pump or main valve," the system can set the risk multiplier to 5.0. However, if it performs a normal auxiliary task that is not sensitive to accuracy, the risk multiplier can be set to 1.0.

[0060] Furthermore, a dynamic adaptive weighting mechanism is introduced during the quantification of comprehensive operating costs. The resource optimization model automatically acquires the real-time attributes of tasks and dynamically adjusts the weighting coefficients based on these attributes. The comprehensive operating cost function after adjusting the weighting coefficients is then used as the new optimization objective. Through this dynamic adaptive weighting mechanism, scheduling decisions are aligned with business priorities in real time. When dealing with unexpected tasks, optimal decisions are made through cost-optimal calculations, enhancing the flexibility and intelligence of decision-making. This transforms production scheduling from a static process of passively executing predetermined plans into a dynamic value-creating process that proactively responds to market changes and internal conditions in real time.

[0061] Based on the cost quantification results, an optimization problem is constructed. The objective function of this optimization problem is the total comprehensive operating cost of all schedulable units. The constraints of the optimization problem are that the combined output of all schedulable units meets the time-sequential load requirements of the production plan, and the final total output equals the planned output. By solving this objective function, the model generates an optimal resource allocation strategy. For example, the optimal strategy might be: allocate 4500 tasks to hydraulic presses in good condition; allocate the remaining 500 tasks to the degraded hydraulic press A2, instructing it to operate under a lower load, and automatically generating a high-priority maintenance work order to inspect hydraulic press A2 within one week. The model parses the resource allocation strategy into work instructions containing precise timing and physical parameters, such as a target pressure of 20 MPa and a holding time of 2 seconds. By quantifying equipment performance degradation and failure risks into comprehensive operating costs, and minimizing these comprehensive operating costs as the optimization objective, the model proactively balances short-term production tasks with long-term equipment health, achieving optimal control of global operating costs and risks.

[0062] Furthermore, the distributed intelligent executor combines the job instructions, real-time status data, and performance degradation data to perform dynamic compensation management and generate scheduling instructions; the schedulable unit dynamically tracks the job instructions based on the scheduling instructions, corresponding to step S5 above, and the specific process includes:

[0063] The schedulable unit of the distributed intelligent actuator receives and parses the work instruction to obtain the target trajectory, which includes time curves of pressure and flow. The target trajectory is input into a simplified dynamic model to generate an initial control output sequence under ideal conditions, such as a second-order model describing the dynamics of hydraulic cylinder pressure, providing a reasonable feedforward open-loop control quantity for the control system. The schedulable unit acquires real-time status data, such as the current oil temperature being 55°C and the viscosity decreasing by 30% compared to the baseline state of 40°C.

[0064] Based on the deviation between the current real-time state data and the baseline state, the system calculates the feedforward compensation adjustment. Specifically, this includes extracting the current oil viscosity and density from the real-time state data and calculating the deviation between the oil viscosity, density, and the baseline state data. The system then processes this deviation based on a preset compensation rule to generate a specific control signal adjustment. This compensation rule can be defined as follows: when the system senses a decrease in real-time oil viscosity relative to the baseline state, it will reduce the intensity of the initial control command by a preset proportional coefficient. The magnitude of this proportional coefficient is linearly positively correlated with the detected viscosity decrease. For example, the proportional coefficient, preset based on experimental calibration, can be set to reduce the intensity of the initial control command by 2% for a 10% decrease in oil viscosity. Therefore, if the system detects a 30% decrease in current oil viscosity, the intensity of the initial control command will be reduced by 6%, preemptively offsetting any pressure overshoot that might result from changes in oil state. Finally, the system superimposes the control signal adjustment and the initial control output sequence to obtain the feedforward compensated control output sequence.

[0065] Using potential fault signatures, adaptive feedback correction is performed on the control output sequence after feedforward compensation. This adaptive feedback correction manifests as real-time adjustment of PID controller parameters. For example, if the potential fault signature is "increased risk of internal leakage in the main pump or main valve," the controller increases the integral gain and decreases the derivative gain through an online algorithm. The adjustment magnitude is proportional to the current performance degradation index. For instance, when the performance degradation index increases from 0 to 1, the integral gain increases by a maximum of 50% from the baseline value, and the derivative gain decreases by a maximum of 30%. When the system detects that the current performance degradation index is 0.58, the integral gain will increase by approximately 29% from the baseline value, and the derivative gain will decrease by approximately 17.4%.

[0066] When responding to and executing scheduling commands, the distributed intelligent actuator uses the sequence generated after feedforward compensation and adaptive feedback correction as the final control output sequence. This final control output sequence is then sent to the physical actuator of hydraulic press A2 to drive the stamping operation. By combining feedforward compensation and adaptive feedback correction, the system ensures that the work commands are executed with high precision and stability despite the dual challenges of changing working conditions and continuous equipment aging. This enhances the system's robustness and reduces the scrap rate and energy consumption caused by control inaccuracies.

[0067] Furthermore, when performing adaptive feedback correction in dynamic compensation management, a control strategy switching mechanism based on fault modes is introduced. The system queries and matches a preset control strategy library based on identified potential fault signatures, automatically selecting and switching to the optimal control strategy corresponding to the potential fault signature. This control strategy library includes adaptive PID algorithms, model predictive control, and robust control. The matching and switching are based on preset rules. For example, when the fault signature is "control loop parameter drift" or "servo valve dynamic characteristic degradation," the system switches to the adaptive PID algorithm; when the confidence level of the fault signature is higher than a threshold and it indicates "increased risk of internal leakage in the main pump or main valve," the system switches to model predictive control; when the system faces increased sensor noise or unmodeled external disturbances, it can switch to robust control. Through this control strategy switching mechanism, targeted control algorithms are used to correct specific physical characteristic changes caused by equipment aging, ensuring the system executes work instructions with high precision and stability. This reduces processing scrap rates and energy consumption caused by control inaccuracies, enhances the production system's adaptability to long-term performance degradation and its overall robustness, and ensures long-term product quality stability.

[0068] This invention assesses equipment performance degradation and quantifies overall operating costs to generate a risk- and cost-optimal resource allocation strategy, thereby improving the economy and foresight of production scheduling decisions. Simultaneously, through dynamic compensation and closed-loop adjustment of distributed intelligent actuators, it achieves precise execution of work instructions, enhancing the system's robustness and reliability under varying operating conditions.

[0069] Example 2

[0070] This invention applies the intelligent flow scheduling method for high-performance oil circuits in sheet metal pipelines proposed in this invention to sheet metal processing workshop B, which has a central high-performance oil circuit station. (See attached document.) Figure 2 The central high-performance hydraulic station simultaneously provides power to hydraulic presses B1, B2, and B3. The production plans of the three hydraulic presses are different, and their demand for hydraulic resources is coupled and competitive. Hydraulic resources include flow rate and pressure.

[0071] The core of the central high-performance hydraulic station is a variable displacement pump unit with a rated pressure of 25MPa and a maximum output flow rate of 400L / min. Within a 60-second production cycle, hydraulic press B1 performs high-precision forming tasks: the first 5 seconds involve rapid downward movement, requiring a flow rate of 200L / min and a pressure of 5MPa; from 5 to 15 seconds, it presses and holds pressure, requiring a flow rate of 20L / min and a pressure of 20MPa, with an accuracy requirement of ±0.2MPa; from 15 to 20 seconds, it rapidly returns, requiring a flow rate of 180L / min and a pressure of 4MPa.

[0072] Hydraulic press B2 performs routine stamping tasks: rapid descent in 10 to 14 seconds, requiring a flow rate of 150 L / min and a pressure of 5 MPa; stamping in 14 to 25 seconds, requiring a flow rate of 30 L / min and a pressure of 18 MPa; and rapid return in 25 to 29 seconds, requiring a flow rate of 140 L / min and a pressure of 4 MPa.

[0073] Hydraulic press B3 performs standby and auxiliary tasks: maintain standby for 0 to 60 seconds, requiring a flow rate of 5L / min and a pressure of 2MPa.

[0074] The system has established a resource optimization model and corresponding distributed intelligent actuators, including a central oil circuit station and three hydraulic presses (B1, B2, and B3). The maximum output flow and rated pressure of the central oil circuit station are entered into the system as key parameters. The resource optimization model receives the production plans of hydraulic presses B1, B2, and B3, analyzes the process steps, and generates three independent time-series load demand sequences. The model superimposes these three time-series load demand sequences based on a time axis vector to generate a global total load prediction curve. By analyzing the global total load prediction, the model identifies potential resource contention conflicts between seconds 10 and 14. The system acquires performance degradation data for each unit to determine whether hydraulic presses B1, B2, and B3 are in a healthy state.

[0075] An optimization problem is constructed, comprising an objective function and constraints. The objective function is to minimize the sum of the total energy consumption and potential quality risk costs of the central oil pipeline station. The constraints are: the instantaneous output flow rate and pressure of the central high-performance oil pipeline station satisfy the global total load prediction curve; and the pressure fluctuation of B1 during the pressure holding phase is less than ±0.2 MPa. By quantifying energy consumption and quality risk as costs and using them as optimization objectives, the system can proactively avoid resource conflicts while ensuring high-precision tasks. Under multiple complex constraints, it finds the most economically optimal scheduling scheme, thereby improving resource utilization efficiency.

[0076] After optimizing the solver calculations, resource allocation strategies for different devices are generated. For example, the central hydraulic station is instructed to increase the system pressure from 20MPa to 20.5MPa at the 9th second, improving the pump's response readiness level as feedforward compensation for the flow demand of B2. New work instructions are generated: delaying the rapid descent start-up time of B2, smoothing the descent speed curve, reducing peak flow, and extending the descent time; keeping the original work instructions for hydraulic press B1 unchanged, and the corresponding intelligent actuator entering a high-precision closed-loop adjustment mode. By transforming the coupled production plans into coordinated work instructions through the resource optimization model, the resource competition and conflict problems caused by overlapping time sequences among multiple devices are resolved, improving resource allocation efficiency in complex coupled scenarios and enhancing the synergy and stability of the production system.

[0077] Each distributed intelligent actuator receives and executes the above-mentioned operation instructions; the central oil circuit station pressurizes in advance according to the instructions; the actuator of B2 controls the servo valve according to the corrected time and speed curve.

[0078] Example 3

[0079] The embodiments of this invention apply the intelligent flow scheduling method for high-performance oil circuits in sheet metal pipelines proposed in this invention to a sheet metal processing workshop C with hydraulic presses C1 and C2 having different performance characteristics.

[0080] Initially, hydraulic press C1 has a performance degradation index of 0.05, indicating good condition. It is currently executing a production task of 500 "standard parts A" with a single-piece time of 30 seconds and a task priority of "normal". Hydraulic press C2 has a performance degradation index of 0.28, exhibits a potential fault signature of "increased risk of internal leakage in the main valve", has a response delay greater than the baseline, and is currently executing a production task of 1000 "auxiliary parts B" with a single-piece time of 25 seconds and a task priority of "normal".

[0081] See Figure 3 The production management system sends an expedited order via a data interface: to produce 10 high-precision parts X, requiring extremely high precision and delivery within half an hour. The resource optimization model receives the expedited order from the production management system and parses it into specific process steps including delivery time and process precision requirements. Simultaneously, the resource optimization model obtains the latest performance degradation data for C1 and C2, confirming that C1 meets the process precision requirements for part X, while C2 does not. After identifying the "expedited" and "high-precision" attributes, the system adjusts the weight of the comprehensive operating cost function in real time; for the part X task, the weight of risk cost is increased. The resource optimization model uses the updated cost function as the objective function to simulate various possible rescheduling schemes and quantify the comprehensive operating cost.

[0082] Simulation Scenario 1: Place the urgent task X in the task queue of hydraulic press C1. Wait for hydraulic press C1, which is in good working order, to complete its current task before starting production of part X. The overall operating cost of this scenario includes current operating costs and delay risk costs. Current operating costs include the routine resource depletion and energy costs incurred by C1 and C2 in executing their tasks. The delay risk cost specifically refers to the fact that urgent task X requires delivery within half an hour; waiting for C1 to complete its current task will cause a delay in urgent task X. The system quantifies this delay as a delay risk cost. This scenario ensures the continuity of routine tasks but cannot meet the critical time constraints of high-priority tasks, leading to increased risk costs.

[0083] Simulation Scheme 2: Assign the urgent task X to hydraulic press C2, whose performance degradation data is 0.28. The overall operating cost of this scheme includes current operating costs and quality risk costs. The current operating cost is the basic cost of system operation. The quality risk cost specifically refers to the following: the system obtains the performance degradation index and potential fault signature of hydraulic press C2 through performance degradation data, determining that C2's performance status does not meet the high-precision requirements of part X. Production of C2 will lead to an increased scrap rate and potential quality problems. The system quantifies these quality problems as high quality risk costs. This scheme meets the time requirements but cannot meet the product quality requirements, resulting in increased risk costs.

[0084] Simulation Scheme 3: The current task of hydraulic press C1, which is in good working order, is prioritized for urgent task X, and then the production task of "Standard Part A" is resumed. The comprehensive operating cost of this scheme includes current operating costs, interruption and mold change costs, and the delay cost of task A. The current operating cost is the basic cost of system operation; the interruption and mold change costs are the production downtime and resource consumption costs caused by production interruption, mold replacement, and equipment restart. The delay cost of task A is the cost incurred due to the postponement of the original plan for Standard Part A. The task priority of Standard Part A is "regular," and the delay cost is lower than the delay risk cost of urgent task X.

[0085] The overall operating cost of simulation scheme 3 includes the added controllable interruption and mold replacement costs, avoiding the uncontrollable delay and quality risk costs of simulation schemes 1 and 2. Calculations show that the overall operating cost of scheme 3 is lower than the first two schemes. The system determines the optimal scheme and generates a new resource allocation strategy: C1 safely suspends the current production of "standard part A" and records the breakpoint information; C1 executes the production task of high-precision part X; the status of the interrupted "standard part A" task is updated to "suspended" and re-queued in the production queue after C1 completes task X. By dynamically adjusting the weight of risk items in the overall operating cost, task attributes are quantified into economic indicators, enabling the system to make optimal trade-offs among multiple potential schemes through cost calculation when dealing with emergencies, thus improving the flexibility and intelligence of scheduling decisions.

[0086] After receiving the instruction, the distributed intelligent actuator of hydraulic press C1 pauses the current task and guides the automation system to complete the mold change within 5 minutes; hydraulic press C1 completes the production of 10 qualified parts X within the time limit, and then automatically resumes the production of standard part A.

[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent flow scheduling of high-performance oil circuits in sheet metal pipeline networks, characterized in that, include: Based on the sheet metal pipeline network containing schedulable units and production management information, a distributed intelligent execution body and resource optimization model are established. The production management information includes historical performance baselines and production plans; The operation data of the schedulable unit is obtained, and the distributed intelligent executor evaluates the operation data to generate real-time status data; Calculate the current output state for the preset test instructions, analyze the drift trend of the current output state from the historical performance baseline, and generate performance degradation data of the schedulable unit; Based on the production plan and the performance degradation data, the comprehensive operating cost is quantified and generated. The comprehensive operating cost is used as the objective function. The resource optimization model calculates and generates a resource allocation strategy. Based on the resource allocation strategy, work instructions are generated. The distributed intelligent executor combines the job instructions, real-time status data, and performance degradation data to perform dynamic compensation management and generate scheduling instructions; the schedulable unit dynamically tracks the job instructions based on the scheduling instructions. The distributed intelligent actuator includes a state evaluation unit and a performance trend prediction unit. The state evaluation unit includes an information network model that receives the operating data and generates real-time state data. The information network model includes a fluid state model. The performance trend prediction unit includes an online learning algorithm that analyzes the output state of preset test job instructions and historical performance baselines to generate performance degradation data. Based on the job instructions and by fusing the real-time state data and the performance degradation data, the unit performs the dynamic compensation management, calculates the final control output, and generates scheduling instructions. The process of generating real-time status data includes: using the operating data as input to the information network model, the operating data including temperature data and pressure data; the information network model performing a nonlinear transformation on the temperature data and pressure data to generate preliminary inference values ​​for real-time status data; the real-time status data including oil viscosity and density; calculating reference values ​​for real-time status data through a fluid state model, the fluid state model including a viscosity-temperature relationship model and a density-temperature-pressure relationship model; comparing the preliminary inference values ​​and the reference values ​​for real-time status data and calculating the difference; determining weights based on the difference; and obtaining the final status data by weighted summation of the preliminary inference values ​​and the reference values ​​for real-time status data. The process of generating a resource allocation strategy includes: parsing the production plan to generate specific process steps, obtaining the hydraulic demand curves for the specific process steps, generating a time-series load demand sequence and a predicted total energy consumption demand based on the hydraulic demand curves; combining the predicted total energy consumption demand and the performance degradation data to quantitatively evaluate the performance status, operating efficiency, and failure risk of the schedulable unit, determining resource loss costs, energy costs, and risk costs, and weighted summing the resource loss costs, energy costs, and risk costs to generate a quantitative comprehensive operating cost; using the total output of the schedulable unit satisfying the time-series load demand sequence as a constraint, constructing an objective function, with minimizing the comprehensive operating cost as the optimization objective, and solving the objective function to generate the resource allocation strategy; The execution of dynamic compensation management to generate scheduling instructions includes: parsing the job instruction to obtain the target trajectory; inputting the target trajectory into a simplified dynamic model to generate an initial control output sequence, wherein the simplified dynamic model includes the state parameters and response characteristics of the schedulable unit in the baseline state; performing feedforward compensation based on the deviation between the real-time state data and the baseline state system data to obtain a feedforward compensated control output sequence; using the potential fault signature of the performance degradation data to perform adaptive feedback correction on the feedforward compensated control output sequence; calculating the final control output sequence; and generating scheduling instructions.

2. The intelligent flow scheduling method for high-performance oil circuits in sheet metal pipelines according to claim 1, characterized in that, The resource optimization model is built based on the demand forecasting model and optimization algorithm, and is realized through the data integration, optimization decision and instruction generation stages. The data integration stage receives and integrates the production plan and the performance degradation data to generate a global decision view that includes the pipeline status and future task load. In the optimization decision-making phase, based on the global decision view, a forward-looking optimization solution is performed to generate the resource allocation strategy, which includes the running roles and work tasks of schedulable units; The instruction generation phase parses the running roles and tasks into job instructions containing timing and physical parameters.

3. The intelligent flow scheduling method for high-performance oil circuits in sheet metal pipelines according to claim 1, characterized in that, The historical performance baseline is the mapping relationship between the baseline job and the required resource consumption when the schedulable unit is in the baseline state.

4. The intelligent flow scheduling method for high-performance oil circuits in sheet metal pipelines according to claim 1, characterized in that, The process of generating performance degradation data includes: calculating a sequence signal and obtaining the required current output state based on a preset test command; comparing the current output state with a historical performance baseline to generate a real-time deviation value sequence; filtering out random noise from the real-time deviation value sequence to obtain a smooth curve, and extracting a dynamic drift trend signal by applying linear regression trend analysis to the smooth curve; extracting feature parameters to generate a feature vector based on the dynamic drift trend signal, the feature parameters including average bias, response delay, overshoot and oscillation amplitude, and drift rate; normalizing the feature parameters using the mapping model in the online learning algorithm and weighted summing to generate a current performance degradation index and a future degradation rate, and matching the feature parameters with a preset fault rule base to identify potential fault signatures; calculating the predicted remaining service life based on the current performance degradation index, the future degradation rate, and a preset failure threshold; and integrating the current performance degradation index, the future degradation rate, the potential fault signature, and the predicted remaining service life into performance degradation data output.

Images