A discrete manufacturing facility efficiency maximization strategy determination system and method

Abstract

The application discloses a discrete manufacturing equipment efficiency maximization strategy determination system and method, effectively achieves the purpose of improving equipment utilization, and comprises a multi-source data acquisition unit, an intelligent decision unit, an execution control unit and a feedback optimization unit.The multi-source data acquisition unit comprises a multi-source data acquisition module and an edge computing module.The intelligent decision unit comprises a prediction module, a multi-objective optimization module and a weight fusion module.The execution control unit comprises a strategy issuing module, an adaptive adjustment module and a dynamic scheduling module.The feedback optimization unit comprises a digital twin simulation module and an incremental learning module.The application has novel structure, ingenious design and simple operation, effectively increases the common manufacturing efficiency of manufacturing equipment, reduces energy consumption cost, improves delivery punctuality rate and reduces carbon emission.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of equipment management method, and relates to a discrete manufacturing equipment efficiency maximization strategy determination system and method. BACKGROUND

[0002] Discrete manufacturing (such as automobile parts, electronic assembly and the like) has the characteristics of complex production process, many product types, large order fluctuation and the like, and the equipment efficiency directly determines the production capacity and cost competitiveness of an enterprise. Traditional efficiency optimization mainly relies on manual experience to adjust equipment parameters or a single dimension index (such as maximizing equipment utilization), and it is difficult to balance economic, technical and environmental targets in a dynamic production environment. With the development of intelligent manufacturing, how to realize global optimization of equipment efficiency through multi-source data fusion and intelligent decision-making has become a core challenge in the industry.

[0003] In addition, existing efficiency optimization mostly ignores multi-dimensional constraints such as order urgency, customer priority, etc., which easily leads to delivery delay or loss of penalty; and does not consider dynamic factors such as equipment health state and time-of-use electricity price, which may cause equipment failure or sharp increase in energy cost in long-term operation, and therefore a discrete manufacturing equipment efficiency maximization strategy determination system and method considering order urgency, customer priority and dynamic factors such as equipment health state and time-of-use electricity price is needed to solve the above problems. SUMMARY

[0004] In view of the above problems, the application provides a discrete manufacturing equipment efficiency maximization strategy determination system and method, which solves the problems in the prior art.

[0005] In order to achieve the above purpose, the technical scheme adopted by the application is as follows:

[0006] A discrete manufacturing equipment efficiency maximization strategy determination system comprises:

[0007] A multi-source data acquisition unit comprising a multi-source data acquisition module for collecting equipment data, order data and environmental data inside a factory building and an edge computing module for data cleaning;

[0008] An intelligent decision-making unit comprising a prediction module, a multi-objective optimization module and a weight fusion module;

[0009] The prediction module is used for predicting order demand, equipment health degree and equipment energy cost;

[0010] The multi-objective optimization module is used for evaluating the economy, technology and environment of the factory building;

[0011] The weight fusion module is configured to combine the subjective weight and the objective weight of the prediction module, the economy, the technology and the environment, and perform weighted fusion, so as to finally generate an optimization strategy of multi-objective optimization.

[0012] The execution control unit comprises a strategy issuing module, an adaptive adjustment module and a dynamic scheduling module.

[0013] The strategy issuing module is configured to send the generated optimization strategy to the adaptive adjustment module after transcoding.

[0014] The adaptive adjustment module is configured to adjust the device running parameters to balance the efficiency and the energy consumption.

[0015] The dynamic scheduling module is configured to adjust the order queue.

[0016] The feedback optimization unit comprises a digital twin simulation module and an incremental learning module; the digital twin simulation module is configured to simulate the optimization strategy and verify the production capacity deviation; and the incremental learning module is configured to feed back the collected daily actual production data to the prediction module and update the prediction module in a timely manner.

[0017] Preferably, the multi-source data acquisition module comprises a vibration sensor, a device on-off state sensor, a current sensor, an enterprise resource planning system, an API interface and a temperature and humidity sensor.

[0018] The application further discloses a discrete manufacturing equipment efficiency maximization strategy determination method based on a discrete manufacturing equipment efficiency maximization strategy determination system, and comprises the following steps:

[0019] A preparation stage and system deployment are performed to configure and calibrate the data acquisition unit, the software and the interface of the system.

[0020] In step S1, multi-source data acquisition and edge preprocessing are performed; the device data, the order data and the environmental data are collected through the multi-source data acquisition module, and the collected data are cleaned through the edge computing module.

[0021] In step S2, multi-dimensional prediction is performed; the order demand, the device health degree and the device energy consumption cost are predicted through the prediction module.

[0022] In step S3, a multi-objective optimization strategy is generated; an economy, a technology and an environment evaluation index system is constructed through the multi-objective optimization module; the subjective weight and the objective weight are fused through the weight fusion module to generate a comprehensive weight; and finally, a plurality of Pareto optimization strategies are generated through the multi-objective optimization module and sent to the strategy issuing module.

[0023] Step S4, policy execution and dynamic adjustment, after the optimization policy is received by the policy issuing module, the policy issuing module converts the optimization policy into device control instructions through the industrial automation standard protocol and sends it to the adaptive adjustment module, which monitors the device load in real time and dynamically adjusts the device processing parameters; the production queue is reordered according to customer level and penalty cost through the dynamic scheduling module;

[0024] Step S5, digital twin verification and model iteration, simulate the execution of the optimization policy on the digital twin platform through the digital twin simulation module, calculate the deviation between the theoretical capacity and the actual capacity, and trigger policy backtracking when the deviation between the actual capacity and the theoretical capacity is >10%; through the incremental learning module, the daily actual production data collected is fed back to the prediction module, and the prediction module is updated incrementally every day.

[0025] Preferably, in step S1, the multi-source data acquisition includes the following steps:

[0026] Step S11, real-time acquisition of device vibration, current, temperature and humidity data through the multi-source data acquisition module, synchronization of production planning scheduling, order urgency, profit rate and delivery cycle of the enterprise resource planning system;

[0027] Step S12, cleaning of raw data by the edge computing module using edge computing nodes to extract key features.

[0028] Preferably, in step S12, the key features include one or more of device comprehensive efficiency, order completion rate, and actual energy consumption cost.

[0029] Preferably, in step S2, multi-dimensional prediction includes:

[0030] Order demand prediction: using a time series model to analyze historical order seasonal fluctuations and market trends, outputting future weekly order quantity prediction results;

[0031] Device health prediction: building a fault warning model based on a long short-term memory network, inputting vibration frequency spectrum peak value and current fluctuation variance, and outputting fault probability threshold alarm;

[0032] Energy consumption cost prediction: combining the production planning scheduling collected by the multi-source data acquisition module and the time-of-use price curve to predict the total energy consumption cost in each period and generate peak-valley energy consumption optimization suggestions.

[0033] Preferably, in step S3, the following steps are included:

[0034] Step S31, determining the subjective weights of the prediction module, economy, technology, and environment through the multi-objective optimization module;

[0035] Step S32, based on the fluctuation of each index in the historical data, the objective weight is calculated by entropy weight method, the subjective weight deviation is corrected, and the final weight is obtained;

[0036] Step S33, a non-dominated sorting genetic algorithm is used to generate three kinds of Pareto optimization strategies based on the final weight obtained in step S32, with the goal of maximizing the comprehensive evaluation value.

[0037] Preferably, in step S31, the influencing factors for evaluating the economy include unit production cost and order profit rate, the influencing factors for evaluating the technology include equipment utilization rate and order on-time delivery rate, and the influencing factors for evaluating the environment include unit product carbon emission and peak-valley electricity price energy consumption ratio.

[0038] Preferably, in step S4, when the adaptive adjustment module dynamically adjusts the equipment processing parameters, the spindle speed and the feed rate are adjusted based on the real-time load and the processing quality threshold.

[0039] Compared with the prior art, the present application has the following beneficial effects:

[0040] 1. The present application has a multi-objective optimization module, which is used to evaluate the economy, technology and environment of the factory building respectively, and then facilitates the generation of multi-objective optimization strategies.

[0041] 2. The present application has a dynamic scheduling module, which facilitates the adjustment of the order queue. DETAILED DESCRIPTION

[0042] Figure 1 The present application is a system structure diagram. CONCRETE EMBODIMENT

[0043] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0044] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application. Figure 1 The specific embodiments of the present application will be further described in detail.

[0045] As shown in the figure, in order to realize the full-dimensional collection of equipment running state, production task and environmental data, a discrete manufacturing equipment efficiency maximization strategy determination system comprises: Figure 1

[0046] ​The multi-source data acquisition unit comprises a multi-source data acquisition module for collecting equipment data, order data and environment data inside the factory building and an edge computing module for data cleaning;

[0047] It should be noted that the multi-source data acquisition module comprises a vibration sensor (for monitoring the mechanical state of the equipment), an equipment startup and shutdown state sensor (for recording the equipment start and stop time), a current sensor (for collecting motor load), an enterprise resource planning system (ERP) interface (for obtaining production plan and order data), an API interface (for connecting with the MES system) and a temperature and humidity sensor (for monitoring environmental parameters); the edge computing module is based on an edge computing node, adopts a sliding window mean filter to eliminate noise data, and removes outliers through a DBSCAN clustering algorithm, so that key features such as cleaned equipment comprehensive efficiency (0EE), order completion rate and actual energy consumption cost can be obtained;

[0048] The intelligent decision-making unit comprises a prediction module, a multi-objective optimization module and a weight fusion module; the prediction module is used for predicting order demand, equipment health degree and equipment energy consumption cost, wherein the order demand prediction is based on an ARIMA time series model to analyze the seasonal fluctuation and market trend of historical orders and output order quantity prediction for the next 7 days; the equipment health degree prediction adopts an LSTM neural network to input vibration frequency spectrum peak value (0-5 kHz) and current fluctuation variance (threshold ± 10%) and output future 24-hour fault probability (threshold > 85% to trigger an alarm); the energy consumption cost prediction combines the time-of-use electricity price curve (peak / flat / valley period) and production plan scheduling to predict total energy consumption cost in each period and generate peak-valley transfer suggestions (in this embodiment, high energy consumption processes can be adjusted to the valley electricity price period);

[0049] The multi-objective optimization module is used for evaluating the economy, technology and environment of the factory building respectively; wherein the economy: unit production cost (yuan / piece), order profit rate (%); the technology: equipment utilization rate (%), order on-time delivery rate (%); the environment: unit product carbon emission (kg / piece), peak-valley electricity price energy consumption ratio (%); a non-dominated sorting genetic algorithm with elitist strategy (NSGA-II) is adopted to generate a Pareto frontier solution set;

[0050] The weight fusion module is used for combining the subjective weight and objective weight of the prediction module, economy, technology and environment for weighted fusion to finally generate an optimization strategy of multi-objective optimization; in this embodiment, the weight fusion module calculates the subjective weight based on an expert scoring matrix through an AHP hierarchical analysis method, combines the objective weight calculated through an entropy weight method, and generates a comprehensive weight through weighted fusion in a 6:4 ratio;

[0051] The execution control unit comprises a strategy issuing module, an adaptive adjustment module and a dynamic scheduling module;

[0052] The policy issuing module converts the optimization strategy into device control instructions (such as G code, Modbus instructions) through the OPC UA protocol, and sends the converted optimization strategy to the adaptive adjustment module;

[0053] The adaptive adjustment module is used for adjusting the device operation parameters to balance efficiency and energy consumption. In the embodiment, the adaptive adjustment module dynamically adjusts the spindle speed (range 500-3000 rpm) and the feed rate (range 10-50 mm / s) based on real-time load (CPU utilization > 80% triggers) and processing quality threshold (tolerance ± 0.05 mm);

[0054] The dynamic scheduling module is used for adjusting the order queue. In the embodiment, the dynamic scheduling module generates a production queue ranking by using a weighted priority algorithm, in which the customer level weight accounts for 70% (VIP customers have priority) and the penalty cost weight accounts for 30%.

[0055] The feedback optimization unit includes a digital twin simulation module and an incremental learning module.

[0056] The digital twin simulation module is used for simulating the optimization strategy and verifying the capacity deviation. In the embodiment, the digital twin simulation module simulates the strategy execution in Siemens NX, calculates the theoretical capacity and actual capacity deviation, and triggers the strategy backtracking when the deviation value is greater than 10%.

[0057] The incremental learning module uses a stochastic gradient descent (SGD) algorithm to feed the actual production data to the prediction module daily to update the model parameters, and is used for feeding the collected daily actual production data to the prediction module and updating the prediction module.

[0058] The application also discloses a discrete manufacturing equipment efficiency maximization strategy determination method based on a discrete manufacturing equipment efficiency maximization strategy determination system. Figure 1 As shown in the figure, in order to maximize the efficiency of different manufacturing equipment, the following steps are included:

[0059] The preparation stage, system deployment, configures and calibrates the data acquisition unit, software and interface of the discrete manufacturing equipment efficiency maximization strategy determination system, including sensor installation, enterprise resource planning system (ERP) interface configuration, and calibrating the clock synchronization accuracy (error < 1 ms) of the edge computing node;

[0060] Step S1, multi-source data acquisition and edge preprocessing, the device data, order data and environment data are collected through the multi-source data acquisition module, and the collected data are cleaned through the edge computing module;

[0061] In step S1, multi-source data acquisition includes the following steps:

[0062] Step S11, real-time acquisition of equipment vibration (sampling rate 10 kHz), current (accuracy ±0.5%), temperature and humidity data (resolution 0.1℃) by the multi-source data acquisition module, synchronization of enterprise resource planning (ERP) system production planning scheduling, order urgency (high / medium / low), profit margin (15% for high-quality orders) and delivery cycle (≤7 days for urgent orders);

[0063] Step S12, cleaning of raw data by the edge computing module using edge computing nodes to extract key features;

[0064] It should be noted that the edge computing module refers to a device deployed at the edge of the network with computing, storage, communication and other capabilities, which is a micro data center in this embodiment, capable of providing data storage and computing services, and cleaning raw data by removing noise, filling missing values and correcting abnormal values;

[0065] The key features include one or more of the device comprehensive efficiency (0EE=availability rate x performance rate x good product rate), order completion rate (actual completion amount / planned amount), and actual energy consumption cost (kWh x electricity price);

[0066] Step S2, multi-dimensional prediction, predicting order demand, equipment health degree and equipment energy consumption cost;

[0067] In step S2, multi-dimensional prediction includes the following steps:

[0068] Step S21, order demand prediction: using a time series model to analyze historical order seasonal fluctuations and market trends, outputting future weekly order quantity prediction results, in this embodiment, using an ARIMA model, R2>0.85;

[0069] Step S22, equipment health degree prediction: building a fault warning model based on a long short-term memory network, inputting vibration frequency spectrum peak value and current fluctuation variance, outputting fault probability threshold alarm, in this embodiment, using an LSTM model, AUC>0.9;

[0070] Step S23, energy consumption cost prediction: combining production planning scheduling and time-of-use price curve collected by the multi-source data acquisition module to predict total energy consumption cost in each period and generate peak-valley energy consumption optimization suggestions;

[0071] Step S3, multi-objective optimization strategy generation, building an economic, technical and environmental evaluation index system; integrating subjective and objective weights to generate a comprehensive weight; finally generating multiple Pareto optimization strategies through a multi-objective optimization module;

[0072] In step S3, the following steps are included:

[0073] In step S31, the subjective weights of the artificial judgment prediction module, economy, technology, and environment are determined by the multi-objective optimization module.

[0074] In step S31, the influencing factors for evaluating economy include unit production cost and order profit rate, the influencing factors for evaluating technology include equipment utilization rate and order on-time delivery rate, and the influencing factors for evaluating environment include unit product carbon emission and peak-valley electricity price energy consumption ratio. In this embodiment, the subjective weight proportions are calculated by AHP hierarchical analysis method as follows: economy (weight 40%), technology (35%), and environment (25%);

[0075] In step S32, the objective weights are calculated based on the fluctuation of each index in the historical data by the entropy weight method, and the subjective weight deviation is corrected. In this embodiment, the comprehensive weights (economy 38%, technology 36%, and environment 26%) are generated after correction by the entropy weight method.

[0076] In step S33, the non-dominated sorting genetic algorithm is used to maximize the comprehensive evaluation value as the target, and three kinds of Pareto optimization strategies are generated. In this embodiment, three kinds of Pareto optimization strategies (strategy A focuses on economy, strategy B balances the three, and strategy C focuses on environment) are generated by the NSGA-II algorithm.

[0077] In step S4, strategy execution and dynamic adjustment, after the optimization strategy is received by the strategy issuing module, the strategy issuing module converts the optimization strategy into device control instructions through the industrial automation standard protocol and sends it to the adaptive adjustment module. The adaptive adjustment module monitors the device load in real time and dynamically adjusts the device processing parameters. The production queue is reordered by the dynamic scheduling module according to the customer level and the penalty cost.

[0078] In step S4, when the adaptive adjustment module dynamically adjusts the device processing parameters, the spindle speed and feed rate are adjusted based on the real-time load and processing quality threshold (e.g., the spindle speed is reduced by 15% to balance energy consumption).

[0079] In step S5, digital twin verification and model iteration, the execution of the optimization strategy is simulated on the digital twin platform by the digital twin simulation module, and the deviation between the theoretical capacity and the actual capacity is calculated. When the deviation between the actual capacity and the theoretical capacity is >10%, the strategy backtracking is triggered. Through the incremental learning module, the daily actual production data collected is fed back to the prediction module, and the prediction module is updated daily (learning rate η = 0.01).

[0080] The application has novel structure, ingenious design, simple and convenient operation, effectively improves the utilization rate of the equipment, further increases the common manufacturing efficiency of the manufacturing equipment, reduces the energy consumption cost, improves the delivery punctuality rate, and reduces the carbon emission.

[0081] Although the application has been described in detail with reference to the foregoing embodiments, those skilled in the art can modify the technical solutions recorded in the foregoing embodiments or make equivalent replacements to some technical features, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. A discrete manufacturing equipment efficiency maximization strategy determination system characterized by: The application relates to a factory intelligent production optimization system. The system comprises a multi-source data acquisition unit, an intelligent decision unit, an execution control unit, and a feedback optimization unit. The multi-source data acquisition unit comprises a multi-source data acquisition module for collecting equipment data, order data and environmental data inside a factory and an edge computing module for data cleaning. The intelligent decision unit comprises a prediction module, a multi-objective optimization module and a weight fusion module. The prediction module is used for predicting order demand, equipment health degree and equipment energy consumption cost. The multi-objective optimization module is used for evaluating the economy, technology and environment of the factory. The weight fusion module is used for combining subjective weights and objective weights of the prediction module, economy, technology and environment and performing weighted fusion to finally generate an optimization strategy of multi-objective optimization. The execution control unit comprises a strategy issuing module, an adaptive adjustment module and a dynamic scheduling module. The strategy issuing module is used for sending the generated optimization strategy to the adaptive adjustment module after conversion. The adaptive adjustment module is used for adjusting equipment operation parameters to balance efficiency and energy consumption. The dynamic scheduling module is used for adjusting an order queue. The feedback optimization unit comprises a digital twin simulation module and an incremental learning module. The digital twin simulation module is used for simulating the optimization strategy to verify production capacity deviation.

2. The system of claim 1, wherein: The incremental learning module is used for feeding back daily actual production data collected to the prediction module and performing periodic update on the prediction module.

3. A method for determining a strategy for maximizing efficiency of a discrete manufacturing device, comprising the system for determining a strategy for maximizing efficiency of a discrete manufacturing device according to any one of claims 1-2, characterized in that: The multi-source data acquisition module comprises a vibration sensor, an equipment on-off state sensor, a current sensor, an enterprise resource planning system, an API interface and a temperature and humidity sensor. The system comprises the following steps: A preparation stage and system deployment, wherein data acquisition units, software and interfaces of the system are configured and calibrated; Step S1, multi-source data acquisition and edge preprocessing, wherein equipment data, order data and environmental data are collected through a multi-source data acquisition module, and the collected data are cleaned through an edge computing module; Step S2, multi-dimensional prediction, wherein order demand, equipment health degree and equipment energy consumption cost are predicted through the prediction module; Step S3, multi-objective optimization strategy generation, wherein an economic, technical and environmental evaluation index system is constructed through the multi-objective optimization module; subjective weights and objective weights are fused through the weight fusion module to generate a comprehensive weight; A plurality of Pareto optimization strategies are finally generated through the multi-objective optimization module and are sent to the strategy issuing module; Step S4, strategy execution and dynamic adjustment, wherein after receiving the optimization strategy, the strategy issuing module converts the optimization strategy into equipment control instructions through an industrial automation standard protocol and sends the equipment control instructions to the adaptive adjustment module, the adaptive adjustment module monitors equipment load in real time and dynamically adjusts equipment processing parameters; the production queue is reordered according to customer grades and default cost through the dynamic scheduling module. Step S5, digital twin verification and model iteration, through the digital twin simulation module, the execution of the optimization strategy is simulated on the digital twin platform, the deviation between the theoretical production capacity and the actual production capacity is calculated, and when the deviation value of the actual production capacity and the theoretical production capacity is greater than 10%, the strategy is triggered back; through the incremental learning module, the daily actual production data collected is fed back to the prediction module, and the prediction module is updated incrementally every day.

4. The method of claim 3, wherein: In step S1, multi-source data acquisition includes the following steps: Step S11, real-time acquisition of equipment vibration, current, temperature and humidity data through the multi-source data acquisition module, synchronization of production planning scheduling, order urgency, profit rate and delivery cycle of the enterprise resource planning system; Step S12, cleaning of raw data by the edge computing module using edge computing nodes to extract key features.

5. The method of claim 4, wherein: In step S12, the key features include one or more of the device comprehensive efficiency, order completion rate, and actual energy consumption cost.

6. The method of claim 3, wherein: In step S2, multi-dimensional prediction includes: Order demand prediction: using time series model to analyze historical order seasonal fluctuations and market trends, outputting future weekly order quantity prediction results; Device health prediction: based on long short-term memory network to build fault warning model, input vibration frequency spectrum peak value, current fluctuation variance, output fault probability threshold alarm; Energy consumption cost prediction: combining production planning scheduling and time-of-use price curve collected by the multi-source data acquisition module, predicting total energy consumption cost in each period and generating peak-valley energy consumption optimization suggestions.

7. The method of claim 3, wherein: In step S3, the following steps are included: Step S31, determine the subjective weights of prediction module, economy, technology and environment through the multi-objective optimization module; Step S32, based on the fluctuation of each index in the historical data, calculate the objective weight by entropy weight method, correct the deviation of subjective weight, and get the final weight; Step S33, using non-dominated sorting genetic algorithm based on the final weight obtained in step S32, taking maximizing comprehensive evaluation value as the goal, generating three kinds of Pareto optimization strategies.

8. The method of claim 7, wherein: In step S31, the influencing factors for evaluating economy include unit production cost and order profit rate, the influencing factors for evaluating technology include equipment utilization rate and order on-time delivery rate, and the influencing factors for evaluating environment include unit product carbon emission and peak-valley electricity price energy consumption ratio.

9. The method of claim 3, wherein: In step S4, when the self-adaptive adjustment module dynamically adjusts the equipment processing parameters, based on real-time load and processing quality threshold, the spindle speed and feed rate are adjusted.

Images