Smart factory energy efficiency optimization method, system and equipment based on industrial Internet of Things

Abstract

The present invention discloses a method, system, and equipment for optimizing energy efficiency of a smart factory based on the Industrial Internet of Things, and relates to the technical field of the Industrial Internet of Things. The method includes: collecting production energy consumption data based on equipment operating parameters and production plan information; constructing a factory digital twin model to determine the equipment start-stop sequence and energy allocation strategy; deploying edge computing nodes, monitoring the equipment vibration spectrum, generating a state prediction diagram within a time sliding window, and performing adaptive compensation optimization; at the same time, formulating an energy flow network based on the equipment start-stop sequence and energy allocation strategy to perform dynamic energy adjustment. The present invention solves the technical problems of low energy utilization efficiency and insufficient equipment operation stability in the prior art of smart factories, and achieves the technical effect of realizing energy efficiency optimization of smart factories based on the Industrial Internet of Things, and improving energy utilization efficiency and equipment operation stability.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial Internet of Things, and in particular to a method, system, and equipment for optimizing energy efficiency in smart factories based on the industrial Internet of Things. Background Art

[0002] With the development of Industrial Internet of Things (IIoT) technology, smart factories are increasingly demanding refined and intelligent energy efficiency management. Traditional factories face challenges in energy management, such as inaccurate energy consumption data collection, poor matching between equipment energy consumption characteristics and production plans, a lack of dynamic optimization mechanisms for energy allocation, and delayed warnings of equipment failures. These issues result in low energy efficiency and make it difficult to adapt to the high-efficiency, stable, and low-consumption requirements of modern production.

[0003] Existing technologies have technical problems such as low energy utilization efficiency of smart factories and insufficient equipment operation stability. Summary of the Invention

[0004] This application provides a smart factory energy efficiency optimization method, system, and equipment based on the Industrial Internet of Things, which are used to solve the technical problems of low energy utilization efficiency and insufficient equipment operation stability in the existing technology of smart factories.

[0005] In view of the above problems, this application provides a smart factory energy efficiency optimization method, system and equipment based on the Industrial Internet of Things.

[0006] The first aspect of the present application provides a method for optimizing energy efficiency of a smart factory based on the Industrial Internet of Things, the method comprising:

[0007] Based on equipment operating parameters and production plan information, production energy consumption data including water resource energy consumption and power resource energy consumption within the same timestamp limit are collected; according to the equipment operating parameters and energy scheduling rule library, a factory digital twin model is constructed, and multi-objective optimization and adjustment are performed in combination with the production energy consumption data to determine the equipment start-stop timing and energy allocation strategy; edge computing nodes are deployed to monitor the equipment vibration spectrum, and the load change trend within the production batch cycle of the production plan information is matched with the abnormal spectrum fluctuation range through cosine similarity, and the early fault sign characteristics under the attention mechanism are set to generate a state prediction diagram within the time sliding window, and adaptive compensation optimization is performed; at the same time, based on the equipment start-stop timing and energy allocation strategy, an energy flow network is formulated to perform dynamic energy adjustment.

[0008] The second aspect of the present application provides a smart factory energy efficiency optimization system based on the Industrial Internet of Things, the system comprising:

[0009] The production energy consumption data acquisition module is used to collect production energy consumption data including water resource energy consumption and power resource energy consumption under the same timestamp limit based on equipment operating parameters and production plan information; the energy allocation strategy determination module is used to build a factory digital twin model based on the equipment operating parameters and energy scheduling rule library, and perform multi-objective optimization and adjustment in combination with the production energy consumption data to determine the equipment start-stop timing and energy allocation strategy; the adaptive compensation optimization module is used to deploy edge computing nodes, monitor the equipment vibration spectrum, match the load change trend within the production batch cycle of the production plan information through cosine similarity, set the early fault sign characteristics under the attention mechanism, generate a state prediction diagram within the time sliding window, and perform adaptive compensation optimization; the energy dynamic adjustment module is used to simultaneously formulate an energy flow network based on the equipment start-stop timing and energy allocation strategy, and perform energy dynamic adjustment.

[0010] The third aspect of the present application provides an electronic device, which includes: a processor; a memory for storing instructions executable by the processor; wherein the processor is used to execute the smart factory energy efficiency optimization method based on industrial Internet of Things provided in this application.

[0011] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0012] Based on equipment operating parameters and production plan information, production energy consumption data is collected. Based on these equipment operating parameters and the energy scheduling rule base, a factory digital twin model is constructed to determine equipment start-up and shutdown sequences and energy allocation strategies. Edge computing nodes are deployed to monitor equipment vibration spectra. Based on load variation trends within the production batch cycle of production plan information, early fault sign features are set using an attention mechanism, state prediction diagrams within a time sliding window are generated, and adaptive compensation optimization is performed. Simultaneously, an energy flow network is developed based on the equipment start-up and shutdown sequences and energy allocation strategies to dynamically adjust energy. This achieves the technical effect of optimizing energy efficiency in smart factories based on the Industrial Internet of Things, improving energy utilization efficiency and equipment operation stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0014] Figure 1 A flow chart of the method for optimizing energy efficiency in a smart factory based on the Industrial Internet of Things provided in an embodiment of the present application.

[0015] Figure 2 Schematic diagram of the structure of the smart factory energy efficiency optimization system based on the Industrial Internet of Things provided in an embodiment of the present application.

[0016] Figure 3 This is a schematic diagram of the structure of an electronic device provided in this application.

[0017] Explanation of the accompanying symbols: production energy consumption data acquisition module 10, energy allocation strategy determination module 20, adaptive compensation optimization module 30, energy dynamic adjustment module 40, processor 21, memory 22, input device 23, output device 24. DETAILED DESCRIPTION

[0018] This application provides a smart factory energy efficiency optimization method, system and equipment based on the Industrial Internet of Things, which is used to solve the technical problems of low energy utilization efficiency and insufficient equipment operation stability in the existing technology of smart factories.

[0019] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0020] Example 1, as Figure 1 As shown, the present application provides a method for optimizing energy efficiency of a smart factory based on the Industrial Internet of Things, the method comprising:

[0021] Step S100: Based on equipment operating parameters and production plan information, production energy consumption data including water resource energy consumption and power resource energy consumption within the same timestamp is collected.

[0022] Specifically, based on equipment operating parameters (such as equipment load rate, operating time, etc.) and production plan information (such as production batch cycle, process schedule, etc.), production energy consumption data under the same timestamp limit are collected. This data specifically includes water resource energy consumption and electricity resource energy consumption; during the collection process, these energy consumption data are bound to the factory production process and production batch cycle, and an energy consumption-process mapping relationship table is established. Through this table, the energy consumption differences of the same equipment in different production stages are identified, and the unit energy consumption benchmark value is extracted. At the same time, the benchmark value is used to automatically mark the energy efficiency abnormal processes and push them to the energy management terminal to ensure that the collected energy consumption data not only covers key resource types, but also can form an accurate association with the specific links of the production process, providing structured data support for subsequent energy efficiency analysis and optimization.

[0023] Step S200: Build a factory digital twin model based on the equipment operating parameters and energy scheduling rule library, and perform multi-objective optimization and adjustment in combination with the production energy consumption data to determine the equipment start-stop timing and energy allocation strategy.

[0024] Specifically, a factory digital twin model is constructed based on the collected equipment operating parameters (such as equipment load rate, operating status of different production stages, etc.) and the preset energy scheduling rule library, and an equipment energy consumption characteristic curve library containing energy consumption parameters of the equipment under different load rates is embedded in the model; then, combined with production energy consumption data (including water resources and electricity resource energy consumption data bound to production processes and batch cycles), the total energy consumption under different product combinations and equipment linkage modes is simulated to generate the energy consumption Pareto frontier and production capacity Pareto frontier; at the same time, the equipment load rate, number of production batches, and number of equipment linkage groups are used as variable parameters, and constraints are set to determine the effective sample set. The non-dominated solution set is obtained through non-dominated sorting, and finally the optimal balance point is selected from the solution set according to the production order priority to determine the equipment start and stop sequence (such as the equipment start-up sequence and operating time in different production stages) and the energy allocation strategy (such as the water resource and electricity resource allocation ratio of each process).

[0025] Step S300: Deploy edge computing nodes, monitor the vibration spectrum of the equipment, use the load change trend within the production batch cycle of the production plan information, match the abnormal spectrum fluctuation range through cosine similarity, set the early fault sign features under the attention mechanism, generate a state prediction diagram within the time sliding window, and perform adaptive compensation optimization.

[0026] Specifically, edge computing nodes are deployed at key equipment in the factory. The nodes integrate wavelet packet decomposition units, which can decompose the monitored equipment vibration spectrum into energy characteristics of multiple frequency bands; combined with the load change trend within the production batch cycle in the production plan information, the cosine similarity algorithm is used to match the abnormal intervals in the vibration spectrum that deviate from the normal fluctuation range, and then set the early fault sign characteristics under the attention mechanism; with the help of long short-term memory networks, the spectrum energy change trend within the production batch cycle is predicted, and combined with the normal fluctuation range, it is determined whether to trigger the equipment load reduction instruction, and at the same time, an equipment status prediction map within the time sliding window is generated. Based on this, the equipment operation is adaptively compensated and optimized to achieve early warning and intervention of potential equipment failures, ensure stable operation of the equipment and optimize energy efficiency.

[0027] Step S400: At the same time, an energy flow network is formulated based on the equipment start-stop sequence and energy distribution strategy to perform dynamic energy adjustment.

[0028] Specifically, an energy flow network is formulated based on the determined equipment start and stop timing and energy allocation strategy. The network covers the entire process of energy supply, transmission and consumption, and at the same time extracts key energy efficiency indicators such as transformer load rate and pipeline pressure loss. According to these key energy efficiency indicators, combined with the peak and valley signals of the power grid (such as the power supply conditions during peak and valley periods) and production plan information, the dynamic adjustment parameters (such as charging and discharging thresholds) and priority scheduling rules of the energy storage units connected to the industrial Internet of Things are configured to dynamically adjust the charging and discharging strategies of the energy storage units and the operating periods of the adjustable loads. By setting a two-dimensional reward function (the first dimension is the deviation rate between the real-time energy consumption and the theoretical minimum energy consumption, and the second dimension is the equipment response delay time), a deep deterministic policy gradient is used to learn the energy scheduling strategy in the continuous action space, and the energy scheduling rule base is iteratively updated with a reinforcement learning algorithm to achieve dynamic optimization and adjustment of the energy flow network.

[0029] In one possible implementation, step S400 further includes:

[0030] Step S410: Based on the equipment start-stop sequence and energy allocation strategy, an energy flow network is formulated, and key energy efficiency indicators are extracted. The key energy efficiency indicators include transformer load rate and pipeline pressure loss.

[0031] Step S420: According to the key energy efficiency indicators, combined with the peak and valley signals of the power grid and the production plan information, the dynamic adjustment parameters and priority scheduling rules of the energy storage unit are configured.

[0032] Specifically, based on the determined equipment start-up and shutdown sequences (such as equipment startup sequence, operating time and shutdown arrangements) and energy allocation strategies (such as the distribution ratio and time period planning of electricity and water resources in various production processes and equipment), an energy flow network covering the factory's energy supply source, transmission path and consumption nodes in each link is constructed to clearly present the flow logic and correlation relationship of energy in the entire production system; on this basis, key energy efficiency indicators that can reflect energy utilization efficiency and transmission status are extracted from the energy flow network, including transformer load rate (used to measure the load-bearing efficiency of power transmission equipment) and pipeline pressure loss (used to evaluate the degree of loss of fluid energy such as water resources during the transmission process). By extracting these indicators, a quantitative basis is provided for subsequent dynamic energy adjustments.

[0033] Based on the extracted key energy efficiency indicators such as transformer load rate and pipeline pressure loss, combined with the real-time peak and valley signals of the power grid (such as power supply capacity limitations during peak power consumption periods) and production plan information (such as the time nodes of production batches and the production capacity requirements of each process), the parameter configuration of the energy storage units connected to the Industrial Internet of Things is dynamically adjusted (including charging and discharging power thresholds, rate limits, etc.), and priority scheduling rules are formulated (such as the energy supply priority corresponding to key production processes and the charging and discharging priority of energy storage units during peak and valley periods) to ensure that the operation of the energy storage units is adapted to the actual energy efficiency status of the energy flow network, and can respond to the peak and valley fluctuations of the power grid and the rhythm of production plans, thereby improving the accuracy and efficiency of energy scheduling.

[0034] In one possible implementation, step S420 further includes:

[0035] Step S421: The energy storage unit is connected to the industrial Internet of Things.

[0036] Step S422: Dynamically adjust the charging and discharging strategy of the energy storage unit and the operating period of the adjustable load.

[0037] Step S423: using a reinforcement learning algorithm to iteratively update the energy scheduling rule base.

[0038] Specifically, the energy storage unit is deeply connected with the Industrial Internet of Things. This connection mechanism enables the energy storage unit to access various dynamic data collected by the Industrial Internet of Things in real time, including but not limited to real-time operating parameters of equipment, production plan change information, power grid peak and valley signals, energy supply and demand balance status, and energy consumption data of other related links; at the same time, the energy storage unit can also synchronously feedback its own real-time charging and discharging status, remaining storage capacity, health and other key information to the Industrial Internet of Things system, forming a two-way data interaction link, providing real-time and accurate data support for the subsequent dynamic adjustment of its charging and discharging strategy, coordinated scheduling with adjustable loads, and iterative optimization of the energy scheduling rule base, ensuring that the operation of the energy storage unit is always linked to the overall energy efficiency optimization goals of the smart factory.

[0039] Based on the connectivity between energy storage units and the Industrial Internet of Things (IIoT), combined with key energy efficiency indicators (such as transformer load factor and pipeline pressure loss), grid peak and valley signals, and production plan information obtained from the IIoT, the charging and discharging strategies of energy storage units and the operating time periods of adjustable loads are dynamically adjusted. The charging and discharging strategies must be adjusted to accommodate grid peak and valley fluctuations. For example, charging can be increased during off-peak periods to reserve energy, while prioritizing the release of stored energy during peak periods to alleviate power supply pressure. Adjustments to the operating time periods of adjustable loads must also align with the production schedule. By staggering the operating time of adjustable loads for non-critical processes, energy supply and demand are balanced, ensuring that energy efficiency is improved and energy consumption is reduced while meeting production needs.

[0040] When using a reinforcement learning algorithm to iteratively update the energy scheduling rule base, a two-dimensional reward function is first set. The first dimension is the deviation rate between real-time energy consumption and the theoretical minimum energy consumption, which is used to measure the gap between actual energy consumption and ideal optimal energy consumption. The second dimension is the device response delay time, which is used to evaluate the device's response speed to scheduling instructions. Based on this two-dimensional reward function, a deep deterministic policy gradient algorithm is adopted to continuously learn energy scheduling strategies in a continuous action space. Through continuous iterative optimization, the energy scheduling rule base can dynamically adapt to changes in factory energy supply and demand, peak and valley fluctuations in the power grid, and production plan adjustments, thereby improving the accuracy and efficiency of energy scheduling.

[0041] In one possible implementation, step S423 further includes:

[0042] Step S4231: Set a two-dimensional reward function, where the first dimension is the deviation rate between real-time energy consumption and theoretical minimum energy consumption, and the second dimension is the device response delay time.

[0043] Step S4232: Based on the two-dimensional reward function, deep deterministic policy gradient is used to learn the energy scheduling strategy in the continuous action space.

[0044] Specifically, a two-dimensional reward function is set as the basis for judging the iterative update of the energy scheduling rule library by the reinforcement learning algorithm. The first dimension is the deviation rate between the real-time energy consumption and the theoretical minimum energy consumption. The deviation rate is determined by calculating the ratio between the real-time energy consumption generated in the actual production process and the theoretical minimum energy consumption simulated based on the factory digital twin model and the equipment energy consumption characteristic curve library. It is used to quantify the degree to which the actual energy consumption deviates from the optimal state; the second dimension is the equipment response delay time, that is, the time interval from the issuance of the energy scheduling instruction to the execution of the instruction by the equipment and the completion of the state adjustment. It is used to measure the response efficiency of the equipment to the scheduling instruction. The setting of the two-dimensional reward function can comprehensively reflect the comprehensive performance of the energy scheduling strategy in energy efficiency optimization and execution efficiency, and provide a clear optimization goal for subsequent strategy learning.

[0045] Based on the above two-dimensional reward function, a deep deterministic policy gradient algorithm is used to learn energy scheduling strategies in a continuous action space: the algorithm constructs an actor-critic network structure, in which the actor network is responsible for outputting specific energy scheduling actions (such as the charge and discharge capacity of energy storage units, the adjustment range of the operating period of adjustable loads, etc.) in an action space composed of continuous variables such as equipment load rate, number of production batches, and number of equipment linkage groups. The critic network evaluates the value of the actions output by the actor network based on the two-dimensional reward function (the deviation rate between real-time energy consumption and theoretical minimum energy consumption, and the equipment response delay time); by continuously interacting with the environment to obtain feedback, the network parameters are continuously updated to optimize action selection, so that the learned energy scheduling strategy can dynamically adapt to peak and valley fluctuations in the power grid and changes in equipment operating status while meeting production plan requirements, ultimately achieving iterative optimization of the energy scheduling rule base and improving the energy efficiency management level of smart factories.

[0046] In one possible implementation, step S100 further includes:

[0047] Step S110: Bind water resource energy consumption and power resource energy consumption with factory production processes and production batch cycles to establish an energy consumption-process mapping relationship table.

[0048] Step S120: Based on the energy consumption-process mapping relationship table, identify the energy consumption differences of the same equipment in different production stages and extract the unit energy consumption benchmark value.

[0049] Step S130: Automatically mark energy efficiency abnormal processes using the unit energy consumption benchmark value and push the mark to the energy management terminal.

[0050] Specifically, the collected water resource energy consumption (such as the water consumption in each production link, the water replenishment and loss of the circulating water system, etc.) and power resource energy consumption (such as the operating power consumption of various equipment, the power loss of transmission lines, etc.) are accurately associated and bound with the specific production processes of the factory (covering the entire process such as raw material pretreatment, processing and assembly, quality inspection, etc.) and the production batch cycle (that is, the complete time interval of each production batch from the beginning of material feeding to the final output of the finished product). By constructing an energy consumption-process mapping relationship table, the specific distribution of different types of energy consumption in each process and each production batch cycle is intuitively presented, thereby achieving accurate correspondence between production energy consumption data and each link and time period in the production process, laying a structured data foundation for subsequent energy consumption analysis and energy efficiency evaluation.

[0051] Based on the established energy consumption-process mapping relationship table, we focus on the energy consumption data of the same equipment in different stages of the production process (such as the startup stage, full-load operation stage, shutdown stage, etc.), and accurately identify the energy consumption difference characteristics of each stage by comparing and analyzing the numerical changes in water resource energy consumption and electricity resource energy consumption in each stage; combined with the rated parameters of the equipment, historical optimal operation records and process standards of the corresponding production batches, we extract the unit energy consumption benchmark value (such as unit product power consumption benchmark, unit time water consumption benchmark, etc.) that can reflect the energy consumption level of the equipment under standard working conditions from these difference characteristics, providing a quantitative reference basis for subsequent judgment of whether the process energy efficiency is abnormal.

[0052] The extracted unit energy consumption benchmark value is used as a reference standard, and it is dynamically compared with the actual energy consumption data of each production process updated in real time in the energy consumption-process mapping relationship table; when the actual water resource energy consumption or electricity resource energy consumption of a process exceeds the preset threshold corresponding to the unit energy consumption benchmark value (such as 120% of the benchmark value), the process is automatically marked as an energy efficiency abnormality process, and an abnormality report containing the name of the abnormal process, the corresponding production batch cycle, the deviation between the actual energy consumption and the benchmark value, the time of the abnormality, etc. is generated simultaneously, and pushed to the energy management terminal in real time through the industrial Internet of Things, so that management personnel can promptly grasp the energy efficiency abnormality and carry out targeted investigation and adjustment.

[0053] In one possible implementation, step S200 further includes:

[0054] Step S210: Embed an equipment energy consumption characteristic curve library in the factory digital twin model, where the equipment energy consumption characteristic curve library contains energy consumption parameters under different load rates.

[0055] Step S220: Based on the equipment energy consumption characteristic curve library, simulate the total energy consumption under different product combinations and different equipment linkage modes to generate the energy consumption Pareto frontier and the production capacity Pareto frontier.

[0056] Step S230: selecting an optimal equilibrium point in a non-dominated solution set from the energy consumption Pareto frontier and the production capacity Pareto frontier according to the production order priority.

[0057] Specifically, an equipment energy consumption characteristic curve library is embedded in the constructed factory digital twin model. This curve library includes the energy consumption parameters corresponding to various types of production equipment in the factory under different load rates (such as low load, medium load, full load and other operating states). Specifically, it covers water resource energy consumption parameters (such as water consumption per unit time, water circulation system energy consumption, etc.) and power resource energy consumption parameters (such as power consumption per unit time, power loss, etc.). By storing the correlation between the equipment load rate and the energy consumption parameters in the form of a curve, the factory digital twin model can accurately simulate the energy consumption performance of the equipment under different operating loads, providing basic data support for subsequent multi-objective optimization and adjustment based on production energy consumption data.

[0058] Based on the equipment energy consumption characteristic curve library embedded in the factory digital twin model (including energy consumption parameters of the equipment under different load rates), the model simulates the total energy consumption under different product combinations (such as the production quantity ratio of multiple products, process route combinations) and different equipment linkage modes (such as the collaborative operation mode of equipment clusters, and the linkage startup sequence); during the simulation process, combined with the correlation between production processes and energy consumption, the total energy consumption and production capacity data corresponding to each combination and mode are calculated, and then the energy consumption Pareto frontier (reflecting the solution set distribution of the optimal energy consumption under a specific production capacity) and the production capacity Pareto frontier (reflecting the solution set distribution of the optimal production capacity under a specific energy consumption) are generated to intuitively present the multi-objective optimization trade-off relationship between energy consumption and production capacity, providing a data basis for the subsequent selection of the optimal balance point.

[0059] Based on production order priorities (such as the delivery deadline for urgent orders and the profit margin weight for high-value orders), the factory digital twin model first calls the data sets of the energy consumption Pareto frontier and the production capacity Pareto frontier to screen out the non-dominated solution set formed by the intersection of the two (that is, each solution in the solution set cannot reduce energy consumption without reducing production capacity, or increase production capacity without increasing energy consumption). Then, a multi-objective decision matrix is ​​constructed based on order priorities. The total energy consumption, production capacity completion rate, and order priority weight corresponding to each solution in the non-dominated solution set are weighted and calculated to obtain a comprehensive score. Finally, the solution with the highest comprehensive score is selected as the optimal balance point. The equipment operating parameters corresponding to this balance point (such as load factor and number of linkage groups) and energy allocation ratio will serve as the specific basis for determining the equipment start-up and shutdown timing and energy allocation strategy, ensuring that the optimal balance between energy consumption and production capacity is achieved while prioritizing the production needs of high-priority orders.

[0060] In one possible implementation, step S230 further includes:

[0061] Step S231: Using the equipment load rate, the number of production batches, and the number of equipment linkage groups as variable parameters, constraints are set in the factory digital twin model.

[0062] Step S232: determining a valid sample set based on the constraint conditions, and performing non-dominated sorting to obtain the non-dominated solution set.

[0063] Specifically, the equipment load rate (i.e., the ratio of the equipment's actual operating load to its rated load), the number of production batches (corresponding to the batch size of different products in the production plan), and the number of equipment linkage groups (referring to the number of equipment combinations that collaborate in the production process) are used as key variable parameters, and corresponding constraints are set in the factory digital twin model. These constraints are determined based on actual production needs and equipment operating specifications. For example, the equipment load rate must be limited to a safe operating range (e.g., no more than 95% of the rated value to prevent equipment overload and damage), the number of production batches must match the upper and lower limits of the total order volume (e.g., no less than the minimum order demand and no more than the maximum production line capacity), and the number of equipment linkage groups is limited by energy supply capacity and production line layout (e.g., the number of linkage groups in the same time period does not exceed the maximum capacity of the energy system). By clearly defining the boundaries of the variable parameters, we ensure that the model simulation process meets the feasibility constraints of actual production scenarios.

[0064] The constraints set in the factory digital twin model (covering the safe range of equipment load rate, the matching limit of production batch quantity and order, the upper limit of the number of equipment linkage groups, etc.) are used to screen out all samples that meet all the constraints from the large amount of sample data generated by model simulation to form a valid sample set. These samples correspond to equipment operation and production arrangement plans that meet the actual production feasibility; then, the samples in the valid sample set are non-dominated sorted, that is, by comparing the performance of each sample in the two target dimensions of energy consumption and production capacity. When a sample has lower energy consumption without reducing production capacity, or higher production capacity without increasing energy consumption, the sample dominates other samples, and vice versa. Finally, all samples that are not dominated by other samples are screened out to form a non-dominated solution set, which provides a basis for the subsequent selection of the optimal balance point from the solution set.

[0065] In one possible implementation, step S300 further includes:

[0066] Step S310: Integrate a wavelet packet decomposition unit on the edge computing node to decompose the vibration spectrum into energy characteristics of multiple frequency bands.

[0067] Step S320: Using the long short-term memory network, the spectrum energy variation trend within the production batch cycle of the production plan information is predicted, and combined with the normal fluctuation range, it is determined whether to trigger the equipment load reduction instruction.

[0068] Specifically, a wavelet packet decomposition unit is integrated into the edge computing nodes deployed at the production site. This unit performs fine processing on the equipment vibration spectrum signals collected in real time by sensors. Utilizing the multi-resolution analysis capabilities of the wavelet packet decomposition algorithm, the original vibration spectrum is decomposed into multiple independent frequency bands according to different frequency intervals, and the energy characteristics corresponding to each frequency band (such as the energy value and energy proportion of each frequency band) are extracted at the same time. Through this decomposition, the vibration energy characteristics of different components (such as bearings, gears, motors, etc.) within a specific frequency range during equipment operation can be separated, providing high-dimensional, fine-grained feature data support for the subsequent identification of abnormal spectrum fluctuation intervals based on load change trends within the production batch cycle. This ensures that the edge computing node can quickly complete the preprocessing and feature extraction of vibration signals locally, improving the real-time and accuracy of equipment status monitoring.

[0069] Based on the multiple frequency band energy features obtained by the edge computing node through the wavelet packet decomposition unit, the long short-term memory network is used to predict the change trend of the equipment vibration spectrum energy within the production batch cycle in the production plan information. The network can effectively capture the dynamic dependence of vibration energy on the time series and accurately output the spectrum energy change curve of each time period within the production batch cycle; at the same time, combined with the normal fluctuation range of the spectrum energy under normal working conditions statistically calculated in the historical operation data of the equipment (such as the mean and standard deviation range of the energy of each frequency band), the predicted spectrum energy change trend is compared with the normal fluctuation range in real time. If the predicted trend exceeds the normal fluctuation range and the duration reaches the preset threshold, or the sudden increase / decrease in the energy value of a certain frequency band exceeds the set standard, it is judged that the equipment has a potential failure risk, and the equipment load reduction instruction is immediately triggered. By reducing the equipment load, the loss caused by abnormal vibration is reduced, thereby ensuring the stability of equipment operation within the production batch cycle.

[0070] Example 2 is based on the same inventive concept as the method for optimizing energy efficiency of a smart factory based on industrial Internet of Things in the above-mentioned embodiment. Figure 2 As shown, this application provides a smart factory energy efficiency optimization system based on the Industrial Internet of Things. The system and method embodiments in the embodiments of this application are based on the same inventive concept. The system includes:

[0071] The production energy consumption data acquisition module 10 is used to collect production energy consumption data including water resource energy consumption and power resource energy consumption within the same timestamp based on equipment operating parameters and production plan information.

[0072] The energy allocation strategy determination module 20 is used to build a factory digital twin model based on the equipment operating parameters and energy scheduling rule library, and perform multi-objective optimization and adjustment in combination with the production energy consumption data to determine the equipment start-stop timing and energy allocation strategy.

[0073] The adaptive compensation optimization module 30 is used to deploy edge computing nodes, monitor the vibration spectrum of the equipment, match the abnormal spectrum fluctuation interval through cosine similarity through the load change trend within the production batch cycle of the production plan information, set the early fault sign characteristics under the attention mechanism, generate the state prediction diagram within the time sliding window, and perform adaptive compensation optimization.

[0074] The energy dynamic adjustment module 40 is used to formulate an energy flow network based on the equipment start-stop sequence and energy distribution strategy to perform dynamic energy adjustment.

[0075] Furthermore, the system is also used to implement the following functions:

[0076] Based on the equipment start-stop timing and energy distribution strategy, an energy flow network is formulated and key energy efficiency indicators are extracted. The key energy efficiency indicators include the transformer load rate and the key energy efficiency indicators of the pipeline pressure loss. According to the key energy efficiency indicators, combined with the peak and valley signals of the power grid and the production plan information, the dynamic adjustment parameters and priority scheduling rules of the energy storage unit are configured.

[0077] Furthermore, the system is also used to implement the following functions:

[0078] The energy storage unit is connected to the industrial Internet of Things; the charging and discharging strategy of the energy storage unit and the operating period of the adjustable load are dynamically adjusted; and the energy scheduling rule base is iteratively updated using a reinforcement learning algorithm.

[0079] Furthermore, the system is also used to implement the following functions:

[0080] A two-dimensional reward function is set up, where the first dimension is the deviation rate between real-time energy consumption and theoretical minimum energy consumption, and the second dimension is the device response delay time. Based on the two-dimensional reward function, a deep deterministic policy gradient is used to learn the energy scheduling strategy in the continuous action space.

[0081] Furthermore, the system is also used to implement the following functions:

[0082] Bind water resource energy consumption and electricity resource energy consumption with the factory production process and production batch cycle to establish an energy consumption-process mapping relationship table; based on the energy consumption-process mapping relationship table, identify the energy consumption differences of the same equipment in different production stages and extract the unit energy consumption benchmark value; use the unit energy consumption benchmark value to automatically mark energy efficiency abnormal processes and push them to the energy management terminal.

[0083] Furthermore, the system is also used to implement the following functions:

[0084] An equipment energy consumption characteristic curve library is embedded in the factory digital twin model, and the equipment energy consumption characteristic curve library contains energy consumption parameters under different load rates. Based on the equipment energy consumption characteristic curve library, the total energy consumption under different product combinations and different equipment linkage modes is simulated to generate the energy consumption Pareto frontier and the production capacity Pareto frontier. According to the production order priority, the optimal balance point in the non-dominated solution set is selected from the energy consumption Pareto frontier and the production capacity Pareto frontier.

[0085] Furthermore, the system is also used to implement the following functions:

[0086] The equipment load rate, the number of production batches, and the number of equipment linkage groups are used as variable parameters, and constraints are set in the factory digital twin model; a valid sample set is determined based on the constraints, and non-dominated sorting is performed to obtain the non-dominated solution set.

[0087] Furthermore, the system is also used to implement the following functions:

[0088] A wavelet packet decomposition unit is integrated on the edge computing node to decompose the vibration spectrum into energy characteristics of multiple frequency bands. Through the long short-term memory network, the spectrum energy change trend within the production batch cycle of the production plan information is predicted, and combined with the normal fluctuation range, it is determined whether to trigger the equipment load reduction instruction.

[0089] Example 3, Figure 3 A structural schematic diagram of an electronic device provided for the method for optimizing energy efficiency of a smart factory based on the Industrial Internet of Things of the present invention shows a block diagram of an exemplary electronic device suitable for implementing an embodiment of the present invention. Figure 3 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention. Figure 3 As shown, the electronic device includes a processor 21, a memory 22, an input device 23 and an output device 24; the number of processors 21 in the electronic device can be one or more. Figure 3 Taking a processor 21 as an example, the processor 21, memory 22, input device 23 and output device 24 in the electronic device can be connected through a bus or other means. Figure 3 The bus connection is taken as an example.

[0090] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0091] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0092] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A smart factory energy efficiency optimization method based on industrial Internet of Things, characterized by: The method comprises: Based on equipment operating parameters and production plan information, collect production energy consumption data including water resource energy consumption and power resource energy consumption within the same timestamp limit; Based on the equipment operating parameters and energy scheduling rule base, a factory digital twin model is constructed, and multi-objective optimization and adjustment are performed in combination with the production energy consumption data to determine the equipment start-stop timing and energy allocation strategy; Deploy edge computing nodes to monitor equipment vibration spectra. Based on the load change trends within the production batch cycle of production plan information, match abnormal spectrum fluctuation intervals using cosine similarity, set early fault sign features using an attention mechanism, generate a state prediction graph within a time sliding window, and perform adaptive compensation optimization. At the same time, an energy flow network is formulated based on the equipment start-stop sequence and energy distribution strategy to perform dynamic energy adjustment; The collection of production energy consumption data including water resource energy consumption and power resource energy consumption within the same time stamp includes: Bind water resource energy consumption and electricity resource energy consumption with factory production processes and production batch cycles to establish an energy consumption-process mapping relationship table; Based on the energy consumption-process mapping relationship table, identifying the energy consumption differences of the same equipment in different production stages and extracting the unit energy consumption benchmark value; Use the unit energy consumption benchmark value to automatically mark energy efficiency abnormal processes and push them to the energy management terminal; Among them, a factory digital twin model is constructed, and multi-objective optimization and adjustment are carried out in combination with the production energy consumption data to determine the equipment start-stop sequence and energy allocation strategy, including: Embedding an equipment energy consumption characteristic curve library in the factory digital twin model, wherein the equipment energy consumption characteristic curve library includes energy consumption parameters under different load rates; Based on the equipment energy consumption characteristic curve library, simulate the total energy consumption under different product combinations and different equipment linkage modes to generate the energy consumption Pareto frontier and the production capacity Pareto frontier; According to the production order priority, the optimal equilibrium point in the non-dominated solution set is selected from the energy consumption Pareto frontier and the production capacity Pareto frontier; Using equipment load rate, production batch number, and equipment linkage group number as variable parameters, setting constraints in the factory digital twin model; Determine a valid sample set based on the constraint conditions, and perform non-dominated sorting to obtain the non-dominated solution set; Among them, edge computing nodes are deployed to monitor the vibration spectrum of equipment, including: Integrating a wavelet packet decomposition unit on the edge computing node to decompose the vibration spectrum into energy characteristics of multiple frequency bands; Through the long short-term memory network, the spectrum energy change trend within the production batch cycle of the production plan information is predicted, and combined with the normal fluctuation range, it is determined whether to trigger the equipment load reduction instruction.

2. The method for optimizing energy efficiency of a smart factory based on the Industrial Internet of Things according to claim 1, wherein: An energy flow network is formulated based on the equipment start-stop sequence and energy distribution strategy to perform dynamic energy adjustment. The method includes: Formulate an energy flow network based on the equipment start-stop sequence and energy allocation strategy, and extract key energy efficiency indicators, including transformer load rate and pipe network pressure loss; According to the key energy efficiency indicators, combined with the peak and valley signals of the power grid and the production plan information, the dynamic adjustment parameters and priority scheduling rules of the energy storage unit are configured.

3. The method for optimizing energy efficiency of a smart factory based on the industrial Internet of Things according to claim 2, wherein: The energy storage unit is connected to the industrial Internet of Things; Dynamically adjust the charging and discharging strategy of the energy storage unit and the operating period of the adjustable load; The energy scheduling rule base is iteratively updated using a reinforcement learning algorithm.

4. The method for optimizing energy efficiency of a smart factory based on the Industrial Internet of Things according to claim 3, wherein: Using a reinforcement learning algorithm to iteratively update the energy scheduling rule base, the method includes: Set up a two-dimensional reward function, where the first dimension is the deviation rate between real-time energy consumption and theoretical minimum energy consumption, and the second dimension is the device response delay time; Based on the two-dimensional reward function, deep deterministic policy gradient is adopted to learn energy scheduling strategy in continuous action space.

5. The smart factory energy efficiency optimization system based on industrial Internet of Things is characterized by: The system is used to implement the smart factory energy efficiency optimization method based on the industrial Internet of Things according to any one of claims 1 to 4, and the system includes: The production energy consumption data acquisition module is used to collect production energy consumption data including water resource energy consumption and power resource energy consumption within the same timestamp based on equipment operating parameters and production plan information; An energy allocation strategy determination module is used to build a factory digital twin model based on the equipment operating parameters and energy scheduling rule library, and to perform multi-objective optimization and adjustment in combination with the production energy consumption data to determine the equipment start-stop timing and energy allocation strategy; The adaptive compensation optimization module is used to deploy edge computing nodes, monitor equipment vibration spectra, and use cosine similarity to match abnormal spectrum fluctuation intervals based on load change trends within the production batch cycle of production plan information. It then sets early fault sign features using an attention mechanism, generates a state prediction graph within a time sliding window, and performs adaptive compensation optimization. The energy dynamic adjustment module is used to formulate an energy flow network based on the equipment start-stop timing and energy distribution strategy to perform dynamic energy adjustment.

6. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; Wherein, the processor is used to execute the smart factory energy efficiency optimization method based on industrial Internet of Things as described in any one of claims 1 to 4.

Images