Energy consumption optimization method and system for kelp drying based on big data comprehensive analysis

Abstract

The application discloses a kelp drying energy consumption optimization method and system combined with big data comprehensive analysis, relates to the field of kelp drying energy consumption, and comprises the following steps: collecting equipment operation state data and environmental parameters in real time by setting edge data collection nodes; using the Internet of Things technology, receiving the data collected by each edge data collection node in real time, and establishing a kelp drying energy consumption database; constructing a kelp drying energy consumption evaluation model, comprehensively evaluating and predicting kelp drying energy consumption; according to the evaluation result of the kelp drying energy consumption evaluation model, obtaining the weight value of the drying energy consumption of each edge data collection node; obtaining the correlation of the drying energy consumption of each edge data collection node, and dynamically adjusting and optimizing the kelp drying energy consumption. The application has the advantages that through multi-modal perception data and a hybrid machine learning model, the kelp drying energy consumption is dynamically evaluated and the waste heat recovery is optimized, and the kelp production capacity loss is significantly reduced.

Description

Technical Field

[0001] This invention relates to the field of energy consumption in kelp drying, specifically to a method and system for optimizing energy consumption in kelp drying by combining big data comprehensive analysis. Background Technology

[0002] Currently, with the continuous growth of global energy demand and the increasing prominence of environmental issues, the energy consumption problem in the kelp drying industry has received increasing attention. On the one hand, high energy costs increase the drying costs of enterprises and reduce their market competitiveness. On the other hand, the large amount of energy consumption also puts considerable pressure on the environment. However, in the existing kelp drying process, energy consumption management is mostly in the post-event statistical analysis stage, lacking real-time prediction and optimization control of energy consumption during the drying process. It is impossible to adjust drying parameters and equipment operating status in a timely manner during the drying process to reduce energy consumption. Therefore, developing a method and system that can predict and optimize kelp drying energy consumption in real time is of great practical significance.

[0003] Existing kelp drying energy management technologies mainly focus on model building for single stages, lacking a global perception data dimension for kelp drying. They cannot comprehensively handle temporal dependencies, spatial correlations, and dynamic attention mechanisms, and lack dynamic optimization capabilities based on real-time weights and energy consumption correlations. This results in rigid waste heat recovery strategies, sluggish system response, and difficulty in adapting to dynamic changes in the drying environment and processes, ultimately leading to problems such as low energy consumption prediction accuracy, poor optimization effects, and energy waste. Summary of the Invention

[0004] To address the aforementioned technical problems, this paper provides a method and system for optimizing kelp drying energy consumption by combining big data comprehensive analysis. This technical solution solves the problems mentioned in the background technology, which lacks a global perception data dimension for kelp drying, cannot comprehensively handle temporal dependencies, spatial correlations and dynamic attention mechanisms, and lacks dynamic optimization capabilities based on real-time weights and energy consumption correlations. This results in rigid waste heat recovery strategies, sluggish system response, and difficulty in adapting to dynamic changes in the drying environment and process, ultimately leading to problems such as low energy consumption prediction accuracy, poor optimization effect and energy waste.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for optimizing kelp drying energy consumption by combining big data comprehensive analysis includes:

[0007] Multimodal sensor arrays are deployed in various kelp drying processes, and edge data acquisition nodes are set up to collect equipment operating status data and environmental parameters in real time.

[0008] Using IoT technology, data collected by each edge data acquisition node is received in real time, and a kelp drying energy consumption database is established.

[0009] Based on the LSTM-Transformer-GNN hybrid machine learning algorithm, a kelp drying energy consumption assessment model is constructed to comprehensively assess and predict kelp drying energy consumption.

[0010] Based on the evaluation results of the kelp drying energy consumption assessment model, the weight values ​​of the drying energy consumption of each edge data acquisition node are obtained.

[0011] The correlation between drying energy consumption of each edge data acquisition node is obtained, and the energy consumption of kelp drying is dynamically adjusted and optimized through waste heat recovery.

[0012] Preferably, the step of constructing a kelp drying energy consumption assessment model based on the LSTM-Transformer-GNN hybrid machine learning algorithm to comprehensively assess and predict kelp drying energy consumption specifically includes:

[0013] Based on the kelp drying process cycle, the window length of the LSTM layer timing is set, wherein the window length of the LSTM layer timing is within the range of the kelp drying process cycle.

[0014] Based on the kelp drying energy consumption database, the data collected by each edge data acquisition node is used as the input.

[0015] Based on the LSTM neural network, the output includes time-series device operating status data and environmental parameters, which are then used as input to the Transformer model.

[0016] By using linear projection, the data collected by each edge data acquisition node in real time is mapped to a space with the same dimension as the Key / Value;

[0017] Based on the Transformer model, the correlation between each data point in the input and other data points is calculated, and the dependency between any data points in the input data is captured.

[0018] Based on the knowledge graph of kelp processing parameters, and based on the mean square error formula, a constraint loss function for the LSTM-Transformer model is established.

[0019] Based on the LSTM-Transformer model, determine and predict the kelp drying energy consumption value of each edge data acquisition node and the total kelp drying energy consumption value within the window length;

[0020] Based on the drying process of each edge data acquisition node in the kelp drying process, the edge weights are adjusted in real time using a CNN neural network to accurately reflect the impact of changes in the drying schedule on energy consumption, and the energy consumption value of kelp drying is further corrected and optimized.

[0021] The expression for setting the window length of the LSTM layer timing is:

[0022]

[0023] In the formula, The window length for the LSTM layer timing. The drying process cycle for kelp, The dominant angular frequency in the kelp drying data. To ensure the window length is an integer time unit, a floor sign is used for rounding down.

[0024] The constrained loss function expression for the LSTM-Transformer model is as follows:

[0025]

[0026] In the formula, The value of the constrained loss function for the LSTM-Transformer model. The number of sample groups. This is the predicted energy consumption value for kelp drying from the model. The target value for energy consumption in kelp drying. These are the weight coefficients with constraints in the kelp process parameter knowledge graph. This is a sample set with constraints in the knowledge graph of kelp processing parameters. The standard boundary thresholds are defined in the kelp processing parameter knowledge graph.

[0027] Preferably, obtaining the weight value of the drying energy consumption of each edge data acquisition node based on the evaluation results of the kelp drying energy consumption evaluation model specifically includes:

[0028] Based on the evaluation results of the kelp drying energy consumption assessment model, the kelp drying energy consumption values ​​of each edge data acquisition node within the window length are obtained.

[0029] Based on the kelp drying energy consumption assessment model, the weights in the LSTM-Transformer-GNN hybrid model were extracted respectively.

[0030] Wherein, LSTM is the historical data weight of each edge data acquisition node, Transformer is the self-attention weight of each edge data acquisition node, and GNN is the spatial weight of each edge data acquisition node.

[0031] Based on the weighting formula, obtain the comprehensive weight value of the drying energy consumption of each edge data acquisition node;

[0032] According to the Softmax normalization formula, the comprehensive weight value of the drying energy consumption of each edge data acquisition node is normalized so that the sum of all comprehensive weight values ​​is 1.

[0033] The comprehensive weighting expression for the drying energy consumption of each edge data acquisition node is as follows:

[0034]

[0035] In the formula, The comprehensive weighted value for drying energy consumption at edge data acquisition nodes. , , These are the historical data weights of the edge data acquisition nodes, the self-attention weights of the edge data acquisition nodes, and the spatial weights of the edge data acquisition nodes, respectively. , , These are the weighting coefficients for historical data, self-attention, and spatial weights of the edge data acquisition nodes, respectively, which can be determined through cross-validation.

[0036] Preferably, the step of obtaining the correlation of drying energy consumption of each edge data acquisition node and dynamically adjusting and optimizing kelp drying energy consumption through waste heat recovery specifically includes:

[0037] Based on the kelp drying energy consumption database, obtain the operating status data and environmental parameters of each edge data acquisition node device;

[0038] Based on the operating status data and environmental parameters of each edge data acquisition node and the drying energy consumption data of each edge data acquisition node, the correlation of drying energy consumption of each edge data acquisition node is determined based on the Pearson correlation formula.

[0039] Based on the correlation of drying energy consumption of each edge data acquisition node and the correlation and coupling of each node, a dynamic adjustment and optimization model for kelp drying energy consumption is established.

[0040] Based on the dynamic adjustment and optimization model of kelp drying energy consumption, the start-up and shutdown and parameters of waste heat recovery are dynamically adjusted.

[0041] Furthermore, this solution proposes a kelp drying energy consumption optimization system that combines big data comprehensive analysis to implement the kelp drying energy consumption optimization method described above, including:

[0042] The data acquisition module is used to deploy a multimodal sensor array in each kelp drying process. By setting up edge data acquisition nodes, it can collect equipment operating status data and environmental parameters in real time. Using Internet of Things (IoT) technology, it can receive the data collected by each edge data acquisition node in real time and establish a kelp drying energy consumption database.

[0043] The prediction and optimization module is used to construct a kelp drying energy consumption assessment model based on the LSTM-Transformer-GNN hybrid machine learning algorithm, comprehensively assess and predict kelp drying energy consumption; obtain the weight value of drying energy consumption of each edge data acquisition node according to the assessment results of the kelp drying energy consumption assessment model; obtain the correlation of drying energy consumption of each edge data acquisition node, and dynamically adjust and optimize kelp drying energy consumption through waste heat recovery optimization.

[0044] Preferably, the data acquisition module includes:

[0045] The data acquisition unit is used to deploy a multimodal sensor array in each drying process of kelp, and to collect equipment operating status data and environmental parameters in real time by setting edge data acquisition nodes;

[0046] The data processing unit is used to receive data collected by each edge data acquisition node in real time using Internet of Things (IoT) technology, and to establish a kelp drying energy consumption database.

[0047] Preferably, the prediction and optimization module includes:

[0048] An evaluation and prediction unit is used to construct a kelp drying energy consumption evaluation model based on the LSTM-Transformer-GNN hybrid machine learning algorithm, and to comprehensively evaluate and predict the kelp drying energy consumption.

[0049] The node weighting unit is used to obtain the weight value of the drying energy consumption of each edge data acquisition node based on the evaluation results of the kelp drying energy consumption evaluation model.

[0050] An energy consumption adjustment unit is used to obtain the correlation of drying energy consumption of each edge data acquisition node, and dynamically adjust and optimize the kelp drying energy consumption through waste heat recovery optimization.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0052] This invention provides a method for optimizing kelp drying energy consumption by combining big data comprehensive analysis. It deploys a multi-modal sensor array to synchronize equipment operating status data and multi-dimensional environmental parameter data, and combines this with edge data acquisition nodes to achieve real-time data collection, constructing a kelp drying energy consumption database covering all process stages. This provides a richer data foundation for subsequent energy consumption analysis. Simultaneously, based on the LSTM-Transformer-GNN hybrid machine learning algorithm, a kelp drying energy consumption assessment model is constructed to comprehensively assess and predict kelp drying energy consumption, achieving in-depth mining and accurate prediction of energy consumption data, significantly improving the reliability of the assessment results. Furthermore, based on the kelp drying energy consumption assessment model, LSTM- The weights in the Transformer-GNN hybrid model are used to obtain the comprehensive weight value of drying energy consumption of each edge data acquisition node according to the weighting formula. This innovatively integrates the temporal modeling capability of LSTM, the self-attention mechanism of Transformer, and the spatial correlation analysis capability of GNN. Finally, by combining the correlation between the comprehensive weight value and energy consumption with the thermo-mass coupling characteristics of the waste heat recovery system, the real-time dynamic adjustment of recovery start-up and shutdown and parameters is realized. This enables the system to adapt to the fluctuations in drying schedule, maximize waste heat utilization and reduce total energy consumption. Thus, by using multimodal sensing data and hybrid machine learning models, the energy consumption of kelp drying can be dynamically evaluated and waste heat recovery optimized, significantly reducing kelp production capacity loss. Attached Figure Description

[0053] Figure 1 This is a flowchart of the kelp drying energy consumption optimization method combining big data comprehensive analysis of the present invention;

[0054] Figure 2 The present invention uses a hybrid machine learning algorithm based on LSTM-Transformer-GNN to construct a kelp drying energy consumption assessment model, and provides a flowchart for comprehensively assessing and predicting kelp drying energy consumption.

[0055] Figure 3 The flowchart below shows how the weight values ​​of the drying energy consumption of each edge data acquisition node are obtained based on the evaluation results of the kelp drying energy consumption evaluation model of the present invention.

[0056] Figure 4 To obtain the correlation of drying energy consumption of each edge data acquisition node in this invention, the kelp drying energy consumption flowchart is dynamically adjusted and optimized through waste heat recovery optimization. Detailed Implementation

[0057] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0058] Reference Figure 1As shown, a method for optimizing kelp drying energy consumption by combining big data comprehensive analysis includes:

[0059] Multimodal sensor arrays are deployed in various kelp drying processes, and edge data acquisition nodes are set up to collect equipment operating status data and environmental parameters in real time.

[0060] Using IoT technology, data collected by each edge data acquisition node is received in real time, and a kelp drying energy consumption database is established.

[0061] Based on the LSTM-Transformer-GNN hybrid machine learning algorithm, a kelp drying energy consumption assessment model is constructed to comprehensively assess and predict kelp drying energy consumption.

[0062] Based on the evaluation results of the kelp drying energy consumption assessment model, the weight values ​​of the drying energy consumption of each edge data acquisition node are obtained.

[0063] The correlation between drying energy consumption of each edge data acquisition node is obtained, and the energy consumption of kelp drying is dynamically adjusted and optimized through waste heat recovery.

[0064] This solution utilizes a multi-modal sensor array to synchronize equipment operating status data and environmental parameters across multiple dimensions. Combined with edge data acquisition nodes, it achieves real-time data collection, constructing a comprehensive kelp drying energy consumption database covering all process stages. This provides a richer data foundation for subsequent energy consumption analysis. Furthermore, based on the LSTM-Transformer-GNN hybrid machine learning algorithm, a kelp drying energy consumption assessment model is built to comprehensively evaluate and predict kelp drying energy consumption. This achieves in-depth data mining and accurate prediction of energy consumption, significantly improving the reliability of the assessment results. Secondly, based on the kelp drying energy consumption assessment model, LSTM-Transformer data is extracted... The weights in the er-GNN hybrid model are determined, and the comprehensive weight value of drying energy consumption of each edge data acquisition node is obtained according to the weighting formula. The model innovatively integrates the temporal modeling capability of LSTM, the self-attention mechanism of Transformer, and the spatial correlation analysis capability of GNN. Finally, by combining the correlation between the comprehensive weight value and energy consumption with the thermo-mass coupling characteristics of the waste heat recovery system, the real-time dynamic adjustment of recovery start-up and shutdown and parameters is realized. This enables the system to adapt to the fluctuations in drying schedule, maximize the utilization rate of waste heat and reduce the total energy consumption. Thus, the goal of dynamically evaluating the energy consumption of kelp drying and optimizing waste heat recovery through multimodal sensing data and hybrid machine learning model is achieved, significantly reducing the loss of kelp production capacity.

[0065] The deployment of a multimodal sensor array in each kelp drying process, and the real-time acquisition of equipment operating status data and environmental parameters by setting edge data acquisition nodes, specifically includes:

[0066] Based on the kelp drying process, a multimodal sensor array is deployed in each drying process to collect real-time equipment operating status data and environmental parameters during each drying process.

[0067] In each drying process, edge data acquisition nodes are set up to filter and normalize the real-time equipment operating status data and environmental parameters collected by the edge data acquisition nodes.

[0068] The simultaneous acquisition mode is adopted to enable all edge data acquisition nodes to operate synchronously under a unified time base, ensuring the time consistency and synchronization of the acquired data.

[0069] Standardized communication format specifications are uniformly set for each edge data acquisition node to ensure that the data uploaded by different nodes maintains a high degree of consistency in format and structure.

[0070] It can be explained that deploying multimodal sensor arrays and edge data acquisition nodes in kelp drying, and constructing a unified data acquisition and communication mechanism, is the foundation for building an efficient energy consumption management system. Its core significance lies in solving the analysis failure problems caused by data fragmentation, noise interference, and heterogeneity in traditional solutions through full-dimensional data perception and standardized processing. At the same time, the simultaneous acquisition mode ensures that each node operates synchronously under a unified time base, avoiding data misalignment caused by clock drift or transmission delay, and ensuring time consistency. This provides an accurate, real-time, and standardized data foundation for kelp drying energy consumption management, and is a key support for improving the accuracy of energy consumption prediction and optimization effect.

[0071] The process of utilizing IoT technology to receive data collected in real time from various edge data acquisition nodes and establish a kelp drying energy consumption database specifically includes:

[0072] Utilizing IoT technology, data collected by each edge data acquisition node is received in real time;

[0073] By using interpolation and data downsampling techniques, the data from each edge data acquisition node is processed to achieve data alignment and positional consistency in time or space;

[0074] Based on the process requirements of kelp drying, the boundary constraints of the data collected by each edge data acquisition node are extracted to establish a knowledge graph of kelp process parameters.

[0075] A database of energy consumption for kelp drying was established based on historical data on kelp drying and a knowledge graph of kelp processing parameters.

[0076] This can be explained by the fact that in kelp drying energy consumption management, the real-time reception of multi-source kelp drying data from edge data acquisition nodes via IoT technology, and the creation of a kelp drying energy consumption database, is the core foundation supporting accurate energy consumption assessment and optimization decisions for kelp drying. Interpolation and data downsampling techniques can effectively address the issue of data location discrepancies. By using the timestamps and spatial locations of the multi-source kelp drying data, consistency across time and space is ensured, providing reliable input for subsequent analysis. Based on this, data boundary constraints are extracted in conjunction with kelp processing requirements, and a kelp process parameter knowledge graph is constructed. This explicitly expresses the boundary constraints between parameters, avoiding energy consumption anomalies caused by parameter out-of-bounds behavior, thereby improving the accuracy and reliability of the data.

[0077] Reference Figure 2 As shown, the kelp drying energy consumption assessment model constructed based on the LSTM-Transformer-GNN hybrid machine learning algorithm, which comprehensively assesses and predicts kelp drying energy consumption, specifically includes:

[0078] Based on the kelp drying process cycle, the window length of the LSTM layer timing is set, wherein the window length of the LSTM layer timing is within the range of the kelp drying process cycle.

[0079] Based on the kelp drying energy consumption database, the data collected by each edge data acquisition node is used as the input.

[0080] Based on the LSTM neural network, the output includes time-series device operating status data and environmental parameters, which are then used as input to the Transformer model.

[0081] By using linear projection, the data collected by each edge data acquisition node in real time is mapped to a space with the same dimension as the Key / Value;

[0082] Based on the Transformer model, the correlation between each data point in the input and other data points is calculated, and the dependency between any data points in the input data is captured.

[0083] Based on the knowledge graph of kelp processing parameters, and based on the mean square error formula, a constraint loss function for the LSTM-Transformer model is established.

[0084] Based on the LSTM-Transformer model, determine and predict the kelp drying energy consumption value of each edge data acquisition node and the total kelp drying energy consumption value within the window length;

[0085] Based on the drying process of each edge data acquisition node in the kelp drying process, the edge weights are adjusted in real time using a CNN neural network to accurately reflect the impact of changes in the drying schedule on energy consumption, and the energy consumption value of kelp drying is further corrected and optimized.

[0086] The expression for setting the window length of the LSTM layer timing is:

[0087]

[0088] In the formula, The window length for the LSTM layer timing. The drying process cycle for kelp, The dominant angular frequency in the kelp drying data. To ensure the window length is an integer time unit, a floor sign is used for rounding down.

[0089] The constrained loss function expression for the LSTM-Transformer model is as follows:

[0090]

[0091] In the formula, The value of the constrained loss function for the LSTM-Transformer model. The number of sample groups. This is the predicted energy consumption value for kelp drying from the model. The target value for energy consumption in kelp drying. These are the weight coefficients with constraints in the kelp process parameter knowledge graph. This is a sample set with constraints in the knowledge graph of kelp processing parameters. The standard boundary thresholds are defined in the kelp processing parameter knowledge graph.

[0092] This can be explained by the fact that traditional LSTM or neural network models struggle to account for both temporal fluctuations and spatial layout influences, leading to significant discrepancies between predictions and actual results, insufficient prediction accuracy, and a disconnect between temporal dependencies and spatial correlations. Therefore, this solution extracts temporal features of equipment operating status and environmental parameters from LSTM layers, using these as input to a Transformer model. A self-attention mechanism is employed to capture arbitrary dependencies between data. Simultaneously, a constrained loss function is constructed using a knowledge graph of kelp process parameters, ensuring that the model output conforms to process constraints and significantly improving prediction accuracy. Furthermore, a CNN neural network dynamically adjusts the edge weights, reflecting the drying output in real time. The impact of process changes on energy consumption is investigated, and the prediction results are further corrected to avoid the accumulation of errors caused by static model assumptions. This enables the model to not only output the local energy consumption values ​​of each edge node, but also to comprehensively evaluate the total energy consumption of the entire process, providing data support for subsequent dynamic waste heat recovery optimization and significantly improving the energy efficiency management level of kelp drying. Among them, the LSTM-Transformer hybrid model is used to process sensor time series data and predict the process-level energy consumption of the future window length, realizing short-term prediction of kelp drying energy consumption. The GNN neural network is used to model the energy consumption dependency between processes, and combined with the drying schedule, the energy consumption trend of the whole batch is predicted, realizing long-term prediction of kelp drying energy consumption.

[0093] The expression for the total energy consumption value of kelp drying is:

[0094]

[0095] In the formula, This represents the total energy consumption for drying kelp. The number of edge data acquisition nodes. For the first Energy consumption value for kelp drying at each edge data acquisition node.

[0096] Reference Figure 3 As shown, obtaining the weight values ​​of the drying energy consumption of each edge data acquisition node based on the evaluation results of the kelp drying energy consumption assessment model specifically includes:

[0097] Based on the evaluation results of the kelp drying energy consumption assessment model, the kelp drying energy consumption values ​​of each edge data acquisition node within the window length are obtained.

[0098] Based on the kelp drying energy consumption assessment model, the weights in the LSTM-Transformer-GNN hybrid model were extracted respectively.

[0099] Wherein, LSTM is the historical data weight of each edge data acquisition node, Transformer is the self-attention weight of each edge data acquisition node, and GNN is the spatial weight of each edge data acquisition node.

[0100] Based on the weighting formula, obtain the comprehensive weight value of the drying energy consumption of each edge data acquisition node;

[0101] According to the Softmax normalization formula, the comprehensive weight value of the drying energy consumption of each edge data acquisition node is normalized so that the sum of all comprehensive weight values ​​is 1.

[0102] The comprehensive weighting expression for the drying energy consumption of each edge data acquisition node is as follows:

[0103]

[0104] In the formula, The comprehensive weighted value for drying energy consumption at edge data acquisition nodes. , , These are the historical data weights of the edge data acquisition nodes, the self-attention weights of the edge data acquisition nodes, and the spatial weights of the edge data acquisition nodes, respectively. , , These are the weighting coefficients for historical data, self-attention, and spatial weights of the edge data acquisition nodes, respectively, which can be determined through cross-validation.

[0105] This can be explained by the fact that kelp drying involves the collaboration of multiple devices and the coupling of complex processes. The energy consumption contribution of different edge data acquisition nodes is significantly affected by historical operating patterns, real-time correlations, and spatial layout. Due to the multidimensionality and complexity of the data collected by each edge data acquisition node, the difficulty of accurately locating the key energy consumption of kelp drying increases. To solve the problems of fuzzy node contribution and scattered optimization targets in traditional energy consumption analysis, this solution uses LSTM to extract the historical data weights of each edge data acquisition node, Transformer to extract the self-attention weights of each edge data acquisition node, and GNN to extract the spatial weights of each edge data acquisition node. By using the three weighting methods of historical data weights, self-attention weights, and spatial weights, and integrating temporal dependence, data correlation, and spatial correlation, a comprehensive weight value is obtained. This allows for the quantification of node energy consumption priority, such as prioritizing high-weight nodes in waste heat recovery, thus achieving accurate quantification of the energy consumption contribution of kelp drying. This provides key support for building a closed loop for energy consumption optimization and improving the precision of energy efficiency management, and provides data correlation analysis for subsequent waste heat recovery optimization.

[0106] Reference Figure 4 As shown, the process of obtaining the correlation of drying energy consumption at each edge data acquisition node and dynamically adjusting and optimizing kelp drying energy consumption through waste heat recovery specifically includes:

[0107] Based on the kelp drying energy consumption database, obtain the operating status data and environmental parameters of each edge data acquisition node device;

[0108] Based on the operating status data and environmental parameters of each edge data acquisition node and the drying energy consumption data of each edge data acquisition node, the correlation of drying energy consumption of each edge data acquisition node is determined based on the Pearson correlation formula.

[0109] Based on the correlation of drying energy consumption of each edge data acquisition node and the correlation and coupling of each node, a dynamic adjustment and optimization model for kelp drying energy consumption is established.

[0110] Based on the dynamic adjustment and optimization model of kelp drying energy consumption, the start and stop of waste heat recovery and kelp drying energy consumption are dynamically adjusted.

[0111] This can be explained by the fact that traditional static waste heat recovery schemes often suffer from insufficient recovery or wasted heat sources because they cannot perceive the energy consumption coupling relationship between nodes. However, in the energy consumption management of kelp drying, quantifying the energy consumption correlation of each edge data acquisition node and dynamically adjusting the waste heat recovery strategy is a key breakthrough in overcoming the isolation of traditional energy efficiency optimization methods and achieving collaborative energy saving throughout the entire process. This scheme establishes a dynamic adjustment and optimization model for kelp drying energy consumption based on the correlation between data and energy consumption, and dynamically adjusts the start and stop of waste heat recovery and kelp drying energy consumption through the model. The formula for the dynamic start and stop of waste heat recovery is as follows:

[0112] The formula for the activation condition is:

[0113] The stopping condition formula is:

[0114] In the formula, The comprehensive weighted value for drying energy consumption at edge data acquisition nodes. To determine the correlation of drying energy consumption among edge data acquisition nodes and the correlation and coupling between nodes, and These are the comprehensive weight value of drying energy consumption of edge data acquisition nodes, the correlation of drying energy consumption of each edge data acquisition node, and the boundary threshold of the correlation and coupling between nodes. To predict the waste heat recovery efficiency of kelp drying energy consumption. The minimum value for predicted waste heat recovery efficiency in kelp drying is determined. Real-time recovery efficiency of waste heat from kelp drying;

[0115] Based on the correlation between data and the correlation between energy consumption, the maximum recovery efficiency of waste heat from kelp drying is obtained. The expression for the maximum recovery efficiency of waste heat from kelp drying is as follows:

[0116]

[0117] In the formula, For from the first The edge data acquisition node to the first Waste heat recovery flow rate of each edge data acquisition node This represents the maximum waste heat recovery flow rate of the edge data acquisition node. , These represent two nodes in the waste heat supply and demand relationship during the kelp drying process, while waste heat recovery refers to the set of all waste heat recovery paths during the kelp drying process. For the first The comprehensive weighted value of drying energy consumption of each edge data acquisition node The first point regarding the supply and demand relationship of waste heat The node and the first The correlation of drying energy consumption at each node and the correlation and coupling between nodes;

[0118] Based on maximizing the recovery efficiency of waste heat from kelp drying energy consumption, a dynamic adjustment and optimization model for kelp drying energy consumption is established.

[0119] The expression for the dynamic adjustment and optimization model of kelp drying energy consumption is as follows:

[0120]

[0121] In the formula, Minimizes energy consumption for kelp drying. This represents the total energy consumption for drying kelp. For recovery efficiency, is a constant term. The hot melt ratio, The outlet temperature for waste heat recovery during kelp drying energy consumption. The required temperature for the waste heat recovery node of kelp drying energy consumption.

[0122] Furthermore, based on the same inventive concept as the aforementioned method for optimizing kelp drying energy consumption by combining big data comprehensive analysis, this solution proposes a kelp drying energy consumption optimization system by combining big data comprehensive analysis, comprising:

[0123] The data acquisition module is used to deploy a multimodal sensor array in each kelp drying process. By setting up edge data acquisition nodes, it can collect equipment operating status data and environmental parameters in real time. Using Internet of Things (IoT) technology, it can receive the data collected by each edge data acquisition node in real time and establish a kelp drying energy consumption database.

[0124] The prediction and optimization module is used to construct a kelp drying energy consumption assessment model based on the LSTM-Transformer-GNN hybrid machine learning algorithm, comprehensively assess and predict kelp drying energy consumption; obtain the weight value of drying energy consumption of each edge data acquisition node according to the assessment results of the kelp drying energy consumption assessment model; obtain the correlation of drying energy consumption of each edge data acquisition node, and dynamically adjust and optimize kelp drying energy consumption through waste heat recovery optimization;

[0125] The data acquisition module includes:

[0126] The data acquisition unit is used to deploy a multimodal sensor array in each drying process of kelp, and to collect equipment operating status data and environmental parameters in real time by setting edge data acquisition nodes;

[0127] The data processing unit is used to receive data collected by each edge data acquisition node in real time using Internet of Things technology, and to establish a kelp drying energy consumption database.

[0128] The prediction and optimization module includes:

[0129] An evaluation and prediction unit is used to construct a kelp drying energy consumption evaluation model based on the LSTM-Transformer-GNN hybrid machine learning algorithm, and to comprehensively evaluate and predict the kelp drying energy consumption.

[0130] The node weighting unit is used to obtain the weight value of the drying energy consumption of each edge data acquisition node based on the evaluation results of the kelp drying energy consumption evaluation model.

[0131] An energy consumption adjustment unit is used to obtain the correlation of drying energy consumption of each edge data acquisition node, and dynamically adjust and optimize the kelp drying energy consumption through waste heat recovery optimization.

[0132] In summary, the advantages of this invention are: by using multimodal sensing data and a hybrid machine learning model, the energy consumption of kelp drying is dynamically evaluated and waste heat recovery is optimized, significantly reducing kelp production capacity loss.

[0133] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing the drying energy consumption of kelp in combination with big data comprehensive analysis, characterized in that, The application relates to a method for dynamically adjusting and optimizing seaweed drying energy consumption. The method comprises the following steps: Deploying a multi-modal sensor array in each drying process of seaweed, collecting equipment running state data and environmental parameters in real time through setting edge data collection nodes; Receiving the data collected by each edge data collection node in real time by using Internet of Things technology, and establishing a seaweed drying energy consumption database; Constructing a seaweed drying energy consumption evaluation model based on an LSTM-Transformer-GNN hybrid machine learning algorithm, and comprehensively evaluating and predicting seaweed drying energy consumption; According to the evaluation result of the seaweed drying energy consumption evaluation model, the weight value of the drying energy consumption of each edge data collection node is obtained; According to the correlation of the drying energy consumption of each edge data collection node, the seaweed drying energy consumption is dynamically adjusted and optimized through waste heat recovery optimization. The method comprises the following steps: According to the seaweed drying process cycle, the window length of the LSTM layer time sequence is set, wherein the window length of the LSTM layer time sequence is within the range of the seaweed drying process cycle; According to the seaweed drying energy consumption database, the data collected by each edge data collection node is taken as an input quantity; Based on the LSTM neural network, the equipment running state data and environmental parameters with time sequence are outputted, and the equipment running state data and environmental parameters with time sequence are taken as the input quantity of the Transformer model; Through linear projection, the data collected by each edge data collection node in real time is mapped to the same dimension space as Key / Value; According to the Transformer model, the correlation between each data and other data in the input quantity is calculated, and the dependency relationship between any data in the input quantity data is captured; According to the seaweed process parameter knowledge graph, a constraint loss function of the LSTM-Transformer model is established based on a mean square error formula; According to the LSTM-Transformer model, the seaweed drying energy consumption value of each edge data collection node and the total seaweed drying energy consumption value within the window length are judged and predicted; According to the drying process of each edge data collection node in the seaweed drying process, the side weight is adjusted in real time based on the CNN neural network, the influence of drying scheduling changes on energy consumption is accurately reflected, and the seaweed drying energy consumption value is further corrected and optimized. ； wherein, is the window length of the LSTM layer timing, is the drying process cycle of kelp, is the dominant angular frequency in the drying data of kelp, is the floor symbol, ensuring that the window length is an integer time unit; The window length expression of the LSTM layer time sequence is: ； In the formula, is the constraint loss function value of the LSTM-Transformer model, is the number of sample groups, is the model kelp drying energy consumption prediction value, is the kelp drying energy consumption target value, is the weight coefficient of the constraint term in the kelp process parameter knowledge graph, is the sample set with the constraint term in the kelp process parameter knowledge graph, is the standard boundary threshold in the kelp process parameter knowledge graph. The constraint loss function expression of the LSTM-Transformer model is: According to the evaluation result of the seaweed drying energy consumption evaluation model, the weight value of the drying energy consumption of each edge data collection node is obtained. According to the evaluation result of the seaweed drying energy consumption evaluation model, the seaweed drying energy consumption value of each edge data collection node within the window length is obtained; According to the seaweed drying energy consumption evaluation model, the weights in the LSTM-Transformer-GNN hybrid model are extracted respectively. Wherein, the LSTM is the historical data weight of each edge data collection node, the Transformer is the self-attention weight of each edge data collection node, and the GNN is the spatial weight of each edge data collection node. According to the weighted formula, the comprehensive weight value of the drying energy consumption of each edge data acquisition node is obtained; According to the Softmax normalization formula, the comprehensive weight value of the drying energy consumption of each edge data acquisition node is normalized, so that the sum of all comprehensive weight values is 1; The comprehensive weight value of the drying energy consumption of each edge data acquisition node is expressed as: ； In the formula, is a comprehensive weight value of the drying energy consumption of the edge data acquisition node, , , respectively are a historical data weight of the edge data acquisition node, a self-attention weight of the edge data acquisition node, and a spatial weight of the edge data acquisition node, , , respectively are bias coefficients of the historical data weight, the self-attention weight, and the spatial weight of the edge data acquisition node, and are determined through cross-validation. 2.The kelp drying energy consumption optimization method combined with big data comprehensive analysis according to claim 1, wherein, The multi-modal sensor array is deployed in each drying process of kelp, and edge data acquisition nodes are set to collect equipment operating state data and environmental parameters in real time, which specifically includes: According to the drying process of kelp, a multi-modal sensor array is deployed in each drying process to collect equipment operating state data and environmental parameters in real time during the process; In each drying process, edge data acquisition nodes are set to filter and normalize the real-time collected equipment operating state data and environmental parameters; A simultaneous time collection mode is adopted to make each edge data acquisition node operate synchronously under a unified time reference, ensuring the time consistency and synchronization of the collected data; For each edge data acquisition node, a standardized communication format specification is uniformly set to ensure that the data uploaded by different nodes are highly consistent in format and structure. 3.The kelp drying energy consumption optimization method combined with big data comprehensive analysis according to claim 2, characterized in that, The Internet of Things technology is used to receive the data collected by each edge data acquisition node in real time, and a kelp drying energy consumption database is established, which specifically includes: The Internet of Things technology is used to receive the data collected by each edge data acquisition node in real time; Interpolation and data downsampling techniques are used to process the data of each edge data acquisition node, achieving alignment and position consistency of the data in time or space; According to the process requirements of kelp drying, the boundary constraints of the data collected by each edge data acquisition node are extracted to establish a kelp process parameter knowledge graph; Based on the historical data of kelp drying and the kelp process parameter knowledge graph, a kelp drying energy consumption database is established.

4. The kelp drying energy consumption optimization method combined with big data comprehensive analysis according to claim 3, characterized in that, The correlation of the drying energy consumption of each edge data acquisition node is obtained through waste heat recovery optimization, and the kelp drying energy consumption is dynamically adjusted and optimized, which specifically includes: According to the kelp drying energy consumption database, the equipment operating state data and environmental parameters of each edge data acquisition node are obtained; Based on the Pearson correlation formula, the correlation of the drying energy consumption of each edge data acquisition node is determined according to the equipment operating state data and environmental parameters of each edge data acquisition node and the drying energy consumption data of each edge data acquisition node; According to the correlation of the drying energy consumption of each edge data acquisition node and the correlation coupling of each node, a dynamic adjustment and optimization model of kelp drying energy consumption is established; According to the dynamic adjustment and optimization model of kelp drying energy consumption, the start and stop and parameters of waste heat recovery are dynamically adjusted.

5. A kelp drying energy consumption optimization system combined with big data comprehensive analysis, characterized in that, The method for optimizing the drying energy consumption of kelp based on big data comprehensive analysis according to any one of claims 1-4, comprising: A data acquisition module is used to deploy a multi-modal sensor array in each drying process of kelp, set edge data acquisition nodes to collect equipment operating state data and environmental parameters in real time, use Internet of Things technology to receive the data collected by each edge data acquisition node in real time, and establish a kelp drying energy consumption database; A prediction and optimization module is configured to construct a kelp drying energy consumption evaluation model based on an LSTM-Transformer-GNN hybrid machine learning algorithm, comprehensively evaluate and predict kelp drying energy consumption, obtain a weight value of the drying energy consumption of each edge data acquisition node according to an evaluation result of the kelp drying energy consumption evaluation model, obtain the correlation of the drying energy consumption of each edge data acquisition node, and dynamically adjust and optimize the kelp drying energy consumption through waste heat recovery optimization.

Images